WHOLE-CABIN BOARDING SIMULATION
DETERMINISTIC STRESS DATASETS FOR MULTI-REGION DEPENDENCY EVALUATION
Version: Stress Dataset Research Log v1

Purpose
-------
This document begins the final controlled validation programme for the completed
Reactive Dependency-Region Management Architecture.

The preceding architecture report is now frozen. Its final implementation
provides reactive congestion detection, persistent dependency-region lifecycle
tracking, three-tick confirmation, state-change-aware evaluation and hierarchical
dependency-region prioritisation.

The present programme does not redesign that architecture. It investigates
whether deterministic input datasets can be constructed to produce simultaneous
confirmed dependency regions and thereby exercise architectural capabilities that
rarely appeared in the conventional benchmark scenarios.

The architecture is therefore the fixed reference implementation. Dataset
construction becomes the independent variable.


======================================================================
RESEARCH BOUNDARY
======================================================================

This research does not reopen or replace the conclusions of the completed
Reactive Dependency-Region Management Architecture.

It introduces one new independent research variable:

    DETERMINISTIC STRESS DATASET CONSTRUCTION

The programme may alter deterministic passenger manifests, seat-distribution
patterns and queue ordering only when those changes are explicitly documented as
properties of the dataset under test.

It does not introduce:

- probabilistic human behaviour;
- passenger hesitation or personal judgement;
- asynchronous tick execution;
- overtaking;
- reverse aisle movement;
- aisle switching;
- movement beyond an assigned row;
- seat reassignment;
- predictive congestion management;
- advance aisle-space reservation;
- airline policy unless later isolated as a separate deterministic dataset
  variable.

The completed architecture must process each dataset without being changed to
accommodate the expected result.


======================================================================
FIXED REFERENCE ARCHITECTURE
======================================================================

The following architectural components remain enabled and unchanged throughout
this research programme:

- synchronous global movement;
- maximum one forward aisle tile per passenger per tick;
- one unique assigned seat per passenger;
- fixed serving aisle for every passenger;
- no movement beyond the assigned row;
- deterministic stand-yield-sit-reseat transactions;
- middle-bank reservation;
- dependency-aware precedence;
- adaptive rear-boundary forward yield;
- guarded forward-yield checks;
- row-bounded congestion clustering;
- persistent dependency-region lifecycle tracking;
- three-tick region confirmation;
- state-change-gated evaluation;
- deterministic hierarchical region prioritisation;
- 1,500-tick no-seat-progress stall detector;
- 30,000 maximum ticks.

Any later change to one of these components would constitute a new architecture
programme and must not be presented as a stress-dataset experiment.


======================================================================
CONTROLLED CONDITIONS
======================================================================

Within every paired experiment, the following must remain identical unless the
experiment explicitly identifies one of them as the dataset variable:

- cabin configuration;
- total seat capacity;
- occupancy rate;
- scenario seed;
- passenger count;
- entry-headway range;
- synchronous execution;
- passenger movement rules;
- seat-event prerequisites;
- blocker movement;
- yield-space rules;
- middle-bank reservation;
- dependency-region detection;
- lifecycle matching, updating and dissolution;
- confirmation threshold;
- state-change signature;
- region-priority hierarchy;
- stall and maximum-tick limits;
- output instrumentation.

Experiment 1 changes the deterministic dataset family while preserving the
entire architecture and the scenario-level cabin, occupancy, headway and seed.


======================================================================
DATASET CONSTRUCTION CONTRACT
======================================================================

A deterministic stress dataset is a reproducible passenger manifest and entry
ordering designed to increase the opportunity for demanding dependency-region
structures without modifying passenger movement after admission.

All stress datasets must obey the following rules:

1. Every passenger retains one unique assigned seat.
2. Every passenger retains one fixed serving aisle.
3. Dataset construction completes before simulation movement begins.
4. Dataset generation is reproducible from the recorded seed.
5. The boarding engine receives a completed manifest and queue; it does not know
   why a passenger was selected or ordered.
6. No stress generator may create two passengers for the same seat.
7. No generator may place a passenger directly into the aisle or a seat-event
   state.
8. No generator may alter movement speed, event duration or blocker rules.
9. No generator may guarantee or fabricate a dependency region.
10. Failure to create simultaneous regions is a valid dataset result and must be
    documented honestly.


======================================================================
NEW TERMINOLOGY
======================================================================

Deterministic Stress Dataset
----------------------------
A reproducible passenger manifest and entry-order dataset deliberately designed
to expose demanding but valid congestion structures to the frozen architecture.

Standard Deterministic Dataset
------------------------------
The established reference generator using seeded random seat selection and
independently shuffled aisle queues.

Stress Dataset Family
---------------------
A named group of deterministic datasets sharing the same construction principle,
such as row-zone concentration, blocker-prone seating or alternating aisle
admission.

Dataset Characteristic
----------------------
A measurable property of an input dataset, such as row-zone concentration,
blocker-prone seat share or left/right aisle balance.

Confirmed Region Concurrency
----------------------------
The number of dependency regions that have simultaneously survived the
three-tick confirmation threshold.

Region Competition
------------------
A tick on which more than one confirmed, state-change-eligible dependency region
requires hierarchical evaluation.

Competition Window
------------------
A continuous tick interval during which two or more confirmed dependency regions
coexist.

Priority Utilisation
--------------------
Observed use of the completed hierarchical prioritisation layer to rank genuine
simultaneous confirmed regions.

Negative Dataset Finding
------------------------
A valid result in which a stress dataset does not produce the intended region
competition. Such a result informs later dataset refinement without justifying
an architecture change.


======================================================================
RESEARCH OBJECTIVES
======================================================================

1. Construct reproducible deterministic datasets capable of producing multiple
   dependency regions within the same simulation.
2. Progress from detected-region concurrency to simultaneous confirmed-region
   concurrency.
3. Produce genuine priority competitions without modifying the completed
   architecture.
4. Identify which input characteristics increase or reduce multi-region
   dependency formation.
5. Separate dataset effects from entry headway and all movement rules.
6. Retain negative findings when apparently stressful datasets simplify rather
   than intensify congestion.
7. Build a reusable catalogue of deterministic benchmark datasets for future
   validation.
8. Complete a final controlled test of the frozen architecture.


======================================================================
EXPERIMENT 1 - BASELINE DETERMINISTIC STRESS DATASET CONSTRUCTION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit Amlani's original movement algorithm operates on an already constructed
one-dimensional occupancy dataset. The completed boarding architecture similarly
processes the manifest and aisle queues that it receives.

Experiment 1 returns to this foundational distinction. It does not change the
movement algorithm. It changes only the deterministic data supplied to that
algorithm and observes whether the same fixed architecture exposes different
congestion structures.

Purpose
-------
Establish a reusable deterministic stress-dataset generator and compare its
first conservative stress family against the standard deterministic dataset.

The first experiment is primarily a foundation and calibration experiment. It
is not required to produce simultaneous confirmed dependency regions. Its role
is to verify that:

- dataset construction is reproducible;
- the frozen architecture remains unchanged;
- paired datasets retain the same scenario-level cabin, occupancy, headway and
  seed;
- dataset characteristics are recorded explicitly;
- region-concurrency and priority evidence can be compared cleanly.

Paired Modes
------------
Each of the 30 scenarios is executed twice.

1. STANDARD DETERMINISTIC DATASET

   The reference generator selects occupied seats from a seeded shuffled seat
   catalogue and independently shuffles the left- and right-aisle queues.

2. BASELINE DETERMINISTIC STRESS DATASET

   The frozen architecture receives a reproducible composite stress manifest.
   Passenger seats are preferentially selected from two separated row zones,
   blocker-prone seats receive greater selection weight and each serving-aisle
   queue is ordered deterministically around the same two zones.

The second mode changes input data only. The architecture, movement contract,
confirmation threshold, state-change gate and priority hierarchy are identical.

Independent Variable
--------------------

    DATASET GENERATION FAMILY

- standard seeded manifest and shuffled queues; versus
- baseline deterministic two-zone stress manifest and stress-ordered queues.

Baseline Stress Family Design
-----------------------------
The first stress family uses three documented dataset characteristics:

1. TWO SEPARATED ROW ZONES

   Seats near approximately one-third and two-thirds of cabin depth receive
   greater deterministic selection weight. The separation is intended to create
   more than one local area of potential congestion rather than one continuous
   cabin-wide concentration.

2. BLOCKER-PRONE SEAT EMPHASIS

   Window and deeper seats receive greater weight than aisle-adjacent seats. This
   increases the potential for valid seated blockers without changing any seat
   event after the manifest is created.

3. DETERMINISTIC ZONE-ORIENTED QUEUE ORDER

   Each aisle queue is ordered reproducibly so passengers associated with the
   two stress zones enter under the existing headway mechanism. The generator
   does not bypass headway and does not insert passengers directly into the
   aisle.

This is intentionally a composite baseline family. Later experiments may isolate
or ablate individual characteristics after the generator and evidence model have
been validated.

Controlled Conditions
---------------------

- identical cabin configuration within each pair;
- identical occupancy rate;
- identical passenger count;
- identical entry-headway range;
- identical scenario seed;
- identical frozen architecture;
- identical synchronous tick execution;
- identical movement constraints;
- identical stall detector and maximum ticks;
- no human behaviour;
- no asynchronous execution;
- no reverse movement, overtaking, aisle switching or assigned-row overshoot.

Dataset Evidence Collected
--------------------------

- dataset family;
- row-zone concentration percentage;
- blocker-prone seat share;
- left/right aisle-assignment balance;
- passenger count and occupancy;
- entry-headway range;
- dependency regions detected;
- regions created, updated, confirmed and dissolved;
- maximum simultaneous active regions;
- regions active at termination;
- state-change evaluations suppressed and reopened;
- priority competitions;
- lower-priority deferrals;
- cluster selections and cluster-started events;
- complete cabins and passengers seated;
- simulation ticks;
- passengers remaining outside;
- residual stall family.

Experimental Hypothesis
-----------------------
A deterministic two-zone, blocker-emphasised dataset may increase spatially
separated dependency activity relative to the standard dataset. However, the
same ordering may also simplify flow by making local seat events complete more
coherently. Either result is scientifically valid.

The primary Experiment 1 question is therefore not whether the stress dataset
improves or worsens boarding. It is whether the dataset layer is reproducible,
measurable and capable of changing dependency-region evidence without changing
the architecture.

Interpretation Safeguards
-------------------------

- A faster stress run is not automatically a failed experiment.
- A slower stress run is not automatically evidence of better stress.
- More detected regions do not necessarily mean more confirmed concurrent
  regions.
- Multiple regions created at different times do not constitute simultaneous
  dependency competition.
- Priority utilisation requires genuine simultaneous confirmed eligible regions.
- The architecture must not be changed merely because the first stress family
  fails to create competition.
- Any improvement or regression in boarding outcome must be reported as a
  dataset effect.
- Composite stress characteristics must be separated in later experiments before
  causal claims are made about a single characteristic.

Primary Evaluation Order
------------------------

1. Correct experiment identity and deterministic replay.
2. Dataset-characteristic differences between paired modes.
3. Maximum simultaneous active regions.
4. Confirmed-region concurrency.
5. Genuine priority competitions and deferrals.
6. Regions created, confirmed and dissolved.
7. Cluster selections and cluster-started events.
8. Complete cabins and passengers seated.
9. Residual stall classifications.
10. Simulation ticks as a secondary measure.

Expected Outcomes
-----------------

Outcome A - Simultaneous confirmed regions appear

The baseline stress family immediately exercises hierarchical prioritisation and
provides a foundation for controlled refinement.

Outcome B - More regions appear, but not simultaneously

The generator increases congestion activity but requires stronger temporal
alignment in a later dataset.

Outcome C - Fewer regions appear

The deterministic zone ordering simplifies cabin flow. This becomes a useful
negative finding and suggests later experiments should separate seat selection
from queue ordering.

Outcome D - No meaningful difference appears

The baseline generator is too conservative. Later experiments may increase one
dataset characteristic at a time while retaining the same frozen architecture.

Diagram Plan
------------
A diagram should be considered only after results are available. A valid diagram
may compare:

- the standard dispersed manifest;
- the two separated stress zones;
- blocker-prone seat emphasis;
- left/right deterministic queues;
- observed region timelines.

The diagram must distinguish input construction from passenger movement and must
not imply that dependency regions were inserted directly into the simulation.


======================================================================
RUNNING CONCLUSIONS
======================================================================

1. The Reactive Dependency-Region Management Architecture is frozen.
2. Dataset construction is now the independent research variable.
3. Experiment 1 establishes the reusable paired stress-dataset framework.
4. The first stress family combines two separated row zones, blocker-prone seat
   emphasis and deterministic zone-oriented queue ordering.
5. Simultaneous dependency competition must arise from valid deterministic
   simulation state; it must not be manufactured by changing the architecture.
6. Human behaviour, asynchronous execution, backward movement, overtaking,
   aisle switching and assigned-row overshoot remain excluded.
7. Negative dataset findings are retained as evidence for later refinement.
8. Future experiments should isolate individual stress characteristics after the
   baseline generator has been validated.



======================================================================
EXPERIMENT 1 FINDINGS - BASELINE DETERMINISTIC STRESS DATASET CONSTRUCTION
======================================================================

Execution Summary
-----------------
Experiment 1 completed 30 deterministic paired scenarios (60 executions) using
the same fixed architecture, cabin characteristics, occupancy, headway and
scenario seeds in each pair.

The standard dataset and baseline composite stress dataset each completed 27 of
30 cabins. However, the internal evidence differed substantially.

Overall Paired Result
---------------------

- Standard-dataset average completion: 99.71%.
- Composite stress-dataset average completion: 96.82%.
- Standard complete cabins: 27 of 30.
- Composite stress complete cabins: 27 of 30.
- Improved stress pairs: 3.
- Unchanged pairs: 24.
- Worse stress pairs: 3.
- Standard total passengers seated: 7,649.
- Composite stress total passengers seated: 7,391.
- Net seated-passenger difference: -258.
- Standard dependency regions detected: 37.
- Composite stress dependency regions detected: 13.
- Standard regions reaching confirmation: 22.
- Composite stress regions reaching confirmation: 9.
- Standard cluster selections: 25.
- Composite stress cluster selections: 6.
- Standard cluster-started events: 31.
- Composite stress cluster-started events: 15.
- Genuine priority competitions: 0 in both modes.

Principal Findings
------------------

1. The dataset-generation layer was successfully separated from the architecture.

Both paired modes used the frozen Reactive Dependency-Region Management
Architecture. The experiment therefore confirmed that manifests and aisle queues
can be varied reproducibly without changing movement, clustering, lifecycle,
confirmation, state-change gating or prioritisation.

2. The composite stress family did not create multi-region competition.

No simultaneous confirmed-region priority competition occurred. The dataset
therefore did not yet exercise the hierarchical prioritisation layer.

3. The composite stress family reduced dependency-region activity.

The stress mode produced 13 detected regions compared with 37 under the standard
dataset, and only 9 confirmed regions compared with 22. Cluster selections and
cluster-started events were also lower.

4. Deterministic concentration can simplify flow rather than intensify it.

The combination of two row zones, blocker-prone seat weighting and zone-oriented
queue ordering appears to have created more coherent local seating sequences in
many scenarios. This reduced the fragmented dependency activity that the stress
family was intended to increase.

5. Boarding outcome alone is not a sufficient stress measure.

Both modes completed 27 cabins, yet the stress mode seated 258 fewer passengers
across the three incomplete cases. Conversely, three standard stalls converted
to complete stress runs. The mixed result confirms that completion must be read
alongside region creation, confirmation and competition evidence.

6. The negative result provides a clear experimental refinement.

Because Experiment 1 changed row concentration, blocker-prone seat share and
queue ordering together, no single characteristic can be identified as the
cause of reduced region activity. The next experiment should isolate row-zone
concentration while returning seat-depth weighting and queue ordering to the
standard deterministic method.

Experiment 1 Conclusion
-----------------------
The reusable deterministic stress-dataset framework is validated. The first
composite family did not produce simultaneous dependency competition and often
simplified congestion structure. This is retained as a valid negative dataset
finding. Experiment 2 will isolate row-zone concentration as the sole dataset
variable.


======================================================================
EXPERIMENT 2 - CONTROLLED ROW-ZONE CONCENTRATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit Amlani's original movement algorithm processes the occupancy dataset it is
given without changing the rules of movement in response to how that dataset was
constructed. Experiment 2 follows the same separation by changing only where
assigned seats are concentrated before movement begins.

Purpose
-------
Determine whether two separated row zones, by themselves, increase spatially
separated dependency activity when blocker-prone seat weighting and specialised
queue ordering are removed.

Paired Modes
------------
Each of the 30 scenarios is executed twice using identical cabin configuration,
occupancy, passenger count, headway, scenario seed and frozen architecture.

1. STANDARD DETERMINISTIC DATASET

   Occupied seats are selected from a seeded shuffled seat catalogue. Left- and
   right-aisle queues are independently shuffled.

2. CONTROLLED ROW-ZONE CONCENTRATION DATASET

   Occupied seats are selected preferentially from two separated row zones near
   one-third and two-thirds of cabin depth. All seat positions inside those zones
   receive equal treatment; window and deep seats receive no additional weight.
   Left- and right-aisle queues are independently shuffled exactly as in the
   standard mode.

Independent Variable
--------------------

    ROW-ZONE CONCENTRATION

- dispersed seeded seat selection; versus
- deterministic concentration around two separated row zones.

Removed Experiment 1 Composite Variables
----------------------------------------

The following Experiment 1 characteristics are deliberately removed:

- blocker-prone seat emphasis;
- zone-oriented queue ordering;
- blocker-first queue tie-breaking.

This permits a narrower causal interpretation of any change in region evidence.

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy and passenger count within each pair;
- identical entry-headway range;
- identical scenario seed;
- independently shuffled aisle queues in both modes;
- identical seat-event and blocker rules;
- synchronous execution;
- no backward movement, overtaking, aisle switching or assigned-row overshoot;
- no human behaviour or asynchronous execution.

Evidence Collected
------------------

- row-zone concentration percentage;
- blocker-prone seat share as a diagnostic control;
- aisle-assignment balance;
- dependency regions detected;
- regions created, confirmed and dissolved;
- maximum simultaneous active regions;
- confirmed-region concurrency;
- priority competitions and lower-priority deferrals;
- cluster selections and starts;
- complete cabins and passengers seated;
- residual stall family;
- simulation ticks as a secondary measure.

Experimental Hypothesis
-----------------------
If separated spatial concentration is sufficient to create independent local
congestion, the row-concentrated dataset should increase concurrent region
activity despite retaining standard shuffled queue order. If region activity
remains lower, row concentration itself may promote coherent completion rather
than competition.

Interpretation Safeguards
-------------------------

- Two row zones do not constitute two dependency regions.
- Regions created at different times do not constitute concurrency.
- A priority competition requires simultaneous confirmed and eligible regions.
- Changes in blocker-prone seat share must be reported, even though no explicit
  blocker weighting is used.
- The architecture must not be changed to force the expected result.
- A negative result will motivate a later isolated seat-depth or admission-order
  experiment rather than reopening the architecture.

Primary Evaluation Order
------------------------

1. Dataset identity and deterministic execution.
2. Row-zone concentration difference.
3. Maximum simultaneous active and confirmed regions.
4. Genuine priority competitions.
5. Regions detected and confirmed.
6. Cluster selections and starts.
7. Complete cabins and seated totals.
8. Residual stall families.
9. Simulation ticks as a secondary measure.

Diagram Plan
------------
A later diagram may compare dispersed seat selection with two separated row
concentration zones while showing that both modes retain independently shuffled
left- and right-aisle queues. It must not depict dependency regions as inserted
or guaranteed by the dataset generator.


======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 1
======================================================================

1. The frozen architecture remains the fixed reference implementation.
2. Experiment 1 validated the reusable deterministic dataset layer.
3. The first composite stress family reduced rather than increased region
   activity and produced no priority competition.
4. Composite dataset variables must now be separated before causal claims are
   made.
5. Experiment 2 isolates row-zone concentration and restores standard shuffled
   queue ordering and neutral seat-depth selection.
6. Simultaneous dependencies must emerge from valid deterministic state.
7. Human behaviour, asynchronous execution and movement-rule changes remain
   excluded.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v2 - EXPERIMENT 2 DESIGN
======================================================================

======================================================================
EXPERIMENT 2 FINDINGS - CONTROLLED ROW-ZONE CONCENTRATION
======================================================================

Purpose Reviewed
----------------
Experiment 2 isolated deterministic row-zone concentration from the composite
stress variables introduced in Experiment 1. Both paired modes retained the
frozen Reactive Dependency-Region Management Architecture, independently
shuffled aisle queues, neutral seat-depth selection and identical movement
rules. Only the spatial distribution of occupied rows changed.

Overall Paired Result
---------------------

- Scenario pairs: 30.
- Scenario executions: 60.
- Standard average completion: 99.71%.
- Row-concentration average completion: 98.92%.
- Improved row-concentration pairs: 2.
- Unchanged pairs: 22.
- Worse row-concentration pairs: 6.
- Net seated-passenger difference: -75.
- Standard complete cabins: 27 of 30.
- Row-concentration complete cabins: 24 of 30.
- Standard stalls converted to complete cabins: 2.
- Standard dependency regions: 37.
- Row-concentration dependency regions: 36.
- Standard confirmed regions: 22.
- Row-concentration confirmed regions: 13.
- Standard priority competitions: 0.
- Row-concentration priority competitions: 7.
- Maximum simultaneous active regions: 2 in both modes.

Principal Findings
------------------

1. Row-zone concentration changed congestion geometry without consistently
   increasing overall dependency-region volume.

The row-concentrated datasets produced almost the same total number of detected
regions as the standard datasets, but fewer regions reached confirmation. This
shows that spatial concentration can change the persistence and timing of
congestion even when total region counts remain similar.

2. The isolated variable produced both beneficial and harmful outcomes.

Two standard stalls converted to complete cabins, while six paired scenarios
seated fewer passengers under row concentration. The result therefore does not
support row concentration as a generally beneficial or generally harmful
boarding arrangement.

3. Genuine simultaneous-region competition was observed.

Unlike the previous benchmark and Experiment 1, the row-concentration datasets
produced seven priority competitions. This is the first direct evidence within
the stress-dataset programme that deterministic manifest construction alone can
exercise the hierarchical prioritisation layer of the frozen architecture.

4. Concurrency did not require human behaviour or asynchronous ticks.

The priority competitions emerged under synchronous deterministic movement,
fixed headway rules and unchanged architecture. Dataset geometry alone was
sufficient to create overlapping confirmed-region eligibility.

5. Row concentration is therefore a useful stress component, but not a complete
   causal explanation.

Because row concentration also changed which seats happened to be selected, the
next experiment must isolate seat-depth distribution directly while restoring
dispersed row selection. This will determine whether blocker-prone seats increase
confirmed-region activity independently of row concentration.

Experiment 2 Conclusion
-----------------------
Controlled row-zone concentration is supported as a valid deterministic stress
variable. It generated the first seven genuine priority competitions without
changing the frozen architecture, human behaviour or synchronous execution.
However, its mixed completion effects and reduced confirmation count show that
it should be treated as one stress-building component rather than a complete
solution for repeatable multi-region generation.


======================================================================
EXPERIMENT 3 - CONTROLLED BLOCKER-PRONE SEAT CONCENTRATION
======================================================================

Connection to the Completed Architecture
----------------------------------------
The Reactive Dependency-Region Management Architecture remains frozen. This
experiment changes only the manifest's seat-depth distribution so that the
existing blocker, yield-space and seat-event rules are exercised more often.
The architecture is not informed that a stress dataset is in use and receives
only the completed deterministic passenger manifest.

Purpose
-------
Determine whether deterministic concentration of window and deep inner seats
increases persistent blocker-dependent seat events, confirmed dependency regions
and simultaneous priority competitions when row selection and aisle queues
remain dispersed.

Controlled-Variable Clarification
---------------------------------
Increasing the probability of blocker-prone assigned seats does not alter a
controlled movement variable. The following remain controlled and unchanged:

- the definition of a blocker;
- the stand-yield-sit-reseat transaction;
- the number of yield tiles required by a given seat event;
- blocker movement direction;
- synchronous tick execution;
- passenger movement speed;
- entry-headway range;
- clustering, lifecycle, confirmation, gating and prioritisation rules.

The assigned-seat distribution is intentionally the independent dataset
variable in Experiment 3. Therefore individual seat assignments cannot remain
identical between the two paired modes; instead, cabin type, occupancy,
passenger count, headway, scenario seed and all architecture and movement rules
remain controlled. This distinction is essential to the new research programme,
where dataset construction is the subject under investigation.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

   Occupied seats are selected from a seeded shuffled catalogue. Left- and
   right-aisle queues are independently shuffled.

2. CONTROLLED BLOCKER-PRONE SEAT CONCENTRATION DATASET

   Occupied seats are selected preferentially by deterministic blocker potential.
   Window seats and deep inner seats are selected before easier aisle-side seats.
   Rows remain dispersed through seeded tie-breaking, and both aisle queues are
   independently shuffled exactly as in the standard mode.

Independent Variable
--------------------

    ASSIGNED-SEAT DEPTH DISTRIBUTION

- seeded neutral seat selection; versus
- deterministic concentration of blocker-prone window and deep inner seats.

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin configuration within each pair;
- identical occupancy rate and passenger count;
- identical entry-headway range;
- identical scenario seed;
- dispersed row selection in both modes;
- independently shuffled aisle queues in both modes;
- unchanged seat-event, blocker and yield-space rules;
- synchronous execution;
- no backward movement, overtaking, aisle switching or assigned-row overshoot;
- no human behaviour or asynchronous execution.

Evidence Collected
------------------

- blocker-prone seat share;
- row-zone concentration as a diagnostic control;
- aisle-assignment balance;
- single- and multiple-blocker seat events;
- temporary aisle reoccupations;
- dependency regions detected;
- regions created, confirmed and dissolved;
- maximum simultaneous active regions;
- priority competitions and lower-priority deferrals;
- cluster selections and starts;
- complete cabins and passengers seated;
- residual stall family;
- simulation ticks as a secondary measure.

Experimental Hypothesis
-----------------------
If blocker-prone seat depth is an important contributor to independent congestion
zones, the concentrated dataset should increase multiple-blocker events,
confirmed dependency regions or priority competitions while row concentration
remains near the standard distribution. A null result would indicate that seat
depth alone is insufficient and that later experiments should combine it with
spatial or admission-pattern stress.

Interpretation Safeguards
-------------------------

- More window seats do not automatically constitute more dependency regions.
- More blocker events do not automatically constitute simultaneous regions.
- A priority competition requires simultaneous confirmed and eligible regions.
- Row-zone concentration must remain a diagnostic control and be reported.
- Different seat assignments are the intended independent-variable change, not
  a breach of the movement contract.
- The architecture must not be modified to force the expected outcome.
- Any worse completion outcome must be reported honestly as a dataset effect.

Primary Evaluation Order
------------------------

1. Dataset identity and deterministic execution.
2. Blocker-prone seat-share difference.
3. Row-zone concentration control.
4. Multiple-blocker events and temporary aisle reoccupations.
5. Confirmed regions and maximum simultaneous regions.
6. Genuine priority competitions.
7. Cluster selections and starts.
8. Complete cabins and seated totals.
9. Residual stall families.
10. Simulation ticks as a secondary measure.

Diagram Plan
------------
A later diagram may compare neutral seat selection with deterministic
window/deep-seat concentration across dispersed rows. It must show that seat-event
rules remain unchanged and must not depict blockers, dependency regions or
priority competitions as inserted directly by the dataset generator.


======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 2
======================================================================

1. The frozen architecture remains the fixed reference implementation.
2. Experiment 1 established the reusable deterministic dataset layer.
3. Experiment 2 generated seven genuine priority competitions through row-zone
   concentration alone.
4. Simultaneous dependencies therefore do not require human behaviour or
   asynchronous execution.
5. Row concentration produced mixed boarding outcomes and fewer confirmed
   regions overall, so it remains a stress component rather than a complete
   benchmark solution.
6. Experiment 3 isolates assigned-seat depth while restoring dispersed row
   selection and standard shuffled queues.
7. Seat-event rules remain controlled; only the deterministic seat-assignment
   distribution changes.
8. No passenger may pass the assigned row or move backwards.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v3 - EXPERIMENT 3 DESIGN
======================================================================


======================================================================
EXPERIMENT 3 FINDINGS - CONTROLLED BLOCKER-PRONE SEAT CONCENTRATION
======================================================================

Purpose Reviewed
----------------
Experiment 3 isolated deterministic assigned-seat depth while restoring dispersed
row selection and independently shuffled aisle queues. The completed Reactive
Dependency-Region Management Architecture remained frozen. The independent
variable was the proportion of window and deep inner seats selected into the
passenger manifest.

Execution Summary
-----------------
Thirty paired scenarios were executed, producing sixty deterministic runs.

Across the complete batch:

- Standard average completion: 99.71%.
- Blocker-prone dataset average completion: 98.96%.
- Improved blocker-prone pairs: 3 of 30.
- Unchanged pairs: 22 of 30.
- Worse blocker-prone pairs: 5 of 30.
- Net seated-passenger difference: -55.
- Standard complete cabins: 27 of 30.
- Blocker-prone complete cabins: 25 of 30.
- Standard dependency regions: 37.
- Blocker-prone dependency regions: 45.
- Standard confirmed regions: 22.
- Blocker-prone confirmed regions: 34.
- Standard cluster selections: 25.
- Blocker-prone cluster selections: 36.
- Standard cluster-started events: 31.
- Blocker-prone cluster-started events: 41.
- Blocker-prone priority competitions: 6.
- Standard priority competitions: 0.

Principal Findings
------------------

1. Seat-depth concentration produced stronger dependency evidence than row
   concentration alone.

The blocker-prone dataset increased detected regions from 37 to 45, confirmed
regions from 22 to 34, cluster selections from 25 to 36 and cluster-started
events from 31 to 41. This supports assigned-seat depth as a meaningful
component of deterministic congestion construction.

2. Genuine hierarchical priority competition was exercised.

Six priority competitions occurred in the blocker-prone dataset. Scenario 19
provided the clearest evidence: two simultaneously active regions reached
confirmation, twelve region rankings were recorded across six competitions, and
the hierarchy selected between left- and right-aisle regions using the completed
deterministic priority score.

3. More blocker-prone seats did not produce a uniform outcome.

Three scenarios improved, twenty-two remained unchanged and five became worse.
Scenario 4 converted a standard rear-boundary stall at 200/204 into a complete
204/204 cabin, while Scenario 7 changed from complete boarding to 209/216 with a
mixed dependency residual. The result confirms that dependency topology, rather
than blocker count alone, controls the final congestion pattern.

4. The dataset changed the geometry of dependency formation.

The blocker-prone manifest produced linked row-event chains, yield-tile
occupancy dependencies, deeper critical-blocker trees and longer-lived confirmed
regions in selected scenarios. In other scenarios, the changed seat geometry
removed a previous bottleneck. Seat-depth concentration therefore changes the
structure and timing of dependencies rather than merely increasing their number.

5. Simultaneous dependencies require spatial separation as well as blocker
   density.

Although six priority competitions were observed, most blocker-prone scenarios
still contained only one active confirmed region at a time. The next experiment
should therefore retain blocker-prone seat depth but deliberately distribute it
across two separated aisle-specific row zones.

6. The frozen architecture remained controlled.

No passenger moved backwards, passed the assigned row, overtook another
passenger, changed aisle or changed assigned seat. Entry headway, synchronous
execution, seat-event prerequisites, blocker movement, lifecycle confirmation,
state-change gating and hierarchical prioritisation were unchanged.

Scientific Interpretation
-------------------------
Experiment 3 establishes Dependency Topology as a central concept for the new
research programme.

Dependency Topology is the spatial arrangement of blocker-dependent seat events
and the resulting dependency chains within a deterministic boarding dataset.
The findings demonstrate that topology cannot be inferred from blocker density
alone. A larger blocker-prone share increases the opportunity for dependencies,
but their independence, coexistence and final effect depend upon where those
seat events occur and how the two aisle flows encounter them.

Experiment 3 Conclusion
-----------------------
Controlled blocker-prone seat concentration is supported as a stronger stress
component than row-zone concentration alone. It increased region and
intervention evidence and produced six genuine hierarchy competitions. However,
it did not reliably create two independent congestion zones. The next experiment
therefore constructs an explicit dual-aisle dependency topology while preserving
the frozen architecture and ordinary deterministic movement rules.


======================================================================
NEW TERMINOLOGY FOR EXPERIMENT 4
======================================================================

Dependency Topology
-------------------
The spatial arrangement of blocker-dependent seat events and their resulting
dependency chains within a deterministic boarding dataset.

Independent Congestion Zone
---------------------------
A locally formed dependency region that remains structurally separate from
another active dependency region, allowing both to coexist without being one
continuous aisle chain.

Dual-Aisle Dependency Topology
------------------------------
A deterministic manifest design that concentrates blocker-prone assigned seats
around two separated row bands, one primarily served by each aisle, while
leaving movement and congestion-management rules unchanged.

Topology Competition
--------------------
A genuine hierarchical priority competition arising because two independently
formed confirmed regions are simultaneously eligible for evaluation.


======================================================================
EXPERIMENT 4 - DUAL-AISLE DEPENDENCY TOPOLOGY CONSTRUCTION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit Amlani's original algorithm investigates how the arrangement of occupied
positions changes movement opportunities within a one-dimensional dataset.
Experiment 4 retains that principle but varies the deterministic input geometry
across two aisle datasets. The completed movement and congestion-management
architecture remains fixed; only the assigned-seat topology supplied to it
changes.

Planned Purpose
---------------
Determine whether a deliberately separated dual-aisle seat topology can produce
more repeatable simultaneous confirmed dependency regions than blocker-prone
seat concentration alone.

Paired Modes
------------
Each scenario is executed twice using the same cabin, occupancy, entry headway,
scenario seed and frozen architecture.

1. STANDARD DETERMINISTIC DATASET

   Seats are selected using the established seeded shuffle and both aisle queues
   are independently shuffled.

2. DUAL-AISLE DEPENDENCY TOPOLOGY DATASET

   Blocker-prone assigned seats are preferentially selected around two separated
   row bands. The earlier band is primarily associated with the left aisle and
   the later band with the right aisle. Middle-bank seats are assigned
   deterministically toward the nearer side of the topology. Both entry queues
   remain independently shuffled.

Only manifest topology changes. The generator does not insert congestion,
create a region directly or determine which region the hierarchy must select.

Independent Variable
--------------------

    DUAL-AISLE SPATIAL DEPENDENCY TOPOLOGY

- standard dispersed deterministic seat selection; versus
- blocker-prone seats distributed across two separated aisle-specific row bands.

Controlled Conditions
---------------------

- completed Reactive Dependency-Region Management Architecture;
- reactive congestion detection;
- persistent lifecycle tracking;
- three-tick lifecycle confirmation;
- state-change-gated evaluation;
- hierarchical region prioritisation;
- cabin configuration and occupancy within each pair;
- scenario seed and entry-headway range;
- synchronous global movement;
- one forward aisle tile per passenger per tick;
- assigned-seat integrity;
- no movement beyond the assigned row;
- no backward movement, overtaking or aisle switching;
- unchanged seat-event, blocker, yield-space and reservation rules;
- independently shuffled entry queues.

Dataset Construction Contract
-----------------------------

1. Two separated target row bands are derived deterministically from cabin size.
2. Left-served blocker-prone seats receive preference around the earlier band.
3. Right-served blocker-prone seats receive preference around the later band.
4. Middle-bank seats use deterministic side assignment only in the topology
   dataset so that the two intended zones remain spatially distinct.
5. Passenger IDs, assigned seats and queue order remain deterministic under the
   scenario seed.
6. The generator does not pre-create region objects or alter runtime movement.
7. A priority competition is counted only when the frozen architecture observes
   two confirmed and eligible regions naturally during execution.

New Evidence Collected
----------------------

- blocker-prone seat share;
- row-zone concentration as a diagnostic control;
- aisle-assignment balance;
- dependency regions created and confirmed;
- maximum simultaneous active regions;
- scenarios containing two active regions;
- priority competitions;
- hierarchical selections;
- left- and right-aisle priority wins;
- lower-priority deferrals;
- cluster selections and starts;
- complete cabins and passengers seated;
- residual stall families;
- simulation ticks as a secondary measure.

Experimental Hypothesis
-----------------------
If simultaneous region competition depends upon spatial separation as well as
blocker density, the dual-aisle topology dataset should produce more scenarios
with two active confirmed regions and more priority competitions than either the
standard dataset or blocker-prone concentration alone.

A null result would indicate that admission timing or another deterministic
interaction is required in addition to seat topology.

Interpretation Safeguards
-------------------------

- The dataset creates favourable input geometry, not congestion objects.
- Two target zones do not prove that two dependency regions will form.
- More active regions do not automatically mean both are confirmed or eligible.
- Priority competitions must be recorded only by the frozen runtime hierarchy.
- Deterministic middle-bank side assignment is part of dataset construction and
  does not permit aisle switching during simulation.
- Any improved or worse boarding result must be reported as a topology effect.
- The architecture must not be modified to force simultaneous dependencies.

Primary Evaluation Order
------------------------

1. Dataset identity and deterministic replay.
2. Maximum simultaneous active regions.
3. Number of genuine priority competitions.
4. Regions ranked and hierarchical selections.
5. Left/right priority wins and lower-ranked deferrals.
6. Confirmed regions and region lifetimes.
7. Cluster selections and starts.
8. Complete cabins and seated totals.
9. Residual stall families.
10. Simulation ticks as a secondary measure.

Diagram Plan
------------
A later diagram may show two separated row bands, one associated primarily with
left-aisle blocker-prone seating and one with right-aisle blocker-prone seating.
It must clearly distinguish dataset geometry from runtime dependency regions and
must not imply that congestion, confirmation or priority is inserted in advance.


======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 3
======================================================================

1. The frozen architecture remains the fixed reference implementation.
2. Experiment 1 established the reusable deterministic dataset layer.
3. Experiment 2 demonstrated that row concentration can create genuine priority
   competition but does not consistently increase confirmed-region evidence.
4. Experiment 3 increased dependency regions, confirmations, cluster activity
   and produced six genuine priority competitions.
5. Blocker density alone does not determine the resulting dependency topology.
6. Experiment 4 retains blocker-prone depth but separates it across two
   aisle-specific row bands.
7. Simultaneous dependencies remain runtime observations, not objects inserted by
   the dataset generator.
8. No passenger may pass the assigned row or move backwards.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v4 - EXPERIMENT 4 DESIGN
======================================================================


======================================================================
EXPERIMENT 4 FINDINGS - DUAL-AISLE DEPENDENCY TOPOLOGY CONSTRUCTION
======================================================================

Purpose Reviewed
----------------
Experiment 4 tested whether two spatially separated, aisle-specific bands of
blocker-prone assigned seats could cause independent dependency regions to coexist
under the completed and frozen architecture.

The experiment changed deterministic manifest topology only. Passenger movement,
entry-headway ranges, synchronous ticks, seat-event rules, clustering, lifecycle
tracking, confirmation, state-change gating and hierarchical prioritisation were
unchanged.

Overall Paired Result
---------------------

- Scenario pairs: 30.
- Standard average completion: 99.71%.
- Dual-aisle topology average completion: 99.81%.
- Improved topology pairs: 3.
- Unchanged pairs: 26.
- Worse topology pairs: 1.
- Net additional passengers seated: +3.
- Standard complete cabins: 27.
- Dual-aisle topology complete cabins: 29.
- Standard dependency regions: 37.
- Dual-aisle topology dependency regions: 7.
- Standard confirmed regions: 22.
- Dual-aisle topology confirmed regions: 6.
- Standard cluster selections: 25.
- Dual-aisle topology cluster selections: 6.
- Priority competitions in either mode: 0.

Principal Findings
------------------

1. Spatial separation alone did not create sustained multi-region competition.

The two deterministic row bands did not produce a genuine priority competition.
Although the intended zones existed in the manifest, the frozen architecture did
not observe two confirmed and simultaneously eligible regions during execution.

2. The topology frequently simplified rather than intensified congestion.

Region creation fell from 37 in the standard datasets to 7 in the topology
datasets. Confirmed regions fell from 22 to 6. This indicates that organising
blocker-prone seats into coherent aisle-specific bands can cause seat events to
resolve in orderly local sequences instead of producing independent sustained
congestion.

3. Boarding outcomes changed in both directions.

Three topology pairs improved and one became worse. Scenario 4 converted a
rear-boundary stall to complete boarding, while Scenario 26 changed from complete
boarding to a 323/342 rear-boundary stall. Scenario 29 converted a 258/274 mixed
residual stall to complete boarding. These contrasting outcomes confirm that
manifest topology changes dependency geometry rather than applying a uniformly
positive or negative pressure.

4. High blocker-prone seat share is not sufficient evidence of congestion.

Several topology datasets recorded very high blocker-prone seat shares while
producing no dependency region at all. Blocker potential becomes consequential
only when admission order, aisle occupancy and seat-event timing combine to make
temporary yield space unavailable.

5. The null priority result identifies the next missing component.

Experiment 4 created spatially separated zones but did not align their active
lifetimes. One zone could form and resolve before the second became confirmed.
The next controlled variable should therefore be temporal overlap between the
left- and right-aisle stress zones, while preserving the existing headway ranges
and frozen movement architecture.

Experiment 4 Conclusion
-----------------------
Dual-aisle dependency topology is not supported as a sufficient standalone method
for producing simultaneous confirmed dependency regions. It remains a useful
spatial component, but the evidence indicates that spatial independence must be
combined with deterministic temporal alignment before the hierarchy can be
reliably exercised.


======================================================================
DISCUSSION - WHY CONGESTION IS BEING DELIBERATELY CREATED
======================================================================

The present programme may initially appear to contradict the preceding
architecture research. The earlier work introduced row-bounded clustering and
reactive dependency-region management to alleviate bottlenecks, whereas the
current work deliberately constructs datasets more likely to produce them.

The objectives are complementary. A completed congestion-management architecture
cannot be fully evaluated unless the benchmark environment produces sufficiently
rich and independent congestion structures. The stress datasets do not create
congestion for operational benefit and do not insert congestion objects into the
simulation. They create controlled deterministic input conditions under which
congestion may arise naturally through the unchanged movement and seat-event
rules.

Once simultaneous confirmed regions can be reproduced reliably, the frozen
hierarchy can be tested on the larger question: which independent congestion
region should receive evaluation first, and whether that prioritisation releases
more of the cabin or avoids a deeper residual stall.


======================================================================
RESEARCH POSITION - FIRST-PRINCIPLES BOARDING EVIDENCE
======================================================================

This research deliberately examines aircraft boarding before introducing airline
boarding policies or probabilistic human-behaviour assumptions. It begins with
seat ownership, aisle occupancy, blocker movement and temporary yield-space
requirements, and develops congestion from those deterministic first principles.

This does not claim that policy or behavioural research is unnecessary. It
establishes a controlled mechanical foundation beneath those later layers. By
first identifying how seat-event dependencies form, evolve and interact, future
policy or behavioural studies can be interpreted against known congestion
mechanisms rather than treating boarding time as an unexplained final outcome.

The overall programme therefore contains three connected levels:

1. Volumes 1 and 2 explain how deterministic cabin congestion forms.
2. The Reactive Dependency-Region Management Architecture explains how that
   congestion is detected, retained, confirmed and prioritised.
3. The Deterministic Stress Dataset programme constructs controlled evidence
   capable of challenging the completed architecture.


======================================================================
EXPERIMENT 5 - DUAL-AISLE DEPENDENCY TIMING ALIGNMENT
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original people-movement algorithm acts upon the occupancy state present
at each deterministic execution step. Experiment 5 preserves that movement
contract while changing only when two deterministic passenger subsets are
presented to the frozen cabin architecture.

The dataset does not delay a passenger in response to observed congestion and
does not predict a future bottleneck. Queue order is constructed before the run
from the scenario seed. Runtime movement then proceeds under the unchanged
synchronous rules.

Planned Purpose
---------------
Determine whether temporal alignment of the two spatial stress zones created in
Experiment 4 can increase the overlap of their dependency-region lifecycles and
thereby produce genuine hierarchical priority competitions.

Paired Modes
------------
Each scenario is executed twice using the same cabin, occupancy, entry-headway
range and scenario seed.

1. STANDARD DETERMINISTIC DATASET

   The established seeded manifest and independently shuffled aisle queues.

2. DUAL-AISLE DEPENDENCY TIMING ALIGNMENT DATASET

   The dual-zone blocker-prone manifest is retained. Each aisle queue is then
   deterministically ordered so passengers associated with its target stress
   band enter during corresponding early waves. The numerical entry-headway range
   remains identical to the paired standard run.

Independent Variable
--------------------

    PRE-RUN TEMPORAL ALIGNMENT OF AISLE-SPECIFIC STRESS ZONES

The experiment changes deterministic queue construction, which is an authorised
dataset variable in this research programme. It does not change admission
headway, movement speed or runtime scheduling.

Controlled Conditions
---------------------

- frozen Experiment 23 congestion-management architecture;
- identical cabin configuration and occupancy;
- identical scenario seed;
- identical entry-headway range;
- one assigned seat per passenger;
- fixed serving aisle after manifest creation;
- synchronous global execution;
- maximum one forward aisle tile per passenger per tick;
- no backward movement, overtaking or aisle switching;
- unchanged blocker, yield-space and middle-bank reservation rules;
- unchanged clustering, lifecycle, confirmation, gating and priority rules;
- unchanged stall detector and maximum tick limit.

Timing-Alignment Contract
-------------------------

1. Two separated target row bands are derived deterministically from cabin size.
2. Blocker-prone seats remain concentrated around their aisle-specific bands.
3. Within each aisle queue, target-band and blocker-prone passengers receive an
   earlier deterministic queue score.
4. Left and right queues are scored independently using corresponding target
   bands so their stress waves can overlap.
5. Entry-headway values are not reduced or bypassed.
6. No passenger order changes after simulation begins.
7. No region is inserted, extended or held alive by the dataset generator.
8. A competition is counted only when the frozen runtime architecture observes
   two confirmed and eligible regions simultaneously.

New Evidence Collected
----------------------

- dependency regions created and confirmed;
- maximum simultaneous active regions;
- priority competitions and lower-priority deferrals;
- hierarchical selections and left/right wins;
- region-lifecycle overlap;
- cluster selections and cluster-started events;
- complete cabins and seated totals;
- residual stall families;
- simulation ticks as a secondary comparison.

Experimental Hypothesis
-----------------------
If Experiment 4 failed primarily because its two spatial zones activated at
different times, deterministic queue alignment should increase lifecycle overlap
and create more genuine priority competitions without changing the architecture
or entry-headway mechanism.

A null result would indicate that spatial and temporal alignment remain
insufficient and that a later dataset may need controlled admission density or a
more explicit multi-wave construction.

Interpretation Safeguards
-------------------------

- Queue alignment is fixed before execution and is not reactive passenger
  behaviour.
- Synchronous ticks remain unchanged.
- Headway remains a controlled scenario characteristic.
- The dataset does not hold passengers to preserve an active region.
- More congestion is not automatically a successful result; the primary target
  is independent simultaneous confirmed regions.
- Any deterioration in boarding completion must be reported as a dataset cost.
- Priority competition must not be inferred from two designed row bands; it must
  be measured by the frozen architecture.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Maximum simultaneous active and confirmed regions.
3. Hierarchical selections and lower-priority deferrals.
4. Region-lifecycle overlap evidence.
5. Confirmed regions and region lifetimes.
6. Cluster selections and starts.
7. Complete cabins and seated totals.
8. Residual stall classifications.
9. Simulation ticks as a secondary measure.

Diagram Plan
------------
A later diagram may show two pre-run deterministic entry queues feeding separated
left- and right-aisle target bands during corresponding time windows. The diagram
must distinguish planned dataset timing from runtime dependency detection and
must not imply asynchronous execution, passenger holding or inserted congestion.

======================================================================
EXPERIMENT 5 FINDINGS - DUAL-AISLE DEPENDENCY TIMING ALIGNMENT
======================================================================

Execution Summary
-----------------
Experiment 5 completed 30 deterministic paired scenarios, producing 60 scenario
executions. Both paired modes used the frozen Experiment 23 architecture,
identical cabin configuration, occupancy, entry-headway range and scenario seed.
Only the pre-run construction of the dual-aisle entry queues changed.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Timing-alignment average completion: 99.85%.
- Average change: +0.14 percentage points.
- Improved pairs: 3 of 30.
- Unchanged pairs: 24 of 30.
- Worse pairs: 3 of 30.
- Net additional passengers seated: +8.
- Standard complete cabins: 27.
- Timing-alignment complete cabins: 27.
- Standard dependency regions: 37.
- Timing-alignment dependency regions: 6.
- Standard confirmed regions: 22.
- Timing-alignment confirmed regions: 6.
- Standard cluster selections: 25.
- Timing-alignment cluster selections: 5.
- Standard cluster-started events: 31.
- Timing-alignment cluster-started events: 4.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Deterministic timing changed outcomes in both directions.

The timing-alignment dataset converted three standard stalls to complete cabins,
but it also introduced three new incomplete outcomes. This confirms that
temporal dataset construction is a genuine stress variable rather than a
one-directional improvement mechanism.

2. Timing could remove a previously stable congestion structure.

Scenario 4 changed from a 200/204 rear-boundary lock to complete boarding.
Scenario 29 changed from a 258/274 mixed residual dependency to complete
boarding. These cases demonstrate that deterministic queue timing can alter the
dependency topology sufficiently to prevent an otherwise stable stall.

3. Timing could also create new linked row-event chains.

Scenario 5 changed from complete boarding to 233/235 with a linked row-event
chain. Scenario 30 changed from complete boarding to 286/288 with another linked
row-event chain. The timing dataset therefore created distinct deterministic
dependencies even though movement rules and entry-headway ranges were unchanged.

4. Broad timing alignment simplified dependency-region activity overall.

Despite the mixed outcome changes, dependency regions fell from 37 to 6 and
confirmed regions fell from 22 to 6. Cluster selections and cluster-started
events also reduced sharply. The aligned queues frequently produced orderly
local seat-event sequences rather than two sustained independent congestion
zones.

5. No hierarchical competition occurred.

The central objective was not achieved. Priority competitions remained zero,
showing that spatial separation plus broad temporal alignment was still
insufficient to create two simultaneously confirmed and eligible regions.

6. The result supports a more constructive two-wave design.

The next dataset should not merely bring both aisle zones forward at similar
times. It should first seed potential seated blockers in each target zone and
then introduce deeper-seat passengers behind them, creating the prerequisites
for independent blocker-dependent congestion on both aisles.

Scientific Interpretation
-------------------------
Experiment 5 confirms that time is a separate deterministic component of
dependency topology. However, simply aligning two broad stress waves does not
guarantee simultaneous dependency regions. The internal order of passengers
within each wave is equally important because congestion requires a specific
sequence: blockers must become seated before later passengers require access to
deeper seats.

This finding preserves the first-principles nature of the programme. No runtime
behaviour was added and no passenger was held in response to congestion. The
evidence arose solely from deterministic pre-run dataset construction.

Experiment 5 Conclusion
-----------------------
Dual-aisle timing alignment is supported as an influential dataset variable but
not as a sufficient standalone method for exercising hierarchical region
priority. The next experiment should use deterministic blocker-seeding and
dependency-trigger waves on both aisles to construct a stronger opportunity for
simultaneous confirmed regions.


======================================================================
DISCUSSION - FROM TIMING ALIGNMENT TO CONSTRUCTIVE CONVERGENCE
======================================================================

The first five stress experiments demonstrate that congestion does not arise
from a single variable in isolation. Row concentration, blocker-prone seat
distribution, spatial separation and timing can each alter the outcome, but the
desired multi-region state requires these elements to occur in the correct
causal order.

For a blocker-dependent row event to form, an aisle-side or intermediate
passenger must first become seated. A later passenger assigned to a deeper seat
must then reach the same row while the required temporary yield tiles are
occupied or otherwise unavailable. To create two independent regions, this
sequence must occur in two separated aisle-specific zones during overlapping
lifecycle windows.

Experiment 6 therefore remains a dataset experiment rather than an architectural
or behavioural intervention. It constructs the causal prerequisites before the
run begins but allows the frozen movement engine to determine whether genuine
congestion actually forms.


======================================================================
EXPERIMENT 6 - CONSTRUCTIVE DEPENDENCY CONVERGENCE
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original people-movement algorithm operates on explicit occupancy states
and deterministic scan order. Experiment 6 applies the same principle to dataset
construction by arranging two pre-run occupancy sequences:

1. passengers likely to become seated blockers;
2. passengers whose deeper seats may later require those blockers to yield.

The movement engine remains unchanged. The dataset creates only the initial
ordering conditions from which the original deterministic architecture may
produce congestion naturally.

Planned Purpose
---------------
Determine whether a deterministic two-wave sequence on each aisle can produce
two independent blocker-dependent congestion regions whose confirmed lifecycles
overlap sufficiently to activate the frozen hierarchical prioritisation layer.

Paired Modes
------------
Each scenario is executed twice with the same cabin, occupancy, entry-headway
range and scenario seed.

1. STANDARD DETERMINISTIC DATASET

   The established seeded manifest with independently shuffled aisle queues.

2. CONSTRUCTIVE DEPENDENCY CONVERGENCE DATASET

   The dual-zone manifest is retained. Each aisle queue is divided
   deterministically into:

   - Wave A: target-zone passengers seated close to the serving aisle, intended
     only to establish potential blockers through normal boarding;
   - Wave B: target-zone passengers assigned to deeper seats, arriving after
     Wave A and therefore capable of generating blocker-dependent seat events;
   - remaining passengers, admitted afterwards under the same headway rules.

Left and right aisle waves are constructed independently but use corresponding
target zones, giving both dependency structures an opportunity to converge in
time.

Independent Variable
--------------------

    PRE-RUN BLOCKER-SEEDING AND DEPENDENCY-TRIGGER WAVE ORDER

This is a deterministic dataset variable. It does not modify seat-event rules,
passenger movement, runtime scheduling or architectural priority.

Controlled Conditions
---------------------

- frozen Experiment 23 congestion-management architecture;
- identical cabin configuration and occupancy;
- identical scenario seed;
- identical entry-headway range;
- one unique assigned seat per passenger;
- fixed serving aisle after manifest creation;
- synchronous global execution;
- maximum one forward aisle tile per passenger per tick;
- no backward movement, overtaking or aisle switching;
- no movement beyond the assigned row;
- unchanged blocker, yield-space and middle-bank reservation rules;
- unchanged clustering, lifecycle, confirmation, gating and hierarchy;
- unchanged stall detector and maximum tick limit;
- no behavioural assumptions or asynchronous execution.

Constructive Convergence Contract
---------------------------------

1. Two separated target row bands are derived from cabin size.
2. The manifest retains deterministic blocker-prone occupancy around both bands.
3. Wave A contains target-band passengers closest to their serving aisle.
4. Wave B contains target-band passengers assigned to deeper seats.
5. Wave A is ordered before Wave B independently on each aisle.
6. The same numerical entry-headway range remains active.
7. No passenger is delayed after execution begins to preserve the waves.
8. No seated blocker is artificially inserted.
9. No congestion region is inserted, extended or protected.
10. A priority competition is recorded only when the frozen runtime architecture
    observes multiple confirmed and eligible regions simultaneously.

New Evidence Collected
----------------------

- dependency regions created and confirmed;
- maximum simultaneous active regions;
- simultaneous confirmed-region competitions;
- hierarchical selections and lower-priority deferrals;
- left/right priority wins and priority ties;
- region lifetimes and overlap;
- blocker-dependent seat-event attempts;
- single- and multiple-blocker seat events;
- cluster selections and cluster-started events;
- complete cabins and seated totals;
- residual stall families;
- simulation ticks as a secondary measure.

Experimental Hypothesis
-----------------------
If the absence of priority competition in Experiment 5 resulted from an
incorrect within-wave causal order, seating potential blockers before deeper-seat
passengers should create more persistent independent dependencies and increase
the probability of simultaneous confirmed-region competition.

A null result would indicate that queue-wave construction alone remains
insufficient and that later work may need a more narrowly targeted benchmark
catalogue or controlled variation in admission density.

Interpretation Safeguards
-------------------------

- Wave order is fixed before execution.
- Wave A does not guarantee that a passenger becomes a blocker.
- Wave B does not guarantee that a seat event becomes blocked.
- The runtime architecture remains solely responsible for detecting regions.
- Headway remains unchanged within each paired scenario.
- No passenger is held because a region has or has not formed.
- More incomplete cabins are a dataset cost, not automatically a success.
- The primary success criterion is genuine simultaneous confirmed-region
  competition, not congestion volume alone.
- Zero competitions must be reported honestly.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Maximum simultaneous active and confirmed regions.
3. Hierarchical selections and lower-priority deferrals.
4. Confirmed-region lifecycle overlap.
5. Regions created, confirmed and dissolved.
6. Blocker-dependent and multiple-blocker seat-event evidence.
7. Cluster selections and starts.
8. Complete cabins and seated totals.
9. Residual stall classifications.
10. Simulation ticks as a secondary measure.

Diagram Plan
------------
A later diagram may show matching left- and right-aisle target zones with Wave A
aisle-side passengers entering first and Wave B deeper-seat passengers following.
The diagram must show that congestion is an observed runtime result rather than
an inserted dataset object and must not imply passenger holding, asynchronous
execution or changed headway.

======================================================================
EXPERIMENT 6 FINDINGS - CONSTRUCTIVE DEPENDENCY CONVERGENCE
======================================================================

Execution Summary
-----------------
Experiment 6 completed 30 deterministic paired scenarios, producing 60 scenario
executions. Both paired modes used the frozen Experiment 23 architecture,
identical cabin configuration, occupancy, scenario seed and entry-headway range.
Only deterministic pre-run manifest and aisle-queue construction changed.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Constructive-convergence average completion: 99.93%.
- Average change: +0.22 percentage points.
- Improved pairs: 3 of 30.
- Unchanged pairs: 26 of 30.
- Worse pairs: 1 of 30.
- Net additional passengers seated: +17.
- Standard complete cabins: 27.
- Constructive-convergence complete cabins: 29.
- Standard dependency regions: 37.
- Constructive-convergence dependency regions: 1.
- Standard confirmed regions: 22.
- Constructive-convergence confirmed regions: 1.
- Standard cluster selections: 25.
- Constructive-convergence cluster selections: 0.
- Standard cluster-started events: 31.
- Constructive-convergence cluster-started events: 0.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Constructive wave ordering changed deterministic outcomes substantially.

The dataset converted Scenarios 4, 17 and 29 from incomplete standard outcomes
to complete cabins. Scenario 29 recovered 16 passengers, changing from 258/274
to 274/274. These results confirm that pre-run causal ordering can alter the
resulting dependency topology without changing movement or architecture.

2. The same ordering also created a new dependency family.

Scenario 10 changed from complete boarding to 337/342 and terminated with a
MOVING_PASSENGER_YIELD_CHAIN. The trigger passenger, moving yield-tile occupant
and downstream waiting row event formed a persistent deterministic chain. This
shows that the dataset is a genuine stress variable rather than a general
boarding improvement.

3. The intended blocker-seeding sequence usually simplified the cabin.

Across the complete batch, dependency regions fell from 37 to 1 and confirmed
regions fell from 22 to 1. The causal ordering generally allowed aisle-side
passengers to sit and deeper-seat passengers to follow in an orderly sequence,
thereby removing many naturally occurring conflicts instead of creating two
independent congestion zones.

4. The central multi-region objective remained unmet.

No priority competition, hierarchical selection or lower-priority deferral was
recorded. Experiment 6 therefore did not produce two simultaneously confirmed
and eligible dependency regions.

5. The limiting factor is now lifecycle overlap rather than congestion creation.

Experiments 2-6 have demonstrated that deterministic datasets can create,
remove and transform congestion. The remaining challenge is narrower: two
independent regions must form and remain active during overlapping lifecycle
windows. Increasing blocker frequency alone is not sufficient.

Scientific Interpretation
-------------------------
Experiment 6 provides strong first-principles evidence that the internal causal
order of a boarding manifest affects dependency topology. Potential blockers,
rearward aisle occupancy and deeper-seat arrivals do not combine mechanically;
their precise order determines whether a dependency forms, resolves quickly or
becomes a persistent yield chain.

The experiment also guards against a misleading assumption: deliberately
seating blockers first does not necessarily intensify congestion. In many cases
it regularises seat access and reduces region formation. A useful multi-region
stress dataset must therefore create both a blocked event and a sufficiently
long-lived local yield-space conflict.

Experiment 6 Conclusion
-----------------------
Constructive dependency convergence is supported as an influential deterministic
dataset technique, but it is rejected as a sufficient method for exercising the
hierarchical prioritisation layer. The next experiment should target the natural
lifetime of two independent regions by creating local downstream yield-corridor
pressure around both aisle-specific target bands.


======================================================================
DISCUSSION - WHY REGION LIFETIME NOW BECOMES THE PRIMARY VARIABLE
======================================================================

The preceding experiments initially asked how independent congestion zones could
be created. Experiment 6 narrows the problem further. The frozen hierarchy can
only compare two regions if both have passed detection and confirmation and are
still eligible at the same tick. Two regions that occur sequentially do not
constitute a priority competition, even if both are individually significant.

The next experiment therefore does not preserve regions artificially and does
not modify lifecycle thresholds. Instead, it changes only the pre-run dataset so
that each target zone contains a natural source of downstream yield-space
pressure. Potential blockers enter first, passengers assigned just beyond the
target zone then occupy the rear corridor, and deeper-seat trigger passengers
follow. If a dependency forms, the occupied corridor may allow it to persist
long enough for the independent region on the other aisle to reach confirmation.

This remains a first-principles dataset investigation. The architecture does not
know which passengers belong to a wave, no passenger is held after execution
begins, and no region is extended by code once it forms.


======================================================================
EXPERIMENT 7 - DUAL-REGION DEPENDENCY LIFECYCLE EXTENSION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original people-movement algorithm represents flow through explicit
occupancy states and deterministic ordering. Experiment 7 uses the same idea at
the dataset level: two independent passenger sequences are constructed so that
normal forward movement may create persistent occupancy dependencies around two
separated row bands.

The passenger movement algorithm remains unchanged. The experiment controls only
which assigned passengers appear earlier in each deterministic aisle queue.

Planned Purpose
---------------
Determine whether local downstream yield-corridor pressure can extend the
natural lifetime of two independently formed dependency regions sufficiently to
produce simultaneous confirmed-region competition.

Paired Modes
------------
Each scenario is executed twice with the same cabin, occupancy, entry-headway
range and scenario seed.

1. STANDARD DETERMINISTIC DATASET

   The established seeded manifest with independently shuffled aisle queues.

2. DUAL-REGION LIFECYCLE EXTENSION DATASET

   Two separated aisle-specific target bands are retained. Each serving-aisle
   queue is constructed in four deterministic phases:

   - Phase 0: target-band passengers close to the serving aisle, capable of
     becoming seated blockers through normal movement;
   - Phase 1: passengers assigned just beyond the target band, capable of
     occupying the downstream/rear yield corridor;
   - Phase 2: deeper-seat target-band passengers, capable of triggering
     blocker-dependent row events;
   - Phase 3: adjacent-row reinforcement passengers, capable of sustaining
     local occupancy pressure while the independent aisle region forms.

Independent Variable
--------------------

    PRE-RUN DUAL-REGION YIELD-CORRIDOR PRESSURE ORDER

This is a deterministic dataset variable. It does not modify runtime movement,
seat-event rules, lifecycle thresholds or architectural priority.

Controlled Conditions
---------------------

- frozen Experiment 23 congestion-management architecture;
- identical cabin configuration and occupancy;
- identical scenario seed;
- identical numerical entry-headway range;
- one unique assigned seat per passenger;
- fixed serving aisle after manifest creation;
- synchronous global execution;
- maximum one forward aisle tile per passenger per tick;
- no backward movement, overtaking or aisle switching;
- no movement beyond the assigned row;
- unchanged blocker, yield-space and middle-bank reservation rules;
- unchanged clustering, lifecycle, confirmation, gating and hierarchy;
- unchanged stall detector and maximum tick limit;
- no human behaviour or asynchronous execution.

Lifecycle-Extension Contract
----------------------------

1. Two target bands are derived deterministically from cabin size.
2. Potential blockers are placed before the local trigger passengers.
3. Rear-corridor passengers are assigned seats beyond each target band and move
   under the ordinary forward-only rules.
4. Deeper-seat triggers then approach the target rows.
5. Adjacent reinforcement passengers follow.
6. The ordering is fixed before tick 1.
7. No passenger is held or released in response to runtime congestion.
8. No dependency region is inserted, protected or artificially prolonged.
9. Confirmation age and lifecycle rules remain unchanged.
10. A competition is recorded only by the frozen runtime hierarchy.

New Evidence Collected
----------------------

- maximum simultaneous active regions;
- regions reaching confirmation;
- priority competitions;
- regions ranked and hierarchical selections;
- lower-priority evaluations deferred;
- left/right priority wins and ties;
- largest region lifetime;
- region updates and dissolutions;
- yield-tile occupied failures;
- multiple-blocker seat events;
- moving-passenger yield chains;
- complete cabins and seated totals;
- residual stall families;
- simulation ticks as a secondary measure.

Experimental Hypothesis
-----------------------
If insufficient lifecycle overlap prevented hierarchical activation in
Experiment 6, deterministic downstream yield-corridor pressure should allow at
least some independently formed regions to persist while a second region reaches
confirmation.

A null result would indicate that the benchmark requires either more narrowly
constructed scenario-specific datasets or a controlled study of admission
density. It would not justify changing lifecycle thresholds or inserting regions.

Interpretation Safeguards
-------------------------

- More congestion is not automatically a successful result.
- A long-lived single region does not satisfy the multi-region objective.
- Two active regions do not count unless the frozen hierarchy regards them as
  confirmed and simultaneously eligible.
- Rear-boundary locks must be distinguished from genuine multi-region overlap.
- Wave labels exist only in pre-run construction and are invisible to runtime.
- Entry-headway values remain those selected for the paired scenario.
- No runtime intervention may be introduced to preserve a region.
- Zero competitions must be reported honestly.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Hierarchical selections and lower-priority deferrals.
3. Maximum simultaneous active and confirmed regions.
4. Largest region lifetime and lifecycle overlap evidence.
5. Regions created, confirmed and dissolved.
6. Yield-corridor and blocker-dependent evidence.
7. Cluster selections and starts.
8. Complete cabins and seated totals.
9. Residual stall classifications.
10. Simulation ticks as a secondary measure.

Diagram Plan
------------
A later diagram may show two separated target bands, each with a seated-blocker
wave, a downstream corridor-pressure wave and a deeper-seat trigger wave. It
must make clear that all ordering is defined before execution and that the
runtime architecture alone determines whether regions form and overlap.

======================================================================
EXPERIMENT 7 FINDINGS - DUAL-REGION DEPENDENCY LIFECYCLE EXTENSION
======================================================================

Execution Summary
-----------------
Experiment 7 completed 30 deterministic paired scenarios, producing 60 scenario
executions. Both modes used the frozen Experiment 23 architecture, the same
scenario seed, cabin, occupancy and entry-headway range. Only deterministic
pre-run queue construction changed.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Lifecycle-extension average completion: 99.89%.
- Improved pairs: 3 of 30.
- Unchanged pairs: 25 of 30.
- Worse pairs: 2 of 30.
- Net additional passengers seated: +11.
- Standard complete cabins: 27.
- Lifecycle-extension complete cabins: 28.
- Standard dependency regions: 37.
- Lifecycle-extension dependency regions: 5.
- Standard confirmed regions: 22.
- Lifecycle-extension confirmed regions: 5.
- Standard cluster selections: 25.
- Lifecycle-extension cluster selections: 5.
- Standard cluster-started events: 31.
- Lifecycle-extension cluster-started events: 5.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. The lifecycle-extension ordering did not extend multi-region lifetimes.

Although the dataset was designed to preserve independent congestion, it reduced
dependency-region creation from 37 to 5 and confirmed regions from 22 to 5.
The structured phases therefore simplified most dependency evolution rather than
keeping two regions alive simultaneously.

2. Additional deterministic order frequently regularised boarding.

The dataset again converted several difficult standard outcomes into complete
boarding, including the established Scenario 4 rear-boundary lock. This confirms
that orderly causal sequencing can remove congestion even when the intention is
to preserve it.

3. Some adverse outcomes remained possible.

Two paired scenarios became worse, showing that the dataset was not simply an
optimisation policy. It genuinely altered dependency topology, but those adverse
cases still formed isolated rather than competing confirmed regions.

4. No hierarchical competition occurred.

Priority competitions and lower-priority deferrals remained zero. The completed
hierarchy therefore still had no occasion to choose between two simultaneously
confirmed eligible regions.

5. Lifecycle duration cannot be increased merely by adding broader reinforcement.

The additional corridor and adjacent-row phases often allowed dependencies to
resolve cleanly before a second independent region matured. More deterministic
structure therefore reduced the phenomenon under investigation.

Scientific Interpretation
-------------------------
Experiment 7 demonstrates that making a dependency sequence longer or more
ordered is not equivalent to extending the lifetime of a detected dependency
region. The runtime region exists only while the actual unresolved occupancy and
seat-event prerequisites remain present. Pre-run ordering that improves causal
flow may therefore shorten or prevent the region even when more passengers are
assigned around the target zone.

The evidence shifts the research question from increasing dependency volume to
preserving causal independence. Two regions must be allowed to form without
shared middle-bank activity, broad queue ordering or downstream passengers
inadvertently resolving one another.

Experiment 7 Conclusion
-----------------------
Dual-region lifecycle extension is not supported as a reliable method for
producing simultaneous confirmed-region competition. The next experiment should
isolate the causal domains of the two candidate regions, using aisle-specific
outer-bank bands and postponing shared middle-bank activity until after the
independent structures have had an opportunity to form.


======================================================================
DISCUSSION - ORDER, LIFETIME AND CAUSAL INDEPENDENCE
======================================================================

Experiments 5-7 show that deterministic order can both create and remove
congestion. Timing alignment, blocker-first construction and lifecycle
reinforcement all changed outcomes, but increasingly ordered datasets often
became easier for the cabin to process.

This is scientifically important. A dependency region is not a permanent object
inserted by the dataset. It is a temporary runtime consequence of unresolved
seat-event prerequisites and occupied aisle space. If the dataset organises
passengers too efficiently, the architecture correctly observes no region.

The next step is therefore not to add more passengers or more phases. It is to
separate the two causal systems. Each aisle should receive its own outer-bank
target band, while shared middle-bank passengers are delayed in the pre-run
manifest. Any resulting regions must still arise naturally through the frozen
movement and seat-event rules.


======================================================================
EXPERIMENT 8 - INDEPENDENT REGION ISOLATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original people-movement algorithm reasons from local occupancy and
explicit movement constraints. Experiment 8 applies the same principle by
constructing two independent local input domains before execution. The left and
right aisle target bands are separated, and their early passengers are drawn
only from the corresponding outer seat banks.

The movement engine does not know that the bands are experimental objects. It
continues to process ordinary passenger IDs, occupied aisle tiles and seat-event
requirements exactly as before.

Planned Purpose
---------------
Determine whether isolating two aisle-specific outer-bank dependency ecosystems
allows both regions to form and remain confirmed simultaneously without merging,
resolving one another or relying upon shared middle-bank interactions.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

   The established seeded manifest and independently shuffled aisle queues.

2. INDEPENDENT REGION ISOLATION DATASET

   Two separated target bands are retained. For each aisle independently:

   - potential blockers in that aisle's outer bank enter first;
   - deeper-seat passengers in the same outer-bank band follow;
   - adjacent outer-bank reinforcement follows;
   - remaining passengers from that outer bank are admitted;
   - middle-bank and other shared passengers enter later.

Independent Variable
--------------------

    PRE-RUN ISOLATION OF TWO AISLE-SPECIFIC OUTER-BANK CAUSAL DOMAINS

Controlled Conditions
---------------------

- frozen Experiment 23 congestion-management architecture;
- identical cabin configuration, occupancy and scenario seed;
- identical entry-headway range;
- synchronous deterministic execution;
- one unique assigned seat per passenger;
- no runtime holding, inserted congestion or artificial region persistence;
- no backward movement, overtaking or aisle switching;
- no passenger moving beyond the assigned row;
- unchanged blocker, yield-space and middle-bank rules;
- unchanged clustering, lifecycle confirmation, state-change gating and
  hierarchical prioritisation;
- unchanged stall detector and maximum tick limit;
- no human-behaviour assumptions.

Isolation Contract
------------------

1. Left and right target bands remain spatially separated.
2. Early left-band passengers use only the left outer bank.
3. Early right-band passengers use only the right outer bank.
4. Middle-bank passengers are excluded from the early isolation phases.
5. Potential blockers enter before deeper-seat triggers in each isolated band.
6. No passenger is held after execution begins.
7. Numerical entry-headway ranges remain unchanged.
8. The runtime architecture alone decides whether a dependency region exists.
9. A competition is counted only when multiple confirmed eligible regions
   coexist under the frozen hierarchy.

Primary Evidence
----------------

- dependency regions created and confirmed;
- maximum simultaneous active regions;
- priority competitions;
- hierarchical selections and lower-priority deferrals;
- left/right priority wins and ties;
- region lifetimes and simultaneous overlap;
- single- and multiple-blocker seat events;
- middle-bank reservation activity;
- cluster selections and cluster-started events;
- complete cabins and seated totals;
- residual stall families;
- simulation ticks as a secondary measure.

Experimental Hypothesis
-----------------------
If shared middle-bank activity and broad queue reinforcement were prematurely
resolving or coupling the two candidate regions, isolating the early outer-bank
causal domains should increase the probability that left and right confirmed
regions coexist.

A null result would indicate that the present benchmark generator cannot
reliably create simultaneous regions through general distribution rules alone.
A later phase may then require a small catalogue of explicitly selected,
deterministic stress manifests rather than further broad generator changes.

Interpretation Safeguards
-------------------------

- Isolation is a pre-run dataset property, not a runtime policy.
- Delaying middle-bank passengers does not change their assigned seats.
- No region is inserted or protected.
- More incomplete cabins are not automatically a successful result.
- The primary criterion remains genuine simultaneous confirmed-region
  competition.
- Zero competitions must be documented without reinterpretation.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Maximum simultaneous active and confirmed regions.
3. Hierarchical selections and lower-priority deferrals.
4. Region-lifecycle overlap.
5. Regions created, confirmed and dissolved.
6. Blocker-dependent seat-event evidence.
7. Cluster selections and starts.
8. Complete cabins and seated totals.
9. Residual stall classifications.
10. Simulation ticks.

Diagram Plan
------------
A later diagram may show two separated outer-bank row bands, one attached to
each aisle, with shared middle-bank passengers placed in a later admission
layer. The diagram must not imply physical barriers, runtime holding or modified
movement rules.

======================================================================
EXPERIMENT 8 FINDINGS - INDEPENDENT REGION ISOLATION
======================================================================

Execution Summary
-----------------
Experiment 8 completed 30 deterministic paired scenarios, producing 60 scenario
executions. The frozen Experiment 23 architecture, scenario seed, cabin,
occupancy and entry-headway range remained identical in each pair. Only the
pre-run isolation of the two aisle-specific causal domains changed.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Independent-isolation average completion: 100.00%.
- Improved pairs: 3 of 30.
- Unchanged pairs: 27 of 30.
- Worse pairs: 0 of 30.
- Net additional passengers seated: +22.
- Standard complete cabins: 27.
- Isolation complete cabins: 30.
- Standard dependency regions: 37.
- Isolation dependency regions: 4.
- Standard confirmed regions: 22.
- Isolation confirmed regions: 4.
- Standard cluster selections: 25.
- Isolation cluster selections: 5.
- Standard cluster-started events: 31.
- Isolation cluster-started events: 5.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Independent isolation produced complete boarding in all 30 stress executions.

The dataset converted all three incomplete standard outcomes to complete cabins,
including the recurring Scenario 4 rear-boundary lock. It introduced no new
incomplete outcome.

2. Isolation strongly suppressed dependency formation.

Dependency regions fell from 37 to 4 and confirmed regions fell from 22 to 4.
The early outer-bank separation therefore prevented most blocker/yield
interactions rather than preserving two independent congestion ecosystems.

3. Multiple-blocker activity was reduced in many scenarios.

The detailed reports show that the isolation ordering frequently allowed
aisle-side and intermediate passengers to seat before deeper-seat arrivals
created a sustained queue. The resulting seat events were simpler and failed
less often.

4. No hierarchical competition occurred.

Priority competitions and lower-priority deferrals remained zero. The hierarchy
again had no occasion to rank two simultaneously confirmed eligible regions.

5. Excessive causal separation is effectively an optimisation.

Although designed as a stress dataset, the isolation manifest behaved like a
highly orderly boarding policy. It removed the interactions required for
dependency regions to form.

Scientific Interpretation
-------------------------
Experiment 8 demonstrates that independent regions cannot be obtained merely by
separating aisle-specific passenger groups. Dependency regions require enough
interaction to create blocked seat-event prerequisites, but not so much coupling
that the regions merge or resolve one another.

The broad generator has therefore reached a practical limit. Repeated global
ordering rules either simplify boarding or create isolated single-region
outcomes. The next stage should move from one universal stress generator to a
small deterministic catalogue of explicitly defined benchmark topologies.

Experiment 8 Conclusion
-----------------------
Independent Region Isolation is rejected as a method for producing simultaneous
confirmed-region competition. It is retained as evidence that deterministic
manifest design can eliminate known stalls and that excessive isolation removes
the phenomenon under study.


======================================================================
DISCUSSION - WHY A BENCHMARK CATALOGUE IS NOW REQUIRED
======================================================================

Experiments 1-8 systematically tested broad deterministic variables: row
concentration, blocker-prone seats, spatial separation, timing, constructive
waves, lifecycle reinforcement and causal isolation.

The evidence now shows that no single general ordering rule reliably produces
the rare multi-region state. Broad rules tend either to regularise boarding or
to generate one dominant local dependency.

A benchmark catalogue is therefore scientifically preferable. Each catalogue
family can represent a named dependency topology while the runtime architecture
remains frozen. This follows normal algorithm-evaluation practice: rare
architectural capabilities are tested with reproducible constructed instances
rather than waiting for random workloads to produce them.


======================================================================
EXPERIMENT 9 - DETERMINISTIC DEPENDENCY BENCHMARK CATALOGUE
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original algorithm separates deterministic input construction from
occupancy-based execution. Experiment 9 preserves that separation. The benchmark
catalogue defines only the starting passenger order; Amit's movement logic and
the frozen dependency-region architecture decide every runtime state.

Planned Purpose
---------------
Determine whether a small reproducible catalogue of distinct deterministic
dependency topologies can exercise a wider range of architectural behaviour than
one universal stress generator, including genuine simultaneous confirmed-region
competition.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

2. DETERMINISTIC DEPENDENCY BENCHMARK CATALOGUE

   The 30 scenarios cycle through five reproducible benchmark families:

   A. Isolated outer-bank blocker/trigger topology.
   B. Tightly localised simultaneous-confirmation candidate.
   C. Separated long-chain topology with adjacent reinforcement.
   D. Controlled interaction topology with delayed middle-bank participation.
   E. Mixed topology retaining limited seeded variation.

Independent Variable
--------------------

    NAMED DETERMINISTIC BENCHMARK TOPOLOGY FAMILY

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy, seed and entry-headway range;
- synchronous deterministic execution;
- unchanged passenger movement and seat-event rules;
- no runtime holding or inserted dependency region;
- no backward movement, overtaking or aisle switching;
- no passenger movement beyond the assigned row;
- unchanged lifecycle confirmation and hierarchy thresholds;
- no behavioural or asynchronous variables.

Benchmark Catalogue Contract
----------------------------

1. Each scenario is assigned one of five families deterministically.
2. Left and right target rows remain reproducible.
3. Potential blockers precede deeper-seat triggers within each family.
4. Middle-bank participation differs by family but is always pre-run.
5. No family changes runtime movement.
6. No family guarantees that a region will form.
7. A successful hierarchy test requires genuine simultaneous confirmed eligible
   regions observed by the frozen architecture.

Primary Evidence
----------------

- priority competitions and lower-priority deferrals;
- maximum simultaneous active regions;
- regions created and confirmed;
- hierarchical selections;
- left/right wins and ties;
- region lifetimes;
- cluster selections and starts;
- residual stall families;
- complete cabins and seated totals;
- family-specific outcome patterns.

Experimental Hypothesis
-----------------------
A catalogue containing several narrowly constructed topologies should provide
better architectural coverage than a single broad generator. At least one
family may produce genuine simultaneous confirmed-region competition even if
other families simplify boarding or remain null.

A complete null result would establish that the present movement model and
region detector require an explicitly hand-authored minimal benchmark instance
rather than family-level manifest rules.

Interpretation Safeguards
-------------------------

- Catalogue families are test instances, not proposed airline policies.
- Better completion is secondary to architectural coverage.
- More congestion alone is not success.
- No runtime state is inserted or protected.
- Family-level null results remain valid evidence.
- Priority competition must be reported exactly as recorded.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Family producing the competition.
3. Maximum simultaneous active and confirmed regions.
4. Hierarchical selections and deferrals.
5. Region lifetimes and overlap.
6. Dependency and cluster evidence.
7. Residual stall topology.
8. Completion and seated totals.
9. Tick counts.

Diagram Plan
------------
A later diagram may present the five benchmark families as separate cabin
miniatures with their left/right target bands and blocker/trigger ordering.
The figure must make clear that these are deterministic test manifests rather
than runtime barriers or airline boarding policies.

======================================================================
EXPERIMENT 9 FINDINGS - DETERMINISTIC DEPENDENCY BENCHMARK CATALOGUE
======================================================================

Execution Summary
-----------------
Experiment 9 completed 30 deterministic paired scenarios, producing 60 scenario
executions. The full experiment was executed twice. The repeated run reproduced
the same principal outcomes, confirming deterministic replay.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Catalogue average completion: 99.93%.
- Improved pairs: 3 of 30.
- Unchanged pairs: 26 of 30.
- Worse pairs: 1 of 30.
- Net additional passengers seated: +17.
- Standard complete cabins: 27.
- Catalogue complete cabins: 29.
- Standard dependency regions: 37.
- Catalogue dependency regions: 3.
- Standard confirmed regions: 22.
- Catalogue confirmed regions: 3.
- Standard cluster selections: 25.
- Catalogue cluster selections: 2.
- Standard cluster-started events: 31.
- Catalogue cluster-started events: 2.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. The catalogue produced reproducible named benchmark behaviour.

The second complete execution reproduced the same principal scenario outcomes,
including the Scenario 4 conversion from rear-boundary lock to complete boarding
and the absence of hierarchical competition.

2. The catalogue successfully produced confirmed dependency regions.

Three catalogue regions reached lifecycle confirmation. This establishes that
narrowly defined benchmark families can produce architecture-visible dependency
objects rather than merely changing total boarding time.

3. The catalogue remained an effective stall-removal mechanism.

Scenario 4 again changed from 200/204 with a persistent rear-boundary lock to
204/204 complete boarding. The catalogue therefore altered the causal topology
enough to remove a known deterministic failure.

4. One adverse catalogue outcome remained.

One paired scenario became incomplete, demonstrating that the catalogue was not
simply an optimisation policy. It could expose a different dependency family
while preserving deterministic execution.

5. No simultaneous confirmed-region competition occurred.

Despite three confirmed regions across the run, they appeared as isolated
events. Priority competitions and lower-priority deferrals remained zero.

Scientific Interpretation
-------------------------
Experiment 9 is supported as a deterministic benchmark generator but remains
incomplete as a hierarchy stress generator. Its strongest contribution is the
demonstration that named topology families can reproducibly create confirmed
regions and convert a genuine stall to success.

However, family-level ordering still specifies only statistical conditions from
which a dependency may emerge. It does not define the intended dependency graph
with sufficient precision to guarantee two concurrent confirmed structures.

Experiment 9 Conclusion
-----------------------
The Deterministic Dependency Benchmark Catalogue is retained as a valid research
instrument. The next experiment should move from broad topology families to an
explicit two-chain dependency specification in which the intended blocker roots,
deep-seat triggers and downstream yield-space occupants are defined before
execution.


======================================================================
DISCUSSION - FROM BENCHMARK FAMILIES TO EXPLICIT TOPOLOGY
======================================================================

Experiments 1-9 show that simultaneous dependency regions are not produced
reliably by increasing congestion probability alone. Row density, seat depth,
timing, isolation and benchmark families can all change outcomes, but the
required graph remains rare.

Experiment 10 therefore changes the level of specification. Instead of defining
a family likely to produce dependencies, it defines two intended local chains:

    potential blocker root
            |
    second blocker
            |
    deep-seat trigger
            |
    downstream yield-space pressure

One chain is constructed for each aisle at separated target rows. The graph is
not inserted into the runtime architecture. It is translated into passenger
assignments and pre-run queue order, after which the frozen movement engine must
either realise or reject the intended topology naturally.


======================================================================
EXPERIMENT 10 - EXPLICIT DETERMINISTIC DEPENDENCY TOPOLOGY GENERATOR
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original algorithm begins with explicit data and then applies local,
deterministic occupancy rules. Experiment 10 follows the same principle. The
input generator names the intended dependency components, but the original
movement and seat-event logic remains solely responsible for creating the actual
runtime state.

Planned Purpose
---------------
Determine whether two explicitly specified and spatially separated dependency
chains can become confirmed simultaneously and activate the frozen hierarchical
prioritisation layer.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

2. EXPLICIT DETERMINISTIC DEPENDENCY TOPOLOGY DATASET

   Each aisle receives one target row containing:

   - an aisle-side potential blocker;
   - a second potential blocker;
   - a deeper-seat trigger passenger;
   - downstream outer-bank passengers capable of occupying required yield space;
   - one adjacent-row deep-seat reinforcement;
   - limited delayed middle-bank participation.

Independent Variable
--------------------

    EXPLICIT PRE-RUN TWO-CHAIN DEPENDENCY TOPOLOGY SPECIFICATION

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy, seed and entry-headway range;
- synchronous deterministic execution;
- unchanged movement and seat-event rules;
- no runtime passenger holding;
- no inserted dependency region;
- no artificial confirmation or lifecycle extension;
- no backward movement, overtaking or aisle switching;
- no movement beyond the assigned row;
- no human-behaviour or asynchronous variables.

Explicit Topology Contract
--------------------------

1. One target chain is defined for each aisle.
2. The target rows are spatially separated.
3. Potential blockers enter before the corresponding deep-seat trigger.
4. Downstream corridor occupants follow early enough to create natural
   yield-space pressure.
5. Adjacent-row reinforcement may extend each local chain.
6. Middle-bank passengers enter only after the outer-chain components.
7. No runtime state is forced.
8. The architecture must detect, confirm and rank any resulting regions itself.
9. Success requires genuine simultaneous confirmed eligible regions.

Primary Evidence
----------------

- priority competitions;
- lower-priority deferrals;
- maximum simultaneous active regions;
- regions confirmed on left and right aisles;
- hierarchical selections and priority wins;
- region overlap duration;
- intended-chain realisation count;
- cluster selections and starts;
- residual stall family;
- complete cabins and seated totals.

Experimental Hypothesis
-----------------------
Explicitly naming both blocker roots, deep-seat triggers and downstream
yield-space pressure should provide more reliable causal coverage than broad
benchmark families. At least one paired scenario may therefore produce two
simultaneously confirmed eligible dependency regions.

A null result would demonstrate that manifest and queue order alone cannot
realise the required graph under the present movement model. That would justify
closing the broad dataset-generator line and documenting the hierarchy as
architecturally valid but not naturally exercised by the available first-
principles boarding model.

Interpretation Safeguards
-------------------------

- The topology is an input specification, not a runtime intervention.
- Intended chains may fail to materialise.
- A single confirmed region is not hierarchy validation.
- More stalls are not automatically successful evidence.
- Repeated deterministic replay must remain exact.
- Zero competitions must be reported without reinterpretation.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Simultaneous confirmed left/right regions.
3. Hierarchical selections and deferrals.
4. Region overlap duration.
5. Intended-chain realisation.
6. Confirmed and dissolved regions.
7. Cluster evidence.
8. Stall topology.
9. Completion and tick counts.

Diagram Plan
------------
A later diagram may show two separated chains, one attached to each aisle, with
blocker roots, deep-seat triggers and downstream yield-pressure passengers.
The diagram must state that the graph is an intended deterministic input
topology and is not inserted into runtime state.

======================================================================
EXPERIMENT 10 FINDINGS - EXPLICIT DETERMINISTIC DEPENDENCY TOPOLOGY
======================================================================

Execution Summary
-----------------
Experiment 10 completed 30 deterministic paired scenarios, producing 60 scenario
executions. Both modes retained the frozen Experiment 23 architecture and the
same cabin, occupancy, scenario seed and entry-headway range.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Explicit-topology average completion: 99.93%.
- Improved pairs: 3 of 30.
- Unchanged pairs: 26 of 30.
- Worse pairs: 1 of 30.
- Net additional passengers seated: +17.
- Standard complete cabins: 27.
- Explicit-topology complete cabins: 29.
- Standard dependency regions: 37.
- Explicit-topology dependency regions: 1.
- Standard confirmed regions: 22.
- Explicit-topology confirmed regions: 1.
- Standard cluster selections: 25.
- Explicit-topology cluster selections: 0.
- Standard cluster-started events: 31.
- Explicit-topology cluster-started events: 0.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Explicit topology did not realise the intended two-chain graph.

The generator named potential blockers, deep-seat triggers and downstream
pressure, but the runtime architecture detected only one region across the
entire topology mode. No pair of confirmed eligible regions coexisted.

2. Strong pre-run specification again regularised boarding.

The explicit ordering converted three standard incomplete outcomes to improved
results, including the recurring Scenario 4 rear-boundary lock. Twenty-nine of
thirty topology cabins completed successfully.

3. Dependency formation was almost completely suppressed.

Regions fell from 37 to 1 and confirmed regions from 22 to 1. No cluster
selection or cluster-started event occurred in topology mode. The prescribed
sequence therefore removed most of the interactions required for dependencies
to emerge.

4. The hierarchy remained unexercised.

Priority competitions and lower-priority deferrals remained zero. This is not a
failure of the hierarchy; the required simultaneous confirmed inputs were never
presented to it.

5. Dependency regions are supported as emergent deterministic objects.

The evidence from Experiments 8-10 shows a consistent pattern: greater global
control over passenger order creates cleaner causal flow and fewer dependency
regions. A dependency graph cannot simply be written into the manifest and
assumed to appear at runtime.

Scientific Interpretation
-------------------------
Experiment 10 establishes a practical upper boundary for global pre-run
specification. Even an explicit two-chain design remains only an intended input
topology. Under the unchanged movement engine, passengers may seat cleanly
before the proposed blockers and triggers become mutually dependent.

The next experiment should therefore retain the natural standard manifest and
most of its seeded queue order, changing only a few local relationships. This
preserves emergence while allowing controlled tests of sensitivity around rows
where blocker-dependent events are plausible.

Experiment 10 Conclusion
------------------------
The Explicit Deterministic Dependency Topology Generator is rejected as a
reliable hierarchy stress generator. It is retained as evidence that excessive
pre-run control suppresses dependency formation and can remove known stalls.


======================================================================
DISCUSSION - RETURNING FROM PRESCRIPTION TO PERTURBATION
======================================================================

Experiments 8-10 progressively increased deterministic control: causal
isolation, benchmark families and explicit dependency chains. Dependency-region
coverage decreased as control increased.

Experiment 11 therefore does not prescribe a graph. It begins with the standard
seeded manifest and shuffled aisle queues, then applies only a few local,
reproducible transpositions around two target row windows. This is analogous to
a sensitivity test: the cabin remains natural except for a narrowly bounded
change whose consequences can be attributed clearly.


======================================================================
EXPERIMENT 11 - CONTROLLED DETERMINISTIC DEPENDENCY PERTURBATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original algorithm evaluates local occupancy changes without globally
rewriting the complete state. Experiment 11 adopts the same principle at the
dataset layer. The standard deterministic queue is preserved and only a small
number of local passenger relationships are perturbed.

Planned Purpose
---------------
Determine whether minimal deterministic changes around two separated row
windows preserve natural dependency formation while increasing the probability
that two confirmed regions overlap.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

2. CONTROLLED DETERMINISTIC DEPENDENCY PERTURBATION DATASET

   The same standard passenger manifest and serving-aisle assignments are used.
   Each aisle queue begins with seeded random shuffling. At most four local
   blocker/deep-seat relationships are then adjusted around one target row.

Independent Variable
--------------------

    SMALL PRE-RUN LOCAL QUEUE PERTURBATIONS

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical passenger manifest, cabin, occupancy and scenario seed;
- identical serving-aisle assignment;
- identical entry-headway range;
- synchronous deterministic execution;
- unchanged movement and seat-event rules;
- no runtime holding or inserted region;
- no backward movement, overtaking or aisle switching;
- no movement beyond the assigned row;
- no behavioural or asynchronous variables.

Perturbation Contract
---------------------

1. The standard manifest is retained.
2. Each aisle begins with normal seeded shuffling.
3. Changes are limited to two narrow row windows.
4. Only adjacent local swaps and one displacement of at most two positions are
   permitted per aisle.
5. No seat, row or serving aisle is changed.
6. Most queue positions remain untouched.
7. No perturbation occurs after execution begins.
8. The frozen architecture alone determines whether regions form.

Primary Evidence
----------------

- priority competitions and deferrals;
- maximum simultaneous active regions;
- confirmed left/right region overlap;
- regions created and confirmed;
- cluster selections and starts;
- sensitivity of known standard region scenarios;
- residual stall families;
- complete cabins and seated totals;
- tick counts as secondary evidence.

Experimental Hypothesis
-----------------------
Minimal local perturbations should preserve more natural interaction than the
explicit topology generator. They may therefore retain ordinary dependency
formation while shifting the timing of two local structures enough to create
simultaneous confirmed-region competition.

A null result would support closure of the general dataset-construction line:
it would show that neither broad ordering nor narrowly bounded queue
perturbation can reliably exercise the hierarchy under the present model.

Interpretation Safeguards
-------------------------

- Perturbations are input sensitivity tests, not airline policies.
- No dependency graph is prescribed.
- A single confirmed region is not hierarchy validation.
- Improved completion is secondary to architectural coverage.
- More congestion is not automatically success.
- Zero competitions must be reported exactly.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Simultaneous confirmed left/right regions.
3. Hierarchical selections and deferrals.
4. Region-overlap duration.
5. Regions created and confirmed.
6. Cluster evidence.
7. Stall topology.
8. Completion and tick counts.

Diagram Plan
------------
A later diagram may show a mostly unchanged shuffled queue with only two or
three highlighted local transpositions around each target row. It must not
suggest runtime passenger holding or a prescribed dependency graph.

======================================================================
EXPERIMENT 11 FINDINGS - CONTROLLED DETERMINISTIC DEPENDENCY PERTURBATION
======================================================================

Execution Summary
-----------------
Experiment 11 completed 30 deterministic paired scenarios, producing 60 scenario
executions. The frozen Experiment 23 architecture, cabin, occupancy, scenario
seed and entry-headway range remained unchanged. Only small bounded pre-run
queue perturbations were introduced.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Perturbation average completion: 99.71%.
- Improved pairs: 0 of 30.
- Unchanged pairs: 30 of 30.
- Worse pairs: 0 of 30.
- Net passenger difference: 0.
- Standard complete cabins: 27.
- Perturbation complete cabins: 27.
- Standard dependency regions: 37.
- Perturbation dependency regions: 36.
- Standard confirmed regions: 22.
- Perturbation confirmed regions: 21.
- Standard cluster selections: 25.
- Perturbation cluster selections: 24.
- Standard cluster-started events: 31.
- Perturbation cluster-started events: 29.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Minimal perturbation preserved the natural dependency landscape.

Unlike Experiments 8-10, the perturbation dataset did not suppress region
formation globally. Aggregate dependency and confirmation counts remained close
to the standard baseline.

2. Small changes could reveal new confirmed structure locally.

Scenario 3 changed from no detected region under the standard dataset to one
confirmed region, one cluster selection and one cluster-started event under the
perturbation dataset.

3. Scenario 4 produced richer evidence without changing the stall outcome.

The standard run produced one confirmed rear-boundary region. The perturbation
run produced two regions and two confirmations, with one cluster selection and
one cluster start, while retaining the same 200/204 rear-boundary lock.

4. Two confirmed regions did not imply a priority competition.

Scenario 4 demonstrated that multiple regions may be created and confirmed at
different times. Priority competition remained zero because the confirmed,
eligible lifecycles did not overlap sufficiently for hierarchical arbitration.

5. Minimal perturbation behaved as a probe rather than an optimisation.

Completion, seated totals and all paired outcomes were unchanged. The
perturbations exposed different internal region evidence without broadly making
boarding easier or harder.

Scientific Interpretation
-------------------------
Experiment 11 marks an important transition. Dependency regions were no longer
being manufactured through broad manifest ordering. Instead, small deterministic
changes revealed sensitivity in naturally emerging dependency structures.

The remaining limitation is temporal overlap. The architecture can observe more
than one confirmed region within a scenario, but not at the same eligible moment.

Experiment 11 Conclusion
------------------------
Controlled deterministic perturbation is supported as a valid scientific probe.
Experiment 12 should preserve this minimal-intervention philosophy while aligning
the queue depth of the two local perturbation windows so their region lifecycles
have a greater opportunity to overlap.


======================================================================
EXPERIMENT 12 - DUAL-WINDOW PERTURBATION OVERLAP ALIGNMENT
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's algorithm retains deterministic local occupancy logic while allowing the
input ordering to be altered independently. Experiment 12 preserves the standard
seeded queues and applies only bounded local swaps plus a small pre-run
cross-aisle queue-depth alignment.

Planned Purpose
---------------
Determine whether two minimally perturbed local windows can mature during
overlapping lifecycle periods and produce genuine hierarchical competition.

Independent Variable
--------------------

    BOUNDED CROSS-AISLE ALIGNMENT OF TWO LOCAL PERTURBATION WINDOWS

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy, scenario seed and headway range;
- synchronous deterministic execution;
- unchanged movement, seat-event, lifecycle and hierarchy rules;
- no runtime holding;
- no inserted or protected region;
- no backward movement, overtaking or aisle switching;
- no movement beyond the assigned row;
- no behavioural or asynchronous variables.

Experiment Contract
-------------------

1. Both aisle queues begin with the standard seeded shuffle.
2. At most six adjacent local swaps are permitted per aisle.
3. One deep trigger may move by no more than four local queue positions.
4. Cross-aisle trigger alignment may move one trigger by no more than three
   positions.
5. No seat, row or aisle assignment changes.
6. No runtime admission delay is introduced.
7. Success requires genuine simultaneous confirmed eligible regions.

Primary Evaluation Order
------------------------

1. Priority competitions.
2. Simultaneous confirmed-region overlap.
3. Hierarchical selections and deferrals.
4. Left/right priority wins and ties.
5. Confirmed regions and region lifetimes.
6. Cluster selections and starts.
7. Completion and residual stall families.
8. Tick counts.

Experimental Hypothesis
-----------------------
If Experiment 11 failed only because its confirmed regions matured at different
times, weak cross-aisle queue-depth alignment should increase lifecycle overlap
without suppressing natural dependency formation.

A null result would indicate that local queue depth alone is insufficient and
that explicit region-overlap instrumentation may be required before designing
further datasets.

======================================================================
EXPERIMENT 12 FINDINGS - DUAL-WINDOW PERTURBATION OVERLAP ALIGNMENT
======================================================================

Execution Summary
-----------------
Experiment 12 completed 30 deterministic paired scenarios, producing 60
executions. Both modes retained the frozen Experiment 23 architecture, identical
scenario seeds, cabin configuration, occupancy and entry-headway ranges. Only
bounded pre-run queue perturbations changed.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Overlap-alignment average completion: 99.71%.
- Improved pairs: 0 of 30.
- Unchanged pairs: 30 of 30.
- Worse pairs: 0 of 30.
- Net passenger difference: 0.
- Complete cabins: 27 in both modes.
- Standard dependency regions: 37.
- Overlap-alignment dependency regions: 35.
- Standard confirmed regions: 22.
- Overlap-alignment confirmed regions: 20.
- Standard cluster selections: 25.
- Overlap-alignment cluster selections: 23.
- Standard cluster-started events: 31.
- Overlap-alignment cluster-started events: 28.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Dual-window alignment preserved overall boarding outcomes.

Every paired scenario seated the same number of passengers in both modes. The
perturbation therefore changed local dependency structure without changing the
overall completion family.

2. Scenario 3 again exposed hidden dependency structure.

The standard execution produced no dependency region. The overlap-alignment
dataset produced two regions, one confirmed region, one cluster selection and
one cluster-started event. This reproduces the Experiment 11 finding that small
deterministic changes can reveal architecture-visible dependencies.

3. The rear-boundary lock remained unchanged.

Scenario 4 retained the same 200/204 rear-boundary stall, dependency-chain length
of four and 1500 ticks without seating progress. The overlap alignment neither
resolved nor deepened the final stable lock.

4. Some naturally occurring evidence was suppressed.

Scenario 5 decreased from two standard regions and one confirmed region to one
unconfirmed region. Across the complete benchmark, regions, confirmations and
cluster activity all decreased slightly.

5. No lifecycle competition occurred.

The two local windows did not mature as simultaneously confirmed eligible
regions. Priority competitions and lower-priority deferrals remained zero.

Scientific Interpretation
-------------------------
Experiment 12 confirms that bounded temporal alignment is insufficient to turn
separate local dependency events into hierarchy competition. Minimal
perturbations can reveal hidden regions, but attempting to align their queue
depths may also suppress naturally occurring evidence.

The result suggests that the missing variable is not simply the distance between
two deep triggers in their entry queues. The two candidate structures must share
a reproducible causal motif while remaining physically independent.

Experiment 12 Conclusion
------------------------
Dual-Window Perturbation Overlap Alignment is retained as evidence that local
dependency structures are sensitive to small deterministic ordering changes.
However, it is rejected as a reliable method for producing simultaneous
confirmed-region competition.


======================================================================
DISCUSSION - FROM WINDOW ALIGNMENT TO EMPIRICAL MOTIF MIRRORING
======================================================================

Experiments 11 and 12 established that natural dependency regions should be
revealed rather than fully prescribed. Experiment 13 therefore avoids another
broad ordering rule and avoids inventing an abstract dependency graph.

Instead, it selects a small motif from ordinary passengers already present in
the standard deterministic manifest:

    shallow potential blocker
    second potential blocker
    deeper-seat trigger
    downstream aisle-pressure passenger

The same type of motif is selected independently for both aisles and placed at a
shared pre-run queue depth. Every other passenger retains the original seeded
relative order. The experiment therefore tests whether an empirically grounded
natural motif can be mirrored across two independent aisle domains.


======================================================================
EXPERIMENT 13 - EMPIRICAL NATURAL-DEPENDENCY MOTIF MIRRORING
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original algorithm depends upon local occupancy relationships rather than
global behavioural assumptions. Experiment 13 follows that principle by
selecting small local passenger motifs from the existing deterministic manifest.
The movement algorithm remains unchanged and must decide whether either motif
actually becomes a dependency region.

Planned Purpose
---------------
Determine whether mirroring the same naturally plausible dependency motif on
both aisles, at a comparable queue depth, produces simultaneous confirmed
regions without prescribing runtime state.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

2. EMPIRICAL NATURAL-DEPENDENCY MOTIF MIRRORING DATASET

   The standard seeded queues are retained except for one small motif per aisle.
   Each available motif contains:

   - a shallow potential blocker near the target row;
   - a second potential blocker;
   - a deeper-seat trigger;
   - an optional downstream passenger capable of contributing aisle pressure.

Independent Variable
--------------------

    MIRRORED PRE-RUN PLACEMENT OF TWO NATURAL PASSENGER MOTIFS

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy, seed and entry-headway range;
- unchanged seat assignments and serving aisles;
- synchronous deterministic execution;
- unchanged movement, blocker and yield-space rules;
- no runtime holding or adaptive admission;
- no inserted dependency region;
- no artificial confirmation or lifecycle extension;
- no backward movement, overtaking or aisle switching;
- no passenger moving beyond the assigned row;
- no behavioural assumptions.

Motif-Mirroring Contract
------------------------

1. Both motifs are selected from passengers already present in the standard
   seeded manifest.
2. No passenger changes seat, row or serving aisle.
3. All non-motif passengers retain their relative seeded order.
4. Each motif contains at least three members before it is moved.
5. Motifs are inserted at one shared deterministic queue depth.
6. No runtime state is monitored to maintain alignment.
7. The frozen architecture alone detects and confirms any region.
8. A hierarchy success requires simultaneous confirmed eligible regions.

New Evidence
------------

- motif size on each aisle;
- natural and mirrored queue depth;
- left/right regions created and confirmed;
- maximum simultaneous active regions;
- priority competitions and deferrals;
- hierarchical selections and aisle wins;
- region overlap duration;
- cluster selections and starts;
- complete cabins and seated totals;
- residual stall families.

Experimental Hypothesis
-----------------------
If the rare hierarchy state depends upon two comparable natural causal motifs
maturing at similar times, empirical motif mirroring should create stronger
cross-aisle lifecycle overlap than general queue-window alignment.

A null result would indicate that even empirically grounded manifest motifs
cannot reliably exercise the hierarchy under the present first-principles model.

Interpretation Safeguards
-------------------------

- A motif is a pre-run passenger-order pattern, not a dependency region.
- The selected passengers may seat without forming congestion.
- Identical motif structure does not guarantee identical runtime timing.
- A single confirmed region is not hierarchy validation.
- More stalls are not automatically successful evidence.
- Zero competitions must be reported exactly.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Simultaneous confirmed left/right regions.
3. Hierarchical selections and deferrals.
4. Region overlap duration.
5. Regions created and confirmed.
6. Cluster selections and starts.
7. Completion and stall family.
8. Tick counts.

Diagram Plan
------------
A later diagram may show one four-passenger motif on each aisle at comparable
queue depth. It must state that the passengers are selected from the ordinary
manifest and that no runtime region is inserted.

======================================================================
EXPERIMENT 13 FINDINGS - EMPIRICAL NATURAL-DEPENDENCY MOTIF MIRRORING
======================================================================

Execution Summary
-----------------
Experiment 13 completed 30 deterministic paired scenarios, producing 60
executions. Both modes retained the frozen Experiment 23 architecture, identical
scenario seeds, cabin configuration, occupancy and entry-headway ranges. The
motif dataset moved only small passenger groups already present in the standard
manifest.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Motif-mirroring average completion: 99.71%.
- Improved pairs: 0 of 30.
- Unchanged pairs: 30 of 30.
- Worse pairs: 0 of 30.
- Net passenger difference: 0.
- Complete cabins: 27 in both modes.
- Standard dependency regions: 37.
- Motif-mirroring dependency regions: 35.
- Standard confirmed regions: 22.
- Motif-mirroring confirmed regions: 21.
- Standard cluster selections: 25.
- Motif-mirroring cluster selections: 25.
- Standard cluster-started events: 31.
- Motif-mirroring cluster-started events: 30.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Motif mirroring preserved all paired completion outcomes.

Every paired scenario seated the same number of passengers. The method therefore
changed local dependency evidence without acting as a broad optimisation or
degradation policy.

2. Scenario 4 produced two confirmed regions.

The standard rear-boundary lock contained one region and one confirmed region.
The motif dataset retained the same 200/204 final stall but produced two regions,
two confirmed regions, one cluster selection and one cluster-started event.

3. Multiple confirmations still did not overlap as eligible competitors.

Scenario 4 is the clearest evidence that the remaining limitation is no longer
region creation or confirmation count. The two confirmed regions matured at
different decision points, so priority competitions and deferrals remained zero.

4. Scenario 19 increased confirmed and selected activity.

The standard execution produced five regions, two confirmed regions and two
cluster selections. Motif mirroring retained five regions but increased
confirmations to three, cluster selections to four and cluster-started events to
five. This shows that natural motifs can enrich lifecycle evidence without
changing the completed outcome.

5. Several scenarios remained exactly reproducible.

Scenarios including 2, 6, 17, 29 and 30 retained the same dependency totals and
outcome families. The motif mechanism therefore did not destabilise the complete
benchmark.

6. The established stall families remained intact.

Scenario 4 retained the four-passenger rear-boundary lock. Scenario 17 retained
a two-passenger rear-boundary lock. Scenario 29 retained the large mixed residual
dependency. Motif mirroring exposed additional lifecycle evidence but did not
artificially remove these first-principles failures.

Scientific Interpretation
-------------------------
Experiment 13 demonstrates that empirically plausible natural motifs are more
effective than fully prescribed topologies at preserving and sometimes
increasing confirmed dependency evidence. However, equal queue depth is not
equal lifecycle timing. The farther target row requires additional aisle travel,
and each motif may also experience different local seat-event delays.

The remaining research question is therefore precise: can two natural motifs be
placed at different pre-run queue depths so that their estimated maturation
opportunities coincide, without any runtime monitoring or intervention?

Experiment 13 Conclusion
------------------------
Empirical Natural-Dependency Motif Mirroring is supported as a stable probe of
dependency-region formation. It produced multiple confirmed regions in a
difficult deterministic scenario but did not activate hierarchy competition.
The next experiment should compensate for target-row travel distance when
placing the two motifs.


======================================================================
DISCUSSION - CONFIRMATION COUNT VERSUS MATURATION COINCIDENCE
======================================================================

A total of two or more confirmed regions within one execution does not establish
hierarchical competition. The regions must be confirmed and eligible at the
same architecture decision point.

Experiment 14 therefore does not change confirmation thresholds, lifecycle
rules or runtime scheduling. It estimates each motif's pre-run maturation
opportunity using two fixed components:

    queue admission depth
    plus
    aisle travel distance to the target row

The farther motif is placed earlier in its queue and the nearer motif later.
This compensates for deterministic travel distance while leaving all runtime
occupancy interactions untouched.


======================================================================
EXPERIMENT 14 - NATURAL-MOTIF LIFECYCLE MATURATION SYNCHRONISATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's algorithm advances passengers through explicit occupancy states one tile
at a time. Experiment 14 uses that same movement principle to estimate pre-run
arrival opportunity: a passenger assigned farther along the aisle requires more
deterministic movement steps before reaching the target row.

The estimate affects only initial queue placement. Amit's movement engine and
the frozen region architecture remain solely responsible for all runtime states.

Planned Purpose
---------------
Determine whether travel-distance-compensated placement of two natural passenger
motifs causes their dependency regions to mature during the same architecture
decision window.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

2. NATURAL-MOTIF LIFECYCLE MATURATION SYNCHRONISATION DATASET

   The standard seeded queues are retained except for one natural motif per
   aisle. Each motif contains up to four existing passengers:

   - shallow potential blocker;
   - second potential blocker;
   - deeper-seat trigger;
   - optional downstream aisle-pressure passenger.

Independent Variable
--------------------

    TRAVEL-DISTANCE-COMPENSATED PRE-RUN MOTIF QUEUE DEPTH

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy, seed and entry-headway range;
- unchanged seats, rows and serving aisles;
- unchanged synchronous execution;
- unchanged movement, blocker and yield-space rules;
- unchanged confirmation and hierarchy thresholds;
- no runtime holding or adaptive admission;
- no inserted region or artificial lifecycle persistence;
- no backward movement, overtaking or aisle switching;
- no behavioural assumptions.

Maturation-Synchronisation Contract
-----------------------------------

1. Both motifs are selected from the standard seeded manifest.
2. All non-motif passengers retain their seeded relative order.
3. A shared estimated arrival target is calculated before execution.
4. Estimated arrival equals queue depth plus target-row travel distance.
5. The farther motif is admitted earlier and the nearer motif later.
6. Queue depths remain bounded within the central half of each aisle queue.
7. No runtime state is read to adjust either motif.
8. The architecture alone decides whether regions form, confirm or compete.
9. Success requires genuine simultaneous confirmed eligible regions.

New Evidence
------------

- natural motif depth and adjusted depth per aisle;
- estimated left/right arrival opportunity;
- maximum simultaneous active regions;
- simultaneous confirmed-region competitions;
- hierarchical selections and deferrals;
- left/right priority wins and ties;
- regions created, confirmed and dissolved;
- region lifetime and overlap;
- cluster selections and starts;
- completion and residual stall family.

Experimental Hypothesis
-----------------------
If Experiment 13 failed because equal queue depth ignored different aisle travel
distances, deterministic distance compensation should increase the probability
that both natural motifs mature within the same decision window.

A null result would indicate that maturation timing is governed primarily by
local occupancy interactions rather than queue depth plus travel distance.

Interpretation Safeguards
-------------------------

- The arrival estimate is approximate and fixed before execution.
- No passenger is held to preserve synchronisation.
- A motif may seat without producing a region.
- Two confirmed regions at separate times do not validate the hierarchy.
- More stalls are not automatically successful evidence.
- Zero competitions must be documented exactly.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Simultaneous confirmed left/right regions.
3. Hierarchical selections and lower-priority deferrals.
4. Maximum simultaneous active regions.
5. Region overlap duration.
6. Regions created and confirmed.
7. Cluster selections and starts.
8. Completion and stall family.
9. Tick counts.

Diagram Plan
------------
A later diagram may show the farther motif placed earlier and the nearer motif
placed later, with both projected toward one estimated maturation window. The
diagram must state that this is a pre-run estimate and not runtime control.

======================================================================
EXPERIMENT 14 FINDINGS - NATURAL-MOTIF LIFECYCLE MATURATION SYNCHRONISATION
======================================================================

Execution Summary
-----------------
Experiment 14 completed 30 deterministic paired scenarios, producing 60
executions. Both modes retained the frozen Experiment 23 architecture, identical
scenario seeds, cabin configuration, occupancy and entry-headway ranges. The
experimental mode changed only the bounded pre-run queue depth of two natural
passenger motifs.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Synchronised average completion: 99.71%.
- Improved pairs: 0 of 30.
- Unchanged pairs: 30 of 30.
- Worse pairs: 0 of 30.
- Net passenger difference: 0.
- Complete cabins: 27 in both modes.
- Standard dependency regions: 37.
- Synchronised dependency regions: 35.
- Standard confirmed regions: 22.
- Synchronised confirmed regions: 21.
- Standard cluster selections: 25.
- Synchronised cluster selections: 25.
- Standard cluster-started events: 31.
- Synchronised cluster-started events: 30.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Travel-distance compensation preserved all paired outcomes.

Every paired scenario seated the same number of passengers and retained the same
completion or stall family. The experiment therefore altered lifecycle evidence
without destabilising the benchmark.

2. Scenario 4 again produced two confirmed regions.

The synchronised mode retained the 200/204 rear-boundary lock but produced two
regions, two confirmations, one cluster selection and one cluster-started event,
compared with one confirmed region in the standard execution.

3. Multiple confirmations remained temporally separate.

Despite two confirmations in Scenario 4, priority competitions and lower-
priority deferrals remained zero. The regions were not simultaneously confirmed
and eligible at one hierarchy decision point.

4. Scenario 5 remained stable.

Both modes produced two regions, one confirmed region, one selection and one
cluster-started event. The synchronised mode increased lifecycle updates slightly
without changing the completed outcome.

5. Queue depth plus travel distance is an incomplete maturation model.

The absence of hierarchy competition indicates that local seat events, aisle
occupancy, blocker release and middle-bank interactions dominate the actual
maturation timing after entry.

Scientific Interpretation
-------------------------
Experiment 14 rules out a straightforward deterministic explanation in which
different target-row travel distances alone prevent lifecycle coincidence.
Pre-run synchronisation can preserve and enrich dependency evidence, but it
cannot control the many local interactions that determine when a region confirms
and how long it remains eligible.

The next experiment should therefore stop attempting to equalise two motifs.
Instead, it should construct two deliberately different natural environments so
that one region develops slowly and persistently while the other develops more
quickly. Such lifecycle diversity may create an overlap window without runtime
holding.

Experiment 14 Conclusion
------------------------
Natural-Motif Lifecycle Maturation Synchronisation is retained as a stable
negative result. It preserved multiple confirmed-region evidence but did not
activate hierarchy competition. The next investigation should vary local
dependency growth characteristics rather than queue-entry timing.


======================================================================
DISCUSSION - FROM SYNCHRONISATION TO LIFECYCLE DIVERSITY
======================================================================

Experiments 11-14 progressively investigated discovery, queue-window alignment,
natural motif mirroring and travel-distance compensation. The findings show that
two regions do not need identical starting conditions. They need overlapping
eligible lifetimes.

Experiment 15 therefore introduces deterministic asymmetry:

    one heavier, blocker-rich and reinforced motif
    one lighter, lower-pressure motif

Both motifs remain natural passenger groups selected from the seeded manifest.
The heavier environment is expected to develop more slowly and persist longer;
the lighter environment may mature earlier and dissolve sooner. Their different
rates may create an overlap that equalised motifs failed to produce.


======================================================================
EXPERIMENT 15 - DETERMINISTIC DEPENDENCY LIFECYCLE DIVERSITY
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original algorithm evaluates local occupancy and dependency conditions
rather than imposing global behaviour. Experiment 15 retains this principle by
constructing two different local input environments while leaving the movement
engine unchanged.

The left motif is blocker-rich and reinforced. The right motif is deliberately
lighter. Any difference in lifecycle growth must arise through ordinary
deterministic occupancy transitions.

Planned Purpose
---------------
Determine whether two natural dependency motifs with deliberately different
local growth characteristics create overlapping confirmed-region lifetimes and
activate the frozen hierarchy.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

2. DETERMINISTIC DEPENDENCY LIFECYCLE DIVERSITY DATASET

   The standard seeded queues are retained except for two natural motifs:

   - Heavy motif: shallow blocker, second blocker, deep trigger, downstream
     pressure passenger and adjacent reinforcement.
   - Light motif: shallow blocker, deep trigger and limited downstream pressure.

Independent Variable
--------------------

    DETERMINISTIC DIFFERENCE IN LOCAL DEPENDENCY-GROWTH ENVIRONMENT

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy, seed and entry-headway range;
- unchanged seats, rows and serving aisles;
- synchronous deterministic execution;
- unchanged movement, blocker and yield-space rules;
- unchanged confirmation and hierarchy thresholds;
- no runtime holding or adaptive admission;
- no inserted region or artificial lifecycle persistence;
- no backward movement, overtaking or aisle switching;
- no behavioural assumptions.

Lifecycle-Diversity Contract
----------------------------

1. Both motifs are selected from passengers already present in the standard
   seeded manifest.
2. The heavy and light motifs are inserted at one shared deterministic depth.
3. All non-motif passengers retain their seeded relative order.
4. The heavy motif contains more blocker and reinforcement components.
5. The light motif contains fewer local pressure components.
6. No runtime state is read or used to preserve either region.
7. The architecture alone determines whether the motifs form regions.
8. Success requires simultaneous confirmed eligible regions.

New Evidence
------------

- heavy and light motif sizes;
- regions created and confirmed by aisle;
- maximum simultaneous active regions;
- priority competitions and deferrals;
- hierarchical selections and aisle wins;
- region lifetimes and overlap duration;
- cluster selections and starts;
- completion and residual stall family;
- tick counts.

Experimental Hypothesis
-----------------------
If identical or synchronised motifs fail because their lifecycle windows rise and
fall together or miss one another narrowly, deterministic lifecycle diversity
may allow a persistent heavy region to overlap a faster light region.

A null result would indicate that queue-manifest construction alone cannot
reliably produce the hierarchy state under the current first-principles model.

Interpretation Safeguards
-------------------------

- Heavy and light describe input composition, not guaranteed runtime behaviour.
- No region is inserted, delayed or protected.
- More congestion is not automatically a successful result.
- A single confirmed region is not hierarchy validation.
- Completion remains secondary to genuine competition evidence.
- Zero competitions must be reported exactly.

Primary Evaluation Order
------------------------

1. Genuine priority competitions.
2. Simultaneous confirmed left/right regions.
3. Hierarchical selections and lower-priority deferrals.
4. Maximum simultaneous active regions.
5. Region overlap duration.
6. Regions created and confirmed.
7. Cluster selections and starts.
8. Completion and stall family.
9. Tick counts.

Diagram Plan
------------
A later diagram may contrast a heavy five-passenger motif with a lighter three-
passenger motif at the same queue depth. The figure must state that these are
pre-run natural passenger groups and not runtime dependency objects.

======================================================================
EXPERIMENT 15 FINDINGS - DETERMINISTIC DEPENDENCY LIFECYCLE DIVERSITY
======================================================================

Execution Summary
-----------------
Experiment 15 completed 30 deterministic paired scenarios, producing 60
executions. Both modes retained the frozen Experiment 23 architecture, identical
scenario seeds, cabin configuration, occupancy and entry-headway ranges. The
experimental mode changed only the composition and placement of one heavy and
one light natural motif.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Lifecycle-diversity average completion: 99.71%.
- Improved pairs: 0 of 30.
- Unchanged pairs: 30 of 30.
- Worse pairs: 0 of 30.
- Net passenger difference: 0.
- Complete cabins: 27 in both modes.
- Standard dependency regions: 37.
- Lifecycle-diversity dependency regions: 36.
- Standard confirmed regions: 22.
- Lifecycle-diversity confirmed regions: 21.
- Standard cluster selections: 25.
- Lifecycle-diversity cluster selections: 25.
- Standard cluster-started events: 31.
- Lifecycle-diversity cluster-started events: 31.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. Lifecycle diversity preserved every paired completion outcome.

All paired scenarios seated identical passenger totals and retained the same
completion or residual-stall families. The heavy/light construction therefore
did not destabilise the frozen movement engine.

2. Scenario 3 revealed a new confirmed dependency region.

The standard execution produced no dependency region. The diversity dataset
produced two regions, one confirmed region, one cluster selection and one
cluster-started event while still completing all 342 passengers.

3. Scenario 4 reverted to one persistent confirmed region.

Unlike Experiments 13 and 14, the diversity dataset did not produce a second
confirmed region in the established rear-boundary-lock scenario. Both modes
retained one region, one confirmation and the same 200/204 final outcome.

4. Aggregate dependency evidence remained effectively unchanged.

Regions decreased only from 37 to 36, confirmations from 22 to 21, while cluster
selections and starts were preserved. Heavy/light motif asymmetry therefore
redistributed local evidence rather than creating lifecycle overlap.

5. No hierarchy competition occurred.

Priority competitions and lower-priority deferrals remained zero. Different
local motif composition did not produce two simultaneously confirmed eligible
regions.

Scientific Interpretation
-------------------------
Experiment 15 demonstrates that lifecycle diversity can expose hidden local
dependency evidence, as shown by Scenario 3, but does not reliably create an
overlap window. Changing motif composition remains a passenger-order technique.
It does not provide a distinct temporal admission layer above the movement
engine.

The evidence therefore supports a transition from passenger-level manifest
manipulation to a deterministic admission-window architecture. This creates a
new input layer while preserving the policy-neutral movement and dependency
systems beneath it.

Experiment 15 Conclusion
------------------------
Deterministic Dependency Lifecycle Diversity is retained as a stable negative
result. The experiment preserved outcomes and revealed one additional local
dependency case, but it did not activate hierarchical arbitration. Further motif
refinement is unlikely to answer the remaining question.


======================================================================
RESEARCH INTERPRETATION - WHY BOARDING POLICY WAS POSTPONED
======================================================================

Real-World Operational Order
----------------------------
In real airline operations, the boarding policy is applied before passengers
enter the aircraft. The airline may assign sections, groups, priorities or other
admission rules, and the resulting passenger stream is then processed by the
cabin.

Scientific Modelling Order
--------------------------
This research deliberately used the opposite order. It first established a
policy-neutral deterministic movement backbone and then studied:

- aisle occupancy;
- seat blockers;
- dependency formation;
- persistent region lifecycles;
- confirmation;
- hierarchy;
- cluster resolution.

Introducing section boarding at the beginning would have made it difficult to
separate intrinsic movement effects from policy-created effects. A rear-first or
zone-based stream could have suppressed, concentrated or relocated dependencies
before the underlying architecture was understood.

Consensus from the Evidence
---------------------------
The evidence supports two valid orders for two different purposes:

For scientific architecture research:

    deterministic movement foundation
        -> dependency analysis
        -> lifecycle and hierarchy
        -> boarding-policy evaluation

For an operational airline simulator:

    airline boarding policy
        -> deterministic admission queue
        -> movement engine
        -> dependency analysis and resolution

These orders are not contradictory. Boarding policy selects the inputs; the
validated deterministic engine processes them.

What the Current Simulator Models
---------------------------------
The simulator provides a first-principles representation of:

- deterministic passenger admission;
- forward, row-bounded aisle movement;
- local occupancy;
- seat-access blockers;
- temporary aisle reoccupation;
- middle-bank coordination;
- dependency-region formation and resolution.

It does not claim to model every real-world behavioural or operational factor.
Human hesitation, conversations, variable baggage handling, instruction
non-compliance, family behaviour and airline commercial priorities remain
outside this architecture.

Future Policy Layer
-------------------
Section boarding, rear-to-front boarding, front-to-back boarding, WilMA,
reverse-pyramid methods and commercial boarding groups should later be treated
as interchangeable deterministic admission strategies above the frozen engine.

This separation allows future policy studies to compare not only total boarding
time but also dependency-region count, confirmation, lifetime, cluster activity,
hierarchy use and residual stall family.


======================================================================
DISCUSSION - THE BRIDGE FROM FOUNDATION TO POLICY
======================================================================

The completed work does not lack a boarding policy. It establishes the layer
that boarding policies require.

Experiment 16 creates the architectural bridge:

    deterministic admission windows
        -> passenger movement
        -> dependency formation
        -> confirmation
        -> hierarchy
        -> cluster resolution

The windows are not yet rear, middle or front cabin sections. They preserve the
ordinary seeded passenger order and introduce only deterministic release
boundaries. Future work can replace the neutral windows with operational
boarding groups without modifying the movement engine.


======================================================================
EXPERIMENT 16 - DETERMINISTIC ADMISSION-WINDOW ARCHITECTURE
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original people-movement algorithm processes explicit occupancy inputs
through deterministic local rules. Experiment 16 adds a separate deterministic
admission layer above that backbone.

The admission layer decides when a passenger becomes eligible to enter. Amit's
movement engine still decides whether the entry tile is free and how the
passenger advances thereafter.

Planned Purpose
---------------
Determine whether neutral deterministic admission windows can produce temporal
separation between passenger cohorts while preserving the policy-free movement
architecture and all controlled variables.

This experiment also establishes the reusable interface through which later
airline boarding policies can be introduced.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

   Passengers use the established seeded aisle queues and ordinary controlled
   entry headway.

2. DETERMINISTIC ADMISSION-WINDOW DATASET

   The same seeded aisle queues are divided into three consecutive cohorts.
   Passengers retain their original relative order. Cohorts two and three receive
   fixed pre-run release ticks.

Independent Variable
--------------------

    DETERMINISTIC COHORT RELEASE WINDOW

Controlled Conditions
---------------------

- frozen Experiment 23 architecture;
- identical cabin, occupancy, seed and passenger assignments;
- identical seeded relative queue order;
- identical numerical entry-headway ranges;
- synchronous deterministic execution;
- unchanged movement, blocker and yield-space rules;
- unchanged clustering, lifecycle confirmation and hierarchy;
- no row-based or seat-based airline policy;
- no runtime congestion response;
- no adaptive release;
- no backward movement, overtaking or aisle switching;
- no behavioural assumptions.

Admission-Window Contract
-------------------------

1. Each aisle queue is divided into three consecutive cohorts.
2. Passenger relative order is unchanged.
3. Window membership depends only on seeded queue position.
4. Release ticks are calculated before execution.
5. Release ticks use cohort size, controlled headway and a fixed deterministic
   pause.
6. A released passenger must still wait for an empty entry tile.
7. No passenger is held after entering the aircraft.
8. Windows do not correspond to front, middle or rear cabin sections.
9. No runtime dependency state can alter a release tick.
10. The frozen movement architecture remains unchanged beneath the new layer.

New Evidence
------------

- cohort release ticks;
- passengers admitted from each window;
- ticks waiting for window release;
- dependency regions created and confirmed;
- maximum simultaneous active regions;
- priority competitions and deferrals;
- cluster selections and starts;
- complete cabins and seated totals;
- residual stall families;
- total and average waiting ticks.

Experimental Hypothesis
-----------------------
Neutral deterministic admission windows may create lifecycle separation that
continuous admission and local motif manipulation could not produce. They may
also reveal whether temporal cohort boundaries suppress or preserve dependency
formation.

The primary architectural objective is not immediate optimisation. It is to
validate a clean admission interface above the frozen movement engine.

Interpretation Safeguards
-------------------------

- Admission windows are architecture, not an airline recommendation.
- Cohorts are based on queue position, not cabin section.
- Better boarding time is secondary to architectural evidence.
- Waiting outside due to a release gate must remain segregated from congestion.
- No policy conclusion may be inferred from this experiment alone.
- Zero hierarchy competitions must be reported exactly.

Primary Evaluation Order
------------------------

1. Correct separation of admission-window and congestion evidence.
2. Genuine priority competitions.
3. Simultaneous confirmed regions.
4. Hierarchical selections and deferrals.
5. Regions created and confirmed.
6. Cluster selections and starts.
7. Completion and residual stall family.
8. Waiting time and tick count.

Future Policy Compatibility
---------------------------
After validation, the admission-window layer can support later deterministic
policies by changing only cohort membership and release order, including:

- rear-to-front sections;
- front-to-back sections;
- window-middle-aisle ordering;
- reverse-pyramid ordering;
- airline-specific boarding groups.

Those future policies remain separate experiments and require operational
interpretation beyond the present architecture.

======================================================================
EXPERIMENT 16 FINDINGS - DETERMINISTIC ADMISSION-WINDOW ARCHITECTURE
======================================================================

Execution Summary
-----------------
Experiment 16 completed 30 deterministic paired scenarios, producing 60
executions. Both modes retained the frozen Experiment 23 movement and dependency
architecture, identical scenario seeds, cabin configuration, occupancy,
passenger assignments, seeded relative queue order and numerical entry-headway
ranges. The experimental mode added three fixed queue-position release windows.

Overall Paired Result
---------------------

- Standard average completion: 99.71%.
- Admission-window average completion: 99.71%.
- Improved pairs: 0 of 30.
- Unchanged pairs: 30 of 30.
- Worse pairs: 0 of 30.
- Net passenger difference: 0.
- Complete cabins: 27 in both modes.
- Standard dependency regions: 37.
- Admission-window dependency regions: 37.
- Standard confirmed regions: 22.
- Admission-window confirmed regions: 22.
- Standard cluster selections: 25.
- Admission-window cluster selections: 25.
- Standard cluster-started events: 31.
- Admission-window cluster-started events: 31.
- Priority competitions: 0 in both modes.

Principal Findings
------------------

1. The admission-window interface was implemented without changing the engine.

The experimental dataset successfully inserted deterministic release gates above
the ordinary entry-tile and movement logic. Passengers still entered only when
the entry tile was free, and every in-cabin rule remained frozen.

2. Queue-position windows produced complete behavioural equivalence.

Across all 30 pairs, completion, dependency-region totals, confirmations,
cluster activity and residual stall families were unchanged. Many scenarios
also retained identical tick counts.

3. Scenario 4 reproduced the exact rear-boundary lock.

Both modes stalled at 200/204 with the same four-passenger chain, critical
blocker, insufficient rear yield space and persistent confirmed region.

4. Temporal separation alone did not alter local dependency topology.

The windows delayed later portions of the same seeded queue but did not change
which destination rows and seat-access demands were grouped together. The
movement engine therefore encountered essentially the same critical local
interactions.

5. No hierarchy competition occurred.

Priority competitions and lower-priority deferrals remained zero. Neutral timing
windows did not produce simultaneous confirmed eligible regions.

Scientific Interpretation
-------------------------
Experiment 16 validates the admission layer as a clean architectural bridge, but
rejects queue-position timing alone as a meaningful dependency variable in the
tested form.

The evidence suggests that the spatial composition of each admitted cohort is
more important than a release delay applied to an otherwise unchanged passenger
sequence. Airlines commonly influence this composition through sections, zones
or seat categories. The next experiment should therefore structure cohorts by
aircraft geography while preserving deterministic execution and the frozen
movement engine.

Experiment 16 Conclusion
------------------------
The Deterministic Admission-Window Architecture is supported as a reusable
interface above the first-principles movement engine. Its neutral queue-position
windows produced no material behavioural change, establishing a clean baseline
for geographically structured admission.


======================================================================
RESEARCH INTERPRETATION - TIMING VERSUS SPATIAL COMPOSITION
======================================================================

Experiment 15 altered local passenger motifs without producing hierarchy
competition. Experiment 16 altered admission timing without changing the
passengers grouped together, and produced exact behavioural equivalence.

Together, these results distinguish two variables:

    when passengers become eligible to enter
    which destination regions coexist in an admitted cohort

The evidence indicates that the second variable is now the stronger candidate.
A timing gate cannot substantially change dependency topology when the same
spatial mix eventually reaches the same critical rows in the same relative
order.

This does not invalidate the admission-window layer. It clarifies its role.
The layer is the interface through which future cohort definitions can be
changed while the deterministic movement and dependency engines remain fixed.


======================================================================
DISCUSSION - THE FIRST POLICY-NEUTRAL GEOGRAPHIC BRIDGE
======================================================================

Experiment 17 is not presented as a recommendation that airlines should board
rear-to-front. It is a controlled topology experiment.

Its purpose is to determine whether grouping passengers by destination region
changes dependency evidence where neutral timing windows did not. The selected
rear-middle-front order is a deterministic stress dataset and a bridge toward
later policy comparisons.

If the geographic dataset changes outcomes, the evidence will show that boarding
policy can influence dependency formation through spatial composition. If it
does not, the architecture will have ruled out another major admission-level
explanation.


======================================================================
EXPERIMENT 17 - DETERMINISTIC GEOGRAPHIC ADMISSION COHORTS
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original movement algorithm remains the lower-level deterministic engine.
Experiment 17 changes only the order in which destination-row groups are made
eligible for entry.

After entry, every passenger remains row-bounded and follows the same occupancy,
blocker, seat-event and dependency rules.

Planned Purpose
---------------
Determine whether deterministic cohorts based on destination geography alter
dependency-region formation, lifecycle confirmation, residual stalls or
hierarchy utilisation compared with the standard seeded queue.

Paired Modes
------------

1. STANDARD DETERMINISTIC DATASET

2. DETERMINISTIC GEOGRAPHIC COHORT DATASET

   Each aisle's seeded passenger list is divided into:

   - rear third;
   - middle third;
   - front third.

   Cohorts are admitted rear, then middle, then front. Seeded relative order is
   preserved inside each cohort.

Independent Variable
--------------------

    DESTINATION-ROW GEOGRAPHIC COHORT MEMBERSHIP AND RELEASE ORDER

Controlled Conditions
---------------------

- frozen Experiment 23 movement and dependency architecture;
- identical cabin, occupancy, seed and passenger assignments;
- identical serving-aisle assignments;
- seeded relative order preserved inside every cohort;
- identical numerical entry-headway ranges;
- synchronous deterministic execution;
- unchanged movement, blocker and yield-space rules;
- unchanged clustering, lifecycle confirmation and hierarchy;
- no runtime congestion response;
- no adaptive release;
- no backward movement, overtaking or aisle switching;
- no behavioural assumptions.

Geographic-Cohort Contract
--------------------------

1. Row thirds are calculated deterministically from cabin length.
2. Rear passengers enter before middle passengers.
3. Middle passengers enter before front passengers.
4. Relative seeded order is unchanged inside each cohort.
5. Fixed release ticks are calculated before execution.
6. A released passenger still requires an empty entry tile.
7. No passenger is held after entering the aircraft.
8. No runtime dependency state changes cohort release.
9. The dataset is a research stress condition, not an airline recommendation.
10. The frozen architecture alone determines all dependency evidence.

New Evidence
------------

- passenger count in each geographic cohort;
- cohort release ticks;
- outside waiting attributable to release gates;
- dependency regions created and confirmed;
- maximum simultaneous active regions;
- priority competitions and deferrals;
- cluster selections and starts;
- completion and residual stall family;
- total and average waiting ticks.

Experimental Hypothesis
-----------------------
If spatial composition is the dominant admission-level variable, rear-middle-
front cohorts should produce measurably different dependency topology from the
standard seeded queue even though movement and entry-headway rules remain
unchanged.

A null result would show that broad row geography alone is also insufficient and
that later policy research must investigate seat category, narrower zones or
combined spatial rules.

Interpretation Safeguards
-------------------------

- This is not yet a commercial boarding-policy recommendation.
- Faster completion is not sufficient evidence of a better policy.
- Outside release-gate waiting must remain separate from in-cabin congestion.
- Geographic ordering may suppress as well as create dependency regions.
- A single confirmed region is not hierarchy validation.
- Zero competitions must be reported exactly.

Primary Evaluation Order
------------------------

1. Correct geographic cohort and release-gate operation.
2. Change in dependency-region topology.
3. Genuine priority competitions.
4. Simultaneous confirmed regions.
5. Hierarchical selections and deferrals.
6. Cluster selections and starts.
7. Completion and residual stall family.
8. Waiting time and tick count.

Future Policy Compatibility
---------------------------
The same admission interface can later test:

- front-to-back sections;
- smaller rear-to-front zones;
- window-middle-aisle groups;
- reverse-pyramid cohorts;
- airline-specific group definitions.

Those later studies should compare operational performance and dependency
evidence separately.


======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 16
======================================================================

1. The frozen movement and dependency architecture remains unchanged.
2. Experiment 16 validated a separate deterministic admission interface.
3. All 30 paired outcomes and architecture metrics were unchanged.
4. Queue-position timing alone did not change dependency topology.
5. Scenario 4 reproduced the identical rear-boundary lock.
6. Priority competitions remained zero.
7. Spatial cohort composition is now the stronger candidate variable.
8. Experiment 17 uses rear-middle-front destination-row cohorts.
9. The geographic dataset is a stress condition, not a policy recommendation.
10. No passenger may move backwards or beyond the assigned row.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v17 - EXPERIMENT 17 DESIGN
======================================================================

======================================================================
EXPERIMENT 17 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The full deterministic verification run completed all 30 paired scenarios
(60 executions). The standard and geographic modes used identical cabins,
occupancy, seeds, passenger assignments, serving aisles, numerical headways,
movement rules, dependency rules and hierarchy controls.

Observed Results
----------------
- Standard passengers seated: 7,649 of 7,671 (99.71%).
- Geographic-cohort passengers seated: 6,961 of 7,671 (90.74%).
- Net passenger difference: -688.
- Successful completions: 46 of 60 executions.
- Geographic mode produced lower completion in 11 of 30 pairs.
- The strongest geographic completion losses included 98.18%, 99.07%,
  57.02%, 79.69%, 47.66%, 60.96%, 64.04% and 60.95% completion cases.
- The longest observed geographic dependency chain contained 28 passengers
  (27 waiting passengers behind the terminal passenger).
- Maximum outside-queue saturation reached 129 waiting passengers; the
  highlighted Experiment 17 case reached 112.
- New residual families included REAR_BOUNDARY_LOCK and
  ENTRY_SATURATION_WITH_INTERNAL_CHAIN.
- Priority competitions and lower-priority deferrals remained zero.

Principal Findings
------------------
1. Broad geographic grouping changed behaviour materially where neutral timing
   windows in Experiment 16 had not.
2. Rear-first thirds concentrated similarly destined passengers and created
   long internal dependency chains before later cohorts could enter.
3. Entry saturation was not merely external delay. In the strongest cases it
   coexisted with a stable internal chain, so the entry tile remained blocked
   while substantial portions of the queue were still outside.
4. The positive scientific result is therefore not that rear-to-front thirds
   are operationally superior. It is that spatial admission composition is a
   causal variable in dependency formation.
5. The same frozen movement engine produced the new behaviour; the independent
   variable was admission geography alone.

Discussion Incorporated Before Experiment 18
---------------------------------------------
Experiment 17 establishes the bridge from asking whether geography matters to
asking how much geographic granularity is appropriate. Broad thirds can be too
coarse: they create large homogeneous waves that may saturate an aisle and form
extended dependency chains. A finer partition may limit the size of each wave
while retaining the spatial separation that Experiment 16 lacked.

Normal entry-headway rules remain applicable inside every cohort. Passengers do
not enter simultaneously. Each released passenger still waits for the ordinary
deterministic headway and an empty entry tile. The cohort mechanism changes
eligibility and spatial composition only.

Experiment 17 Conclusion
------------------------
Experiment 17 is the first admission experiment in this series to produce a
major behavioural change. It supports spatial composition as a dominant
admission-level variable and justifies a controlled granularity study.


======================================================================
EXPERIMENT 18 - FINE-GRAINED DETERMINISTIC GEOGRAPHIC ADMISSION ZONES
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic movement engine remains unchanged. Experiment 18
modifies only the spatial admission granularity above that engine. Once a
passenger enters, all row-bounded movement, blocker, yield-space, seat-event,
dependency-region, lifecycle and hierarchy rules are identical.

Planned Purpose
---------------
Determine whether dividing the cabin into six smaller deterministic geographic
zones can reduce the extreme admission saturation observed with Experiment 17's
three broad cohorts while preserving useful dependency diversity.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINE-GRAINED GEOGRAPHIC ZONE DATASET.

Fine-Grained Zone Contract
--------------------------
- Every cabin is divided longitudinally into six deterministic zones.
- Zones enter from rear to front.
- Seeded relative order is preserved inside each zone.
- Normal deterministic entry-headway rules still apply within every zone.
- Fixed release ticks are calculated before execution.
- A released passenger still requires an empty entry tile.
- No runtime congestion state changes the zone order or release schedule.
- No passenger is held or redirected after entry.

Independent Variable
--------------------
    GEOGRAPHIC ADMISSION GRANULARITY: SIX ZONES INSTEAD OF THREE COHORTS

Primary Evaluation Order
------------------------
1. Correct six-zone mapping and deterministic release.
2. Outside-queue saturation compared with Experiment 17.
3. Completion loss and converted-success cases.
4. Dependency-chain length and residual stall family.
5. Dependency-region creation and confirmation diversity.
6. Genuine priority competitions and deferrals.
7. Cluster selections and starts.
8. Total waiting and tick count.

Experimental Hypothesis
-----------------------
Finer deterministic geographic grouping may reduce the size of each admitted
wave and therefore reduce extreme entry saturation. It may nevertheless retain
more spatial structure than Experiment 16's timing-only windows, allowing useful
dependency differences to remain visible.

Interpretation Safeguards
-------------------------
- Finer zones are not assumed to be better.
- Reduced outside waiting does not by itself prove reduced in-cabin congestion.
- A completion improvement must be checked for displaced dependency costs.
- The six-zone order is a controlled stress dataset, not an airline policy
  recommendation.
- Zero hierarchy competitions must be reported exactly if observed.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 17
======================================================================
1. Timing-only windows produced behavioural equivalence in Experiment 16.
2. Broad geographic thirds produced a major behavioural change in Experiment 17.
3. Spatial admission composition is therefore a validated causal variable.
4. Broad cohorts can create rear-boundary locks and entry saturation with long
   internal dependency chains.
5. Normal entry headway remains active inside every cohort and zone.
6. Experiment 18 tests six smaller rear-to-front zones.
7. The movement and dependency architecture remains frozen.
8. The next question is granularity, not whether geography matters.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v18 - EXPERIMENT 18 DESIGN
======================================================================


======================================================================
EXPERIMENT 18 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The full deterministic run completed all 30 paired scenarios (60 executions).
Both modes used identical cabins, occupancy, scenario seeds, passenger manifests,
serving aisles, numerical entry-headway ranges, synchronous movement, seat-event
rules, dependency-region lifecycle and hierarchy controls. Only the pre-run
admission ordering and six-zone release schedule differed.

Observed Results
----------------
- Standard passengers seated: 7,492 of 7,789 (96.19%).
- Fine-zone passengers seated: 6,209 of 7,789 (79.72%).
- Net fine-zone passenger difference: -1,283.
- Successful completions: standard 19 of 30; fine-zone 12 of 30.
- Total successful executions: 31 of 60.
- Fine-zone completion was lower in 16 of 30 pairs.
- Fine-zone mode converted three incomplete standard cases to success, showing
  that the effect was not uniformly negative.
- The largest single paired loss was 153 seated passengers.
- The largest observed increase in the outside queue was 117 passengers.
- The longest observed unresolved dependency chain contained 35 passengers.
- Fine-zone mode created 22 dependency regions, of which 20 reached confirmation.
- Fine-zone hierarchy activity was substantial: 3,710 priority competitions and
  3,446 lower-priority deferrals were recorded.
- Fine-zone cluster intervention remained rare: one selection and two starts.
- Dominant residual families included REAR_BOUNDARY_LOCK and
  ENTRY_SATURATION_WITH_INTERNAL_CHAIN.

Principal Findings
------------------
1. Increasing geographic granularity was not monotonically beneficial.
2. Six uninterrupted rear-to-front zones often concentrated passengers into
   narrow destination bands and generated severe local aisle saturation.
3. The strongest cases combined a blocked entry tile, a long internal dependency
   chain and a large outside queue. External delay and internal congestion were
   therefore parts of the same causal structure rather than separate effects.
4. Experiment 18 did not invalidate Experiment 17. Together they show that
   geography matters, but that excessive spatial localisation can become
   counterproductive.
5. The experiment identified a non-linear operating region: too little spatial
   structure produced little change in Experiment 16, broad structure produced
   major change in Experiment 17, and finer uninterrupted structure produced
   severe saturation in Experiment 18.
6. Unlike the earlier admission experiments, Experiment 18 exercised the frozen
   hierarchy at scale. The large number of genuine competitions and deferrals is
   important architecture-validation evidence even though operational completion
   was frequently worse.
7. The scarcity of cluster starts despite extensive hierarchy activity shows
   that prioritisation can rank competing regions without guaranteeing that a
   currently executable release exists.

Discussion Incorporated Before Experiment 19
---------------------------------------------
The next experiment should not add still more zones. Experiment 18 has already
shown that smaller geographic groups, when released as uninterrupted local waves,
can intensify rather than relieve congestion.

The stronger next question is whether the spatial signal can be retained while
limiting the instantaneous size of each local wave. This requires a controlled
change to release pacing, not a change to the movement engine or an adaptive
runtime policy.

Normal entry-headway rules remain active for every passenger. A micro-batch gate
therefore adds an eligibility boundary above the existing entry rule; it does not
replace numerical headway or allow simultaneous insertion.

Experiment 18 Conclusion
------------------------
Experiment 18 establishes a turning point in the admission programme. Geographic
structure is causal, but finer subdivision alone is not a solution. Excessive
local concentration can create very long dependency chains, substantial outside
queues and repeated hierarchy competitions. The next experiment must preserve
geographic organisation while deterministically bounding each release wave.


======================================================================
EXPERIMENT 19 - PACED GEOGRAPHIC MICRO-BATCH ADMISSION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic movement backbone remains unchanged. Experiment
19 introduces only a pre-execution admission schedule above that backbone. Once
a passenger enters, the same one-tile synchronous movement, row limit, blocker
yield, seat event, reservation, dependency-region lifecycle and hierarchy rules
continue without modification.

Planned Purpose
---------------
Determine whether fixed deterministic micro-batch pacing can preserve the useful
spatial dependency diversity of geographic admission while reducing the extreme
entry saturation produced by Experiment 18's uninterrupted six-zone waves.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. PACED GEOGRAPHIC MICRO-BATCH DATASET.

Paced Geographic Contract
-------------------------
- The cabin remains divided into six longitudinal destination-row zones.
- Zone order remains rear to front.
- Seeded relative order is preserved inside every zone.
- Each zone is divided before execution into fixed deterministic micro-batches.
- Every micro-batch receives a pre-calculated earliest release tick.
- A fixed pause separates consecutive micro-batches.
- A fixed zone-transition pause separates adjacent geographic zones.
- Normal deterministic numerical entry headway remains active.
- The entry tile must still be empty before insertion.
- The schedule does not inspect runtime congestion, regions or seat events.
- No passenger is held, redirected or reordered after entering the aircraft.

Independent Variable
--------------------
    FIXED MICRO-BATCH PACING WITHIN SIX GEOGRAPHIC ZONES

Primary Evaluation Order
------------------------
1. Correct zone, micro-batch and pre-run release construction.
2. Outside-queue saturation compared with Experiment 18.
3. Completion loss and converted-success cases.
4. Longest dependency-chain length.
5. Entry-saturation and rear-boundary residual families.
6. Confirmed-region concurrency and hierarchy competitions.
7. Lower-priority deferrals, cluster selections and starts.
8. Total waiting and completion ticks.

Experimental Hypothesis
-----------------------
If Experiment 18's main failure mechanism was the uninterrupted size of each
local release wave, fixed micro-batch pauses should reduce entry saturation and
shorten internal dependency chains while retaining more geographic structure
than the timing-only windows of Experiment 16.

A negative result would show that simple time pacing is insufficient once the
queue remains strongly ordered by destination geography. That would justify a
later experiment changing composition between successive geographic releases
rather than merely inserting longer pauses.

Interpretation Safeguards
-------------------------
- Micro-batch pacing is fixed and non-adaptive.
- A lower outside queue alone is not sufficient evidence of improvement.
- Longer elapsed time may be an intentional pacing cost and must be reported.
- Hierarchy activation and operational completion remain separate outcomes.
- The dataset is a controlled research condition, not an airline recommendation.
- Any reduction in dependency chains must be checked for displaced waiting.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 18
======================================================================
1. Timing-only windows were behaviourally neutral in Experiment 16.
2. Broad geographic cohorts changed dependency topology in Experiment 17.
3. Six uninterrupted geographic zones frequently intensified saturation in
   Experiment 18.
4. Geographic granularity is therefore non-linear rather than monotonically
   beneficial.
5. Experiment 18 produced genuine hierarchy utilisation at scale.
6. Priority ranking did not guarantee an executable cluster intervention.
7. Normal numerical headway remains active inside all cohorts, zones and batches.
8. Experiment 19 retains six zones but bounds their release into fixed
   deterministic micro-batches.
9. Runtime movement and dependency architecture remain frozen.
10. The next question is whether pacing can separate useful spatial structure
    from destructive local concentration.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v19 - EXPERIMENT 19 DESIGN
======================================================================



======================================================================
EXPERIMENT 19 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic run executed all 30 paired scenarios (60 total
executions). The standard and paced modes retained identical cabin geometry,
occupancy, scenario seed, passenger manifest, serving-aisle assignment, entry
headway range, synchronous movement, seat-event rules, dependency-region
lifecycle and hierarchy controls. Only the precomputed six-zone micro-batch
admission schedule differed.

Observed Results
----------------
- Successful completions across both modes: 29 of 60.
- Incomplete or stalled executions: 31 of 60.
- Paced micro-batches converted at least one incomplete standard case to full
  success; Scenario 3 improved from 196/204 to 204/204.
- Several ordinary cases remained neutral, including Scenario 2 at 204/204 in
  both modes.
- Near-complete regressions also occurred, including Scenario 1 at 304/306
  compared with a 306/306 standard completion.
- Severe local saturation remained possible. Scenario 4 seated 151/259 in the
  paced mode while 77 passengers remained outside and a 30-passenger dependency
  chain occupied the right aisle.
- The most severe recorded paced case ended at 169/324, a paired loss of 155
  passengers with 121 additional passengers outside.
- Rear-boundary locks, linked row-event chains and entry saturation with internal
  chains all remained present.

Principal Findings
------------------
1. Fixed micro-batch pacing partially recovered behaviour lost under Experiment
   18's uninterrupted fine geographic zones.
2. The recovery validates admission-wave size as a causal variable: geography
   alone was not responsible for every deterioration observed in Experiment 18.
3. Pacing was not sufficient as a universal remedy. Some scenarios still formed
   long, spatially concentrated aisle chains with large outside queues.
4. The result is therefore mixed rather than negative: pacing can convert a
   stalled case to success, remain neutral in ordinary cases, or still permit a
   severe saturation event depending on topology and seeded composition.
5. The evidence points to admission-wave shape, not merely batch size. Sharp
   boundaries between successive zones can still create concentrated local
   transitions even when each zone is internally paced.
6. Normal numerical entry headway remained active throughout. The observed
   changes came from eligibility order and pacing, not simultaneous insertion or
   movement-engine modification.

Discussion Incorporated Before Experiment 20
---------------------------------------------
Experiment 20 should retain both six-zone geography and fixed micro-batches but
change the shape of the transition between adjacent zones. Rather than waiting
for one zone's full batch sequence to finish before beginning the next, a small
precomputed overlap should introduce the first micro-batch of the next adjacent
zone before the final micro-batch of the current zone.

This is not runtime adaptation. The complete sequence is created before the
simulation begins, seeded order remains preserved inside every zone, and no
passenger is redirected after entry. The purpose is to smooth abrupt zone
boundaries while preserving deterministic reproducibility.

Experiment 19 Conclusion
------------------------
Experiment 19 confirms that deterministic pacing can recover some completions
and reduce some dependency structures, but fixed pauses alone do not eliminate
entry saturation. The next question is whether a smoother deterministic
inter-zone transition can retain spatial organisation without creating sharp
local release fronts.


======================================================================
EXPERIMENT 20 - OVERLAPPING GEOGRAPHIC MICRO-BATCH ADMISSION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen. The new
logic operates only before execution by constructing the eligibility queue and
release ticks. Once admitted, every passenger follows the same one-tile
synchronous movement, row limit, blocker-yield, seat-event, reservation,
dependency-region lifecycle and hierarchy rules used in the preceding
experiments.

Planned Purpose
---------------
Test whether a small fixed overlap between adjacent geographic zones can smooth
the admission profile and reduce the severe entry-saturation cases that remained
under Experiment 19's strictly sequential micro-batch zones.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. OVERLAPPING GEOGRAPHIC MICRO-BATCH DATASET.

Overlapping Admission Contract
------------------------------
- The cabin remains divided into six longitudinal destination-row zones.
- Overall progression remains rear to front.
- Seeded relative order is preserved inside each zone.
- Each zone is divided before execution into deterministic micro-batches.
- The first micro-batch of the next adjacent zone is inserted before the final
  micro-batch of the current zone.
- The overlap sequence and release ticks are fixed before execution.
- Normal deterministic numerical entry headway remains active.
- The entry tile must still be empty before insertion.
- Runtime congestion, dependency regions and seat events do not alter release.
- No passenger is reordered inside a zone, redirected after entry or permitted
  to switch aisle.

Independent Variable
--------------------
    FIXED ADJACENT-ZONE MICRO-BATCH OVERLAP

Primary Evaluation Order
------------------------
1. Correct deterministic overlap construction and replay.
2. Successful completions and conversions relative to the standard mode.
3. Outside-queue accumulation.
4. Longest dependency-chain length and critical-tree depth.
5. Entry-saturation, linked-row-event and rear-boundary residual families.
6. Confirmed-region concurrency and hierarchy competitions.
7. Cluster selections and executable starts.
8. Waiting cost and elapsed ticks.

Experimental Hypothesis
-----------------------
If sharp transitions between sequential geographic zones contributed to the
remaining Experiment 19 saturation cases, a small precomputed overlap should
smooth local admission density, reduce abrupt destination-band fronts and lower
the frequency or severity of entry-saturation chains.

A negative result would show that deterministic overlap merely redistributes the
same spatial concentration, and that future work should change passenger
composition inside micro-batches rather than only their temporal arrangement.

Interpretation Safeguards
-------------------------
- The overlap is fixed and non-adaptive.
- Improved completion must be checked against waiting and elapsed-time cost.
- Lower outside waiting alone does not prove reduced internal congestion.
- Rear-boundary locks are geometric constraints and may remain unaffected.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 19
======================================================================
1. Timing-only windows were neutral in Experiment 16.
2. Broad geographic cohorts changed dependency topology in Experiment 17.
3. Uninterrupted fine zones frequently intensified saturation in Experiment 18.
4. Fixed micro-batch pacing partially recovered completion in Experiment 19.
5. Pacing did not prevent all severe entry-saturation chains.
6. Admission-wave size and transition shape are now separate causal variables.
7. Normal entry headway remains active in every mode.
8. Experiment 20 introduces only a fixed adjacent-zone overlap.
9. Runtime movement and dependency architecture remain frozen.
10. The next question is whether smoother zone boundaries outperform strictly
    sequential geographic micro-batches.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v20 - EXPERIMENT 20 DESIGN
======================================================================



======================================================================
EXPERIMENT 20 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic run executed all 30 paired scenarios (60 total
executions). The standard and overlapping modes retained identical cabin
geometry, occupancy, scenario seed, passenger manifest, serving-aisle
assignment, numerical entry headway, synchronous movement, seat-event rules,
dependency-region lifecycle and hierarchy controls. Only the precomputed
adjacent-zone overlap schedule differed.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers across the paired set.
- The overlapping mode seated 6,332 of 7,996 passengers, an aggregate loss of
  1,595 seated passengers.
- The overlapping mode was worse in 16 of 30 paired scenarios, improved three,
  and was equal in eleven.
- No incomplete standard case was converted to full success.
- Scenario 14 produced the largest paired loss: 36/259 compared with 259/259,
  with 160 additional passengers remaining outside.
- Scenario 9 seated only 33/216 under overlap, a loss of 183 passengers.
- Entry saturation with an internal chain was the dominant overlap residual
  family, appearing in 12 paired scenarios.
- Some neutral and modestly improved cases remained, including Scenario 22 at
  285/288 compared with 271/288, but these did not offset the dominant decline.

Principal Findings
------------------
1. Fixed adjacent-zone overlap did not smooth the admission profile reliably.
   It frequently allowed neighbouring geographic waves to coexist in the same
   aisle before the earlier wave had dispersed.
2. The resulting wave merger recreated the conditions of a large concentrated
   release: long occupied aisle chains, persistent yield-tile occupation and
   large outside queues.
3. Small deterministic waves are beneficial only while they remain sufficiently
   independent. Premature overlap removes that protection.
4. The result is a clear negative finding for fixed temporal overlap rather than
   for geographic organisation itself.
5. Normal numerical entry headway remained active throughout. The regression
   arose from eligibility-wave interaction, not simultaneous insertion or any
   movement-engine change.
6. Rear-boundary locks remained a separate geometric limitation and were not
   expected to disappear under admission overlap.

Discussion Incorporated Before Experiment 21
---------------------------------------------
The next experiment should stop using a fixed clock to decide when the next
geographic zone begins. Instead, it should apply a declared deterministic
completion milestone. The next adjacent zone becomes eligible only after a
fixed proportion of the current zone has entered the aircraft.

This is state-triggered but not optimisation-driven. The threshold is fixed
before execution, does not inspect congestion scores, dependency regions or seat
events, and does not choose among alternative policies at runtime. Its purpose
is to preserve wave independence while avoiding arbitrary timing assumptions.

Experiment 20 Conclusion
------------------------
Experiment 20 demonstrates that premature deterministic overlap is strongly
counterproductive. Admission-wave independence is more important than merely
smoothing a visual timing boundary. The next question is whether a fixed
completion-triggered release rule can preserve that independence while allowing
controlled progression between geographic zones.


======================================================================
EXPERIMENT 21 - COMPLETION-TRIGGERED GEOGRAPHIC ADVANCEMENT
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen. The new
logic governs only which preassigned passenger may become eligible at the
aircraft entrance. Once admitted, every passenger follows the same one-tile
synchronous movement, row limit, blocker-yield, seat-event, middle-bank
reservation, dependency-region lifecycle and hierarchy rules used throughout
the programme.

Planned Purpose
---------------
Test whether geographic admission waves can remain sufficiently independent
when progression to the next zone is controlled by a fixed completion milestone
rather than by fixed release ticks or premature overlap.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. COMPLETION-TRIGGERED GEOGRAPHIC DATASET.

Completion-Triggered Admission Contract
---------------------------------------
- The cabin remains divided into six longitudinal destination-row zones.
- Overall progression remains rear to front.
- Seeded relative order is preserved inside every zone.
- Each aisle tracks its own zone progression independently.
- The next frontward non-empty zone is unlocked only after at least 80 percent
  of the currently active zone for that aisle has entered the aircraft.
- Once more than one zone is unlocked, passenger selection follows a fixed
  deterministic alternation across unlocked non-empty zones.
- The 80 percent threshold and alternation rule are declared before execution.
- Normal deterministic numerical entry headway remains active.
- The entry tile must still be empty before insertion.
- Congestion scores, dependency regions, seat events and hierarchy decisions do
  not alter the threshold or select a different policy.
- No passenger changes seat, serving aisle or destination after entry.

Independent Variable
--------------------
    FIXED ZONE-COMPLETION MILESTONE FOR FRONTWARD ADVANCEMENT

Primary Evaluation Order
------------------------
1. Correct deterministic threshold and replay behaviour.
2. Successful completions and converted-success cases.
3. Aggregate seated passengers and outside-queue accumulation.
4. Longest dependency-chain length and critical-tree depth.
5. Entry-saturation and linked-row-event residual families.
6. Rear-boundary residuals, which remain geometrically distinct.
7. Confirmed-region concurrency, hierarchy competitions and deferrals.
8. Waiting cost and elapsed ticks.

Experimental Hypothesis
-----------------------
If Experiment 20 failed because the next wave was released before the current
wave had dispersed sufficiently, an 80 percent completion milestone should
preserve more wave independence, reduce simultaneous geographic concentration
and recover completion relative to fixed overlap.

A negative result would show that a simple completion proportion is still too
coarse and that later work should examine deterministic composition inside each
release wave rather than only the timing or trigger of zone advancement.

Interpretation Safeguards
-------------------------
- The milestone is fixed and non-optimising.
- Runtime state is used only to evaluate the declared completion proportion.
- No congestion-sensitive or predictive admission decision is introduced.
- Improved completion must be checked against additional waiting time.
- A lower outside queue alone does not establish reduced internal dependency.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 20
======================================================================
1. Timing-only windows were neutral in Experiment 16.
2. Broad geographic cohorts changed dependency topology in Experiment 17.
3. Uninterrupted fine zones frequently intensified saturation in Experiment 18.
4. Fixed micro-batch pacing partially recovered completion in Experiment 19.
5. Premature adjacent-zone overlap caused a strong aggregate regression in
   Experiment 20.
6. Small admission waves help only while they remain sufficiently independent.
7. Normal entry headway remains active in every mode.
8. Experiment 21 replaces fixed overlap timing with an 80 percent completion
   milestone.
9. The trigger is deterministic and non-optimising; movement architecture stays
   frozen.
10. The next question is whether completion-triggered advancement preserves wave
    independence more reliably than fixed temporal overlap.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v21 - EXPERIMENT 21 DESIGN
======================================================================


======================================================================
EXPERIMENT 21 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic run executed all 30 paired scenarios (60 total
executions). Standard and completion-triggered modes retained identical cabin
geometry, occupancy, scenario seed, passenger manifest, serving-aisle
assignment, numerical entry headway, synchronous movement, seat-event rules,
dependency-region lifecycle and hierarchy controls. Only the declared
completion-triggered geographic admission policy differed.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers.
- Completion-triggered mode seated 5,638 of 7,996 passengers, an aggregate loss
  of 2,289 seated passengers.
- Completion-triggered mode was worse in 17 of 30 pairs, improved four and was
  equal in nine.
- Four incomplete standard cases were converted to success, showing that the
  mechanism was not uniformly defective.
- Scenario 26 produced the largest loss: 31/353 compared with 353/353, with 251
  additional passengers outside and a 35-passenger dependency chain.
- The largest improvement was Scenario 22 at +17 seated passengers.
- Entry saturation with an internal chain was the completion-triggered residual
  family in 15 scenarios; 13 scenarios completed fully.
- Scenario 29 converted a 257/259 rear-boundary lock into 259/259 success.

Principal Findings
------------------
1. An 80 percent entry-completion threshold did not reliably preserve wave
   independence.
2. The threshold measured how much of a zone had entered, but not whether those
   passengers had dispersed spatially inside the aisle.
3. Consequently, a newly unlocked frontward zone could still merge with a dense
   unresolved rearward wave and create long entry-saturation chains.
4. The severe failures again showed dominant NEXT_AISLE_TILE_OCCUPIED evidence,
   large outside queues and linked row-event dependencies.
5. Neutral and converted-success cases confirm that completion-triggering is
   scenario dependent rather than universally invalid.
6. Normal numerical entry headway remained active. The negative result arose
   from admission-wave composition and internal spatial interaction, not from a
   movement-engine change.

Discussion Incorporated Before Experiment 22
---------------------------------------------
Experiments 18-21 have now tested increasingly refined timing, pacing, overlap
and completion-trigger rules. The repeated residual family indicates that the
next controlled variable should be passenger composition inside the active
wave, rather than another variation of when a zone becomes eligible.

Experiment 22 therefore returns to the broad rear/middle/front cohorts associated
with the constructive result in Experiment 17. Within each active cohort, it
interleaves eligible passengers from opposite serving aisles in a fixed
left/right sequence where possible. This changes spatial composition while
retaining deterministic cohort progression and ordinary entry headway.

Experiment 21 Conclusion
------------------------
Completion percentage is an incomplete proxy for wave dispersal. A zone may be
80 percent admitted while its passengers still occupy a long, tightly coupled
aisle chain. The next experiment therefore isolates opposite-aisle composition
inside broad cohorts rather than refining the release clock again.


======================================================================
EXPERIMENT 22 - COARSE-COHORT OPPOSITE-AISLE INTERLEAVING
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen. The new
logic affects only the order in which preassigned passengers become eligible at
the entrance. Once admitted, passengers use the same one-tile synchronous
movement, row limit, blocker-yield, seat-event, middle-bank reservation,
dependency-region lifecycle and hierarchy rules.

Planned Purpose
---------------
Test whether broad geographic cohort benefits can be retained while reducing
single-aisle concentration through deterministic opposite-aisle interleaving.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. COARSE-COHORT OPPOSITE-AISLE INTERLEAVED DATASET.

Admission Contract
------------------
- The cabin is divided into three broad destination-row cohorts: rear, middle
  and front.
- Cohorts progress strictly rear to middle to front.
- A cohort advances only when both serving-aisle queues for that cohort are
  empty.
- Seeded relative order is preserved within each aisle/cohort queue.
- Where both aisles have eligible passengers, successful admissions alternate
  left then right in a fixed deterministic sequence.
- If the preferred aisle is temporarily unable to admit because of its entry
  cooldown or occupied entry tile, the other eligible aisle may proceed.
- The next successful admission preference is always the opposite of the side
  most recently admitted.
- Normal deterministic numerical entry headway remains active independently for
  each aisle.
- Entry tiles must remain empty before insertion.
- No congestion score, dependency region, seat event or hierarchy decision
  changes the cohort or aisle preference.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    OPPOSITE-AISLE COMPOSITION INSIDE BROAD GEOGRAPHIC COHORTS

Primary Evaluation Order
------------------------
1. Deterministic replay and correct rear/middle/front progression.
2. Aggregate seated passengers and converted-success cases.
3. Left/right admission balance within active cohorts.
4. Outside-queue accumulation and entry saturation.
5. Longest dependency-chain length and critical-tree depth.
6. Linked row-event and rear-boundary residual families.
7. Confirmed-region concurrency, hierarchy competitions and deferrals.
8. Waiting cost and elapsed ticks.

Experimental Hypothesis
-----------------------
If the repeated failures in Experiments 18-21 were amplified by one-sided
admission-wave composition, deterministic opposite-aisle interleaving inside
broad cohorts should reduce sustained concentration in a single aisle while
preserving the geographic structure that proved useful in Experiment 17.

A negative result would show that aisle alternation alone cannot control local
row-event dependencies and that the broad-cohort benefit depends on more than
left/right composition.

Interpretation Safeguards
-------------------------
- Interleaving is fixed and deterministic, not congestion adaptive.
- Numerical headway remains active and may delay either side independently.
- A balanced admission count does not by itself prove reduced internal
  dependency.
- Rear-boundary locks remain a separate geometric limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 21
======================================================================
1. Timing-only windows were neutral in Experiment 16.
2. Broad geographic cohorts changed dependency topology in Experiment 17.
3. Fine zones frequently intensified saturation in Experiment 18.
4. Micro-batch pacing partially recovered completion in Experiment 19.
5. Adjacent-zone overlap caused strong regression in Experiment 20.
6. An 80 percent completion trigger also regressed strongly in Experiment 21.
7. Completion percentage does not measure spatial dispersal inside the aisle.
8. Entry saturation with internal chains remained the recurring failure family.
9. Experiment 22 returns to broad cohorts and changes wave composition through
   deterministic opposite-aisle interleaving.
10. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v22 - EXPERIMENT 22 DESIGN
======================================================================


======================================================================
EXPERIMENT 22 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic run executed all 30 paired scenarios (60 total
executions). Standard and opposite-aisle-interleaved modes retained identical
cabin geometry, occupancy, scenario seed, passenger manifest, serving-aisle
assignment, numerical entry headway, synchronous movement, seat-event rules,
dependency-region lifecycle and hierarchy controls. Only the declared admission
composition inside the broad rear/middle/front cohorts differed.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers.
- Coarse-cohort opposite-aisle interleaving seated 7,101 of 7,996 passengers,
  an aggregate loss of 826 seated passengers.
- Interleaving improved two pairs, was worse in 12 and was equal in 16.
- One incomplete standard case was converted to success.
- Scenario 1 improved from 221/235 to 232/235 and reduced the outside queue by
  eight passengers.
- In Scenario 1 the residual family changed from entry saturation with an
  internal chain to a rear-boundary lock, showing that interleaving can improve
  admission-wave behaviour in an individual case without resolving the final
  geometric boundary condition.
- Several already successful scenarios remained fully successful, confirming
  that the mechanism was not intrinsically invalid.

Principal Findings
------------------
1. Opposite-aisle alternation did not improve the architecture overall.
2. Aisle balance alone was not a sufficient proxy for reduced internal
   dependency.
3. Alternation weakened the local geographic cohesion that had been constructive
   in Experiment 17 and distributed seat-event activity across both aisles
   without reliably reducing row-event coupling.
4. The aggregate regression shows that broad geographic structure matters more
   than mechanically equalising left/right admissions.
5. The isolated improvements remain useful evidence: admission-wave composition
   can change the residual family even when it does not produce full completion.
6. Normal numerical entry headway remained active throughout; the result arose
   from deterministic queue composition rather than a movement-engine change.

Discussion Incorporated Before Experiment 23
---------------------------------------------
Experiments 17-22 have now tested geographic scale, timing windows, fine zones,
micro-batch pacing, overlap, completion-triggered advancement and opposite-aisle
interleaving. The evidence indicates that further variations of admission timing
or aisle alternation are unlikely to be the strongest next step.

Experiment 23 therefore retains the successful coarse rear/middle/front
geographic structure and changes only passenger composition inside each cohort.
Passengers are ordered by fixed blocker-complexity class, with simpler seat
topologies released before more complex topologies while seeded order is
preserved within each class.

Experiment 22 Conclusion
------------------------
Deterministic opposite-aisle alternation is not a general substitute for
geographic cohesion. The next controlled question is whether preserving the
cohort while ordering its internal seat topology can reduce blocker interaction
without reopening the movement architecture.


======================================================================
EXPERIMENT 23 - COARSE-COHORT BLOCKER-COMPLEXITY ORDERING
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen. The new
logic changes only the pre-execution order in which already assigned passengers
become eligible at the entrance. Once admitted, every passenger follows the same
one-tile synchronous movement, row limit, blocker-yield, seat-event,
middle-bank reservation, dependency-region lifecycle and hierarchy rules.

Planned Purpose
---------------
Test whether broad geographic cohort benefits can be retained while reducing
early blocker interaction by admitting lower-complexity seat topologies before
higher-complexity topologies inside each cohort.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. COARSE-COHORT BLOCKER-COMPLEXITY-ORDERED DATASET.

Admission Contract
------------------
- The cabin is divided into three broad destination-row cohorts: rear, middle
  and front.
- Cohorts progress strictly rear to middle to front.
- Each aisle retains its own queue and ordinary numerical entry headway.
- Passenger complexity is calculated before execution from seat depth relative
  to the assigned serving aisle.
- Complexity class 0 contains aisle-adjacent or zero-crossing seats.
- Complexity class 1 contains seats one position deeper.
- Complexity class 2 contains deeper, multiple-blocker-prone topologies.
- Inside each aisle/cohort queue, classes progress from 0 to 1 to 2.
- Seeded relative order is preserved within every complexity class.
- Entry tiles must remain empty before insertion.
- No runtime congestion score, dependency region, seat event or hierarchy
  decision changes the declared class ordering.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    BLOCKER-COMPLEXITY COMPOSITION INSIDE BROAD GEOGRAPHIC COHORTS

Primary Evaluation Order
------------------------
1. Deterministic replay and correct cohort/class ordering.
2. Aggregate seated passengers and converted-success cases.
3. Outside-queue accumulation and entry saturation.
4. Seat-event requests, failed starts and blocker interaction.
5. Longest dependency-chain length and critical-tree depth.
6. Linked row-event and rear-boundary residual families.
7. Confirmed-region concurrency, hierarchy competitions and deferrals.
8. Waiting cost and elapsed ticks.

Experimental Hypothesis
-----------------------
If broad geographic cohesion is beneficial but mixed blocker complexity causes
premature seat-event coupling, admitting simpler seat topologies before deeper
ones should allow early passengers to clear the aisle and create more usable
yield space before complex blocker transactions begin.

A negative result would show that simple-first composition merely postpones or
concentrates the difficult seat events and that blocker complexity must not be
interpreted as an independently optimisable admission variable.

Interpretation Safeguards
-------------------------
- Complexity is fixed from seat topology before movement begins.
- The classes do not predict actual blockers present at runtime.
- Seeded order is preserved within each class.
- Numerical headway remains active independently for both aisles.
- A lower early blocker count does not by itself establish improved completion.
- Rear-boundary locks remain a separate geometric limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 22
======================================================================
1. Broad geographic cohorts remain the strongest constructive admission result.
2. Fine geographic subdivision and several timing refinements regressed.
3. Micro-batch pacing produced partial recovery but did not eliminate severe
   saturation.
4. Opposite-aisle interleaving improved isolated cases but lost 826 passengers
   in aggregate.
5. Aisle balance is not equivalent to preserved geographic cohesion.
6. Experiment 23 retains broad cohorts and tests fixed blocker-complexity
   composition inside each cohort.
7. Simple, intermediate and complex classes are determined before execution.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v23 - EXPERIMENT 23 DESIGN
======================================================================


======================================================================
EXPERIMENT 23 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic run executed all 30 paired scenarios (60 total
executions). Standard and blocker-complexity-ordered modes retained identical
cabin geometry, occupancy, scenario seed, passenger manifest, serving-aisle
assignment, numerical entry headway, synchronous movement, seat-event rules,
dependency-region lifecycle and hierarchy controls. Only the fixed pre-execution
ordering of seat-topology classes inside the broad cohorts differed.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers.
- Coarse-cohort blocker-complexity ordering seated 6,466 of 7,996 passengers,
  an aggregate loss of 1,461 seated passengers.
- Complexity ordering improved three pairs, was worse in 19 and was equal in
  eight.
- No incomplete standard case was converted to success.
- Scenario 1 improved from 221/235 to 232/235 and reduced the outside queue by
  eight passengers.
- Scenario 1 changed from entry saturation with an internal chain to a
  rear-boundary lock, showing that topology ordering can improve an individual
  admission wave without generalising across the dataset.
- Scenario 2 regressed from 216/216 to 213/216 and ended in a linked row-event
  chain.
- Scenario 3 regressed from 306/306 to 130/306, with 141 passengers still
  outside, 28 moving in the aisle and seven waiting at rows.

Principal Findings
------------------
1. Fixed simple-to-complex ordering was a clear aggregate regression.
2. Seat-topology complexity is not an independently optimisable admission
   variable in this architecture.
3. Grouping all simple passengers before intermediate and complex passengers
   weakened the natural seeded spatial relationships inside each broad cohort.
4. Difficult seat events were postponed and then concentrated rather than
   eliminated.
5. The strongest failures returned to entry saturation, long aisle queues and
   linked row-event chains.
6. The isolated Scenario 1 improvement remains valid evidence that composition
   can alter the residual family, but it does not support general adoption.
7. Normal numerical headway remained active and the movement engine remained
   frozen throughout.

Discussion Incorporated Before Experiment 24
---------------------------------------------
Experiments 17-23 have now tested equal broad cohorts, finer zones, fixed timing,
micro-batches, overlapping waves, completion-triggered advancement, opposite-
aisle interleaving and blocker-complexity ordering. The repeated regressions show
that preserving broad geographic cohesion and seeded internal order is more
important than imposing additional ordering classes.

Experiment 24 therefore restores seeded order inside each aisle/cohort and
changes only the positions of the two broad longitudinal boundaries. Instead of
equal thirds, deterministic geometry-weighted boundaries allocate approximately
25 percent of rows to the front cohort, 35 percent to the middle cohort and 40
percent to the rear cohort. This tests boundary placement without changing the
movement architecture or introducing runtime adaptation.

Experiment 23 Conclusion
------------------------
Simple-first blocker ordering does not generalise. It can postpone and
concentrate complex seat events while breaking useful spatial relationships.
The next controlled variable is cohort boundary geometry, with passenger order
otherwise restored.


======================================================================
EXPERIMENT 24 - GEOMETRY-WEIGHTED COARSE-COHORT BOUNDARIES
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen. The new
logic changes only the pre-execution row boundaries used to place passengers
into rear, middle and front admission cohorts. Once admitted, every passenger
uses the same one-tile synchronous movement, row limit, blocker-yield,
seat-event, middle-bank reservation, dependency-region lifecycle and hierarchy
rules.

Planned Purpose
---------------
Test whether the constructive broad-cohort behaviour can be retained or improved
by moving the cohort boundaries away from equal thirds while preserving seeded
passenger order inside each cohort.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. GEOMETRY-WEIGHTED COARSE-COHORT DATASET.

Admission Contract
------------------
- The cabin retains three broad destination-row cohorts.
- Cohorts progress strictly rear to middle to front.
- The front cohort contains approximately 25 percent of cabin rows.
- The middle cohort extends to approximately 60 percent of cabin rows, giving
  it approximately 35 percent of rows.
- The rear cohort contains the remaining approximately 40 percent of rows.
- Exact integer boundaries are calculated deterministically from each cabin's
  row count before execution.
- Each aisle retains its own seeded queue inside every cohort.
- Seeded relative order is preserved; no complexity sort or aisle alternation
  is applied.
- Ordinary numerical entry headway remains active independently for each aisle.
- Entry tiles must be empty before insertion.
- No runtime congestion, completion, dependency or hierarchy signal changes a
  boundary or passenger's eligibility.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    PRECOMPUTED LONGITUDINAL COHORT-BOUNDARY GEOMETRY

Primary Evaluation Order
------------------------
1. Deterministic replay and correct geometry-weighted cohort membership.
2. Aggregate seated passengers and converted-success cases.
3. Comparison with the equal-third broad-cohort result.
4. Outside-queue accumulation and entry saturation.
5. Longest dependency-chain length and critical-tree depth.
6. Rear-boundary, linked row-event and internal-chain residual families.
7. Confirmed-region concurrency, hierarchy competitions and deferrals.
8. Waiting cost and elapsed ticks.

Experimental Hypothesis
-----------------------
If equal thirds place too much pressure on a narrow rear operating region,
allocating a larger share of cabin rows to the rear cohort may preserve local
geographic cohesion while reducing abrupt density changes at the rear/middle
boundary.

A negative result would show that broad-cohort benefit is not improved merely by
moving static boundaries and that the equal-third division is already an
adequate coarse representation.

Interpretation Safeguards
-------------------------
- Boundaries are derived only from cabin row count before execution.
- The 25/35/40 proportions are a controlled research partition, not an airline
  recommendation.
- Seeded passenger order is restored and preserved inside each cohort.
- Numerical headway remains active independently for both aisles.
- Rear-boundary locks remain a separate physical limitation.
- No runtime adaptation or congestion optimisation is introduced.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 23
======================================================================
1. Broad geographic cohorts remain the strongest constructive admission result.
2. Fine subdivision and multiple timing refinements produced major regressions.
3. Micro-batch pacing gave partial recovery but did not generalise fully.
4. Opposite-aisle interleaving and blocker-complexity ordering both weakened
   aggregate performance.
5. Experiment 23 lost 1,461 seated passengers relative to standard mode.
6. Seat complexity cannot be treated as an independent admission priority.
7. Experiment 24 restores seeded internal order and changes only static cohort
   boundary geometry.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v24 - EXPERIMENT 24 DESIGN
======================================================================


======================================================================
EXPERIMENT 24 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Standard and geometry-weighted modes retained identical
cabin geometry, occupancy, scenario seed, manifest, aisle assignment, numerical
headway, movement rules, seat-event rules and dependency architecture. Only the
two precomputed broad cohort boundaries differed.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers.
- Geometry-weighted cohort boundaries seated 7,677 of 7,996 passengers.
- The aggregate change was a loss of 250 seated passengers.
- Geometry weighting improved four pairs, was worse in nine and was equal in
  17.
- Three incomplete standard cases were converted to complete success.
- Pair 1 improved from 221/235 to 235/235.
- Pair 12 improved from 225/235 to 235/235.
- Pair 5 improved from 239/245 to 245/245.
- Pair 25 produced the largest regression: 276/282 fell to 137/282, with an
  outside-queue increase of 114 passengers.
- Pair 11 fell from 270/282 to 171/282.
- The longest severe geometry-weighted chain reached 31 passengers in Pair 25.

Principal Findings
------------------
1. Moving the static broad boundaries produced genuine converted-success cases.
2. The result did not generalise: aggregate completion still declined.
3. Geometry-weighted partitioning is therefore a mixed structural result rather
   than a major breakthrough.
4. Larger rear allocation can remove some entry-saturation stalls, but it can
   also concentrate passengers into long rearward dependency waves.
5. Broad geographic cohesion remains valuable, but boundary placement alone is
   not sufficient to control local interaction.
6. Normal entry headway remained active and the movement engine stayed frozen.

Discussion Incorporated Before Experiment 25
---------------------------------------------
The previous discussion correctly identified structural partitioning as more
promising than further global reordering, but the full Experiment 24 dataset
shows both benefits and regressions. The next experiment therefore retains the
same geometry-weighted broad boundaries and seeded order, while preventing
adjacent local regions inside a broad cohort from admitting simultaneously.

Experiment 24 Conclusion
------------------------
Static boundary geometry can convert selected failures to success, but the
25/35/40 partition lost 250 passengers overall. The next controlled variable is
local activation hierarchy, not another change to global passenger order.


======================================================================
EXPERIMENT 25 - HIERARCHICAL LOCAL-REGION ACTIVATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic one-dimensional people-movement backbone remains
frozen. Experiment 25 changes only which precomputed local admission queue may
supply the next passenger. Once admitted, every passenger follows the same
one-tile synchronous movement, row limit, blocker-yield, seat-event,
middle-bank reservation, dependency-region lifecycle and hierarchy rules.

Planned Purpose
---------------
Test whether the mixed Experiment 24 result can be improved by retaining its
geometry-weighted broad cohorts while preventing adjacent local admission waves
from becoming active together.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. HIERARCHICAL LOCAL-REGION DATASET.

Admission Contract
------------------
- The 25/35/40 front/middle/rear broad boundaries from Experiment 24 are
  retained.
- Each broad cohort is divided deterministically into a frontward and rearward
  local region.
- The rearward local region activates first.
- Only after both aisle queues for that local region are empty does the
  frontward local region activate.
- Only after both local regions are empty does progression move to the next
  broad cohort.
- Left and right aisles share the same active cohort and local-region level.
- Seeded relative order is preserved inside each aisle/local-region queue.
- Ordinary numerical entry headway remains active independently in each aisle.
- Entry tiles must be empty before insertion.
- Runtime congestion, dependency, progress and hierarchy evidence cannot alter
  region membership or activation order.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    PRECOMPUTED HIERARCHICAL LOCAL-REGION ACTIVATION

Primary Evaluation Order
------------------------
1. Deterministic replay and correct local-region membership.
2. Aggregate seated passengers and converted-success cases.
3. Comparison with Experiment 24's geometry-weighted result.
4. Outside-queue accumulation and entry saturation.
5. Longest dependency-chain length and critical-tree depth.
6. Rear-boundary, linked row-event and internal-chain residual families.
7. Confirmed-region concurrency, hierarchy competitions and deferrals.
8. Waiting cost and elapsed ticks.

Experimental Hypothesis
-----------------------
If Experiment 24's regressions were caused by adjacent local waves interacting
inside the same broad cohort, sequential local activation should preserve broad
geographic cohesion while reducing simultaneous local dependency formation.

A negative result would show that additional local hierarchy simply recreates
fine-zone fragmentation and that the stronger Experiment 17 behaviour depends
on allowing the broad cohort to remain internally unconstrained.

Interpretation Safeguards
-------------------------
- All local boundaries and activation order are fixed before execution.
- No runtime optimisation or congestion-sensitive release is introduced.
- Seeded relative order is preserved within each local queue.
- Numerical headway remains active independently for both aisles.
- Rear-boundary locks remain a separate physical limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 24
======================================================================
1. Broad geographic cohorts remain the strongest constructive admission result.
2. Geometry-weighted boundaries converted three failures to success.
3. The same geometry-weighted design lost 250 passengers in aggregate.
4. Static boundary placement is therefore influential but not sufficient.
5. Severe regressions remained associated with entry saturation and long
   rearward dependency chains.
6. Experiment 25 retains Experiment 24 geometry and seeded order.
7. Experiment 25 changes only local activation hierarchy inside each broad
   cohort.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v25 - EXPERIMENT 25 DESIGN
======================================================================


======================================================================
EXPERIMENT 25 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Both modes retained the same manifest, cabin geometry,
occupancy, numerical headway, movement rules, blocker-yield rules, seat-event
rules, dependency architecture and deterministic replay seed. Only the
precomputed local-region activation hierarchy differed.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers.
- Hierarchical local-region mode seated 5,913 of 7,996 passengers.
- The aggregate change was a loss of 2,014 seated passengers.
- Hierarchical activation improved four pairs, was worse in 15 and was equal in
  11.
- No incomplete standard case was converted to complete success.
- Pair 22 produced the largest improvement: 271/288 increased to 285/288.
- Pair 1 improved from 221/235 to 232/235 and changed the residual from
  ENTRY_SATURATION_WITH_INTERNAL_CHAIN to REAR_BOUNDARY_LOCK.
- Pair 26 produced the largest regression: 353/353 fell to 40/353, with 242
  additional passengers remaining outside.
- Pair 27 fell from 353/353 to 85/353.
- Pair 7 fell from 274/274 to 44/274.
- The longest observed dependency chain reached 35 passengers.

Principal Findings
------------------
1. The promising early-pair pattern did not generalise across the full run.
2. Local hierarchy reduced internal saturation in selected cases, but it also
   created severe admission starvation and long dependency chains elsewhere.
3. Empty-queue advancement does not mean the active local wave has dispersed.
4. Additional local fragmentation can recreate the same fine-zone problem seen
   in Experiment 18.
5. Broad geographic cohesion remains beneficial, but strict local sequencing
   inside each broad cohort is not sufficient by itself.
6. Ordinary numerical entry headway remained active in every local queue.
7. The movement and dependency architecture remained frozen.

Discussion Incorporated Before Experiment 26
---------------------------------------------
The earlier discussion described Experiment 25 as promising based on the first
few pairs. The full 30-pair result changes that interpretation. Experiment 25 is
a strong negative result overall, despite several useful local improvements.

The next controlled test retains the same hierarchy but changes only the
advancement condition. Instead of advancing immediately when both admission
queues are empty, Experiment 26 requires a fixed proportion of the current
local region to be seated first.

Experiment 25 Conclusion
------------------------
Hierarchical local-region activation did not improve the dataset overall.
Advancing on queue exhaustion alone allowed a new local wave to begin while the
previous wave was still spatially active. Experiment 26 therefore tests a fixed
local seating gate without changing region geometry or passenger order.


======================================================================
EXPERIMENT 26 - HIERARCHICAL LOCAL COMPLETION GATES
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen.
Experiment 26 changes only the condition that permits the next precomputed local
admission queue to become active. Once admitted, passengers use the same
one-tile synchronous movement, row limit, blocker-yield, seat-event,
middle-bank, dependency-region lifecycle and hierarchy rules.

Planned Purpose
---------------
Test whether Experiment 25 failed because queue exhaustion was treated as local
completion even though many passengers from the active region were still moving
or waiting at rows.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. HIERARCHICAL LOCAL COMPLETION GATE DATASET.

Admission Contract
------------------
- Experiment 25's 25/35/40 broad cohort geometry is retained.
- Each broad cohort retains its rearward and frontward local regions.
- Rearward local regions still precede frontward local regions.
- The next local region activates only when:
  1. both admission queues for the active local region are empty; and
  2. at least 80% of all passengers assigned to that local region are seated.
- Empty local regions advance immediately.
- The 80% threshold is fixed before execution and cannot change at runtime.
- Left and right aisles share the same active cohort and local-region level.
- Seeded relative order is preserved within each aisle/local-region queue.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- Runtime congestion cannot change region membership, activation order or the
  threshold.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    FIXED 80% LOCAL SEATING GATE BEFORE REGION ADVANCEMENT

Primary Evaluation Order
------------------------
1. Deterministic replay and gate correctness.
2. Aggregate seated passengers and converted-success cases.
3. Comparison with Experiment 25.
4. Outside-queue accumulation and admission starvation.
5. Longest dependency-chain length and critical-tree depth.
6. Rear-boundary, linked row-event and entry-saturation residual families.
7. Waiting cost and elapsed ticks.
8. Confirmation, hierarchy and suppression evidence.

Experimental Hypothesis
-----------------------
If Experiment 25's regressions were caused by advancing while the previous local
wave was still spatially active, requiring 80% of that region to be seated
should improve wave separation.

A negative result would show that completion gating compounds local
fragmentation and that the broad-cohort architecture should not be subdivided
further.

Interpretation Safeguards
-------------------------
- The 80% gate is fixed and identical across every scenario.
- No runtime optimisation selects a different threshold.
- Local-region membership and activation order remain precomputed.
- Numerical headway remains active independently for both aisles.
- Rear-boundary locks remain a separate physical limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 25
======================================================================
1. Experiment 25 lost 2,014 seated passengers overall.
2. Four local improvements did not offset 15 worse pairs.
3. No stalled standard case converted to complete success.
4. Queue exhaustion was not a reliable proxy for local wave completion.
5. Strict local hierarchy can recreate severe fine-zone fragmentation.
6. Experiment 26 retains all geometry and seeded ordering.
7. Experiment 26 changes only the local advancement condition.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v26 - EXPERIMENT 26 DESIGN
======================================================================


======================================================================
EXPERIMENT 26 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Both modes retained the same manifest, cabin geometry,
occupancy, numerical entry headway, movement rules, blocker-yield rules,
seat-event rules, dependency architecture and replay seed. Only the fixed 80%
local seating gate differed from Experiment 25.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers.
- Hierarchical local-completion-gate mode seated 5,668 of 7,996 passengers.
- The aggregate change was a loss of 2,259 seated passengers.
- The completion-gate mode improved four pairs, was worse in 14 and was equal
  in 12.
- Two incomplete standard cases were converted to complete success.
- Pair 1 improved from 221/235 to 235/235.
- Pair 11 improved from 270/282 to 282/282.
- Pair 22 improved from 271/288 to 285/288 but remained incomplete.
- Pair 26 produced the largest regression: 353/353 fell to 38/353.
- Pair 27 fell from 353/353 to 85/353.
- Pair 6 fell from 324/324 to 84/324.
- Pair 25 fell from 276/282 to 36/282.
- The longest observed dependency chain reached 34 passengers.

Principal Findings
------------------
1. The 80% gate produced two genuine converted-success cases.
2. Those successes did not generalise across the complete dataset.
3. The gate amplified admission starvation in many scenarios because later
   regions remained locked behind an unresolved active region.
4. A seating percentage can become a hard barrier when the remaining active
   passengers are themselves trapped in a dependency chain.
5. The result is more polarised than Experiment 25: selected cases improve
   strongly, while several complete standard runs collapse severely.
6. Local completion gating therefore compounds fragmentation rather than
   reliably protecting wave independence.
7. Normal numerical entry headway remained active throughout.
8. The movement and dependency architecture remained frozen.

Discussion Incorporated Before Experiment 27
---------------------------------------------
The earlier discussion correctly identified Pair 1 as an important success, but
the full run shows that Experiment 26 is a strong negative result overall. The
80% gate solved selected entry-saturation cases while creating severe local
admission starvation elsewhere.

The local-fragmentation branch has now tested queue exhaustion and fixed seating
completion. Neither generalised. Experiment 27 therefore returns to the best
constructive architecture already observed: three broad geographic cohorts with
no internal subdivision or completion gate.

Experiment 26 Conclusion
------------------------
A fixed 80% local seating gate is not a robust improvement. It can convert
specific failures to success, but it also locks most of the cabin behind a
small unresolved local region. Further threshold tuning would risk becoming a
parameter-search exercise rather than a new architectural test.


======================================================================
EXPERIMENT 27 - CONFIRMATORY COARSE-COHORT REPLICATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen.
Experiment 27 changes only the precomputed admission sequence by restoring three
broad rear, middle and front geographic cohorts. Once admitted, every passenger
uses the same one-tile synchronous movement, row boundary, blocker-yield,
seat-event, middle-bank reservation, dependency-region lifecycle and hierarchy
rules.

Planned Purpose
---------------
Confirm whether the strongest constructive admission-policy result remains
reproducible under the current frozen architecture after the negative
Experiments 18-26 branch.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. CONFIRMATORY COARSE-COHORT DATASET.

Admission Contract
------------------
- The cabin is divided deterministically into equal front, middle and rear row
  thirds.
- Rear cohort admission completes before middle cohort admission begins.
- Middle cohort admission completes before front cohort admission begins.
- There are no local subregions.
- There is no completion percentage, overlap, micro-batching, complexity
  ordering or runtime-adaptive release.
- Seeded relative order is preserved inside each aisle/cohort queue.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- Entry tiles must be empty before insertion.
- Runtime congestion cannot change cohort membership or progression order.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    CONFIRMATORY THREE-COHORT REAR-TO-FRONT ADMISSION

Primary Evaluation Order
------------------------
1. Deterministic replay and correct cohort membership.
2. Aggregate seated passengers and converted-success cases.
3. Reproducibility of the earlier broad-cohort finding.
4. Outside-queue accumulation and entry saturation.
5. Dependency-chain length and critical-tree depth.
6. Rear-boundary, linked row-event and internal-chain residual families.
7. Waiting cost and elapsed ticks.
8. Confirmation, hierarchy and suppression evidence.

Experimental Hypothesis
-----------------------
If broad geographic cohesion was the genuinely constructive property identified
earlier, restoring equal-thirds cohorts without local fragmentation should
outperform the local-region and completion-gate variants.

A negative result would indicate that the earlier broad-cohort advantage was
not stable under the current frozen architecture.

Interpretation Safeguards
-------------------------
- Experiment 27 is a confirmatory replication, not a new optimisation layer.
- Equal cohort boundaries are calculated before execution.
- No local hierarchy or percentage gate remains.
- Numerical entry headway remains active independently for both aisles.
- Rear-boundary locks remain a separate physical limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 26
======================================================================
1. Experiment 26 lost 2,259 seated passengers overall.
2. Two converted-success cases did not offset 14 worse pairs.
3. The fixed 80% gate can become a hard admission barrier.
4. Queue exhaustion and seating percentage both failed as robust local
   advancement criteria.
5. Further local-threshold tuning is not justified at this stage.
6. Experiment 27 closes the local-fragmentation branch.
7. Experiment 27 restores equal-thirds rear-to-front coarse cohorts.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v27 - EXPERIMENT 27 DESIGN
======================================================================


======================================================================
EXPERIMENT 27 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Standard and confirmatory modes shared the same scenario
seed, manifest, cabin geometry, occupancy, numerical headway, seat assignments,
aisle assignments, movement rules and dependency architecture. Only the
equal-thirds rear-to-middle-to-front admission policy differed.

Observed Results
----------------
- Standard mode seated 7,927 of 7,996 passengers.
- Confirmatory coarse-cohort mode seated 7,301 of 7,996 passengers.
- The aggregate change was a loss of 626 seated passengers.
- Coarse cohorts improved two pairs, were worse in 12 and were equal in 16.
- One incomplete standard case converted to complete success.
- Pair 1 improved from 221/235 to 232/235.
- Pair 5 improved from 239/245 to 245/245 and converted to success.
- Pair 11 produced a major regression: 270/282 fell to 131/282.
- Pair 25 produced the largest regression: 276/282 fell to 122/282.
- Pair 28 fell from 353/353 to 334/353.
- The longest observed dependency chain reached 35 passengers.

Principal Findings
------------------
1. The broad-cohort benefit was reproduced in selected scenarios.
2. The result did not generalise across the full 30-pair dataset.
3. Equal-thirds cohorts were substantially more stable than the local-region
   and completion-gate variants, but still lost 626 passengers overall.
4. Broad geographic cohesion can remove entry saturation and convert a failure
   to success, as shown in Pair 5.
5. The same strict cohort progression can also concentrate a rearward wave and
   produce large internal chains in other manifests.
6. The confirmatory result is therefore mixed rather than a validation of a
   universally superior policy.
7. Ordinary numerical entry headway remained active throughout.
8. The movement and dependency architecture remained frozen.

Discussion Incorporated Before Experiment 28
---------------------------------------------
The early scenarios again suggested a strong recovery, but the complete result
shows that the effect remains scenario-sensitive. Experiment 27 is better than
Experiments 25 and 26, yet it does not establish a robust aggregate advantage.

Rather than introducing another architectural variation, Experiment 28 repeats
the same broad-cohort test with a new deterministic experiment seed. This
separates architecture from scenario-set dependence.

Experiment 27 Conclusion
------------------------
Broad geographic cohorts remain the most defensible constructive admission
concept in this branch, but their aggregate benefit was not confirmed on the
current 30-pair sample. Independent-seed replication is required before any
strong general conclusion is justified.


======================================================================
EXPERIMENT 28 - INDEPENDENT-SEED COARSE-COHORT REPLICATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen.
Experiment 28 retains Experiment 27's equal-thirds rear-to-middle-to-front
admission policy and changes only the reproducible experiment seed used to
generate a new paired scenario set.

Planned Purpose
---------------
Test whether Experiment 27's mixed result is stable across an independent set
of deterministic manifests, cabin configurations, occupancy levels and
headway selections.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. INDEPENDENT-SEED COARSE-COHORT DATASET.

Experimental Seed
-----------------
    2183790101901

Admission Contract
------------------
- Equal front, middle and rear row thirds are retained.
- Rear cohort admission completes before middle cohort admission begins.
- Middle cohort admission completes before front cohort admission begins.
- Seeded relative order is preserved inside each aisle/cohort queue.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- There are no local subregions, completion gates, overlap, micro-batches,
  complexity classes or runtime-adaptive release rules.
- Within every pair, standard and cohort modes share the same new scenario
  seed, manifest, seat assignments, aisle assignments and headway.
- Runtime congestion cannot alter cohort membership or progression.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    INDEPENDENT DETERMINISTIC SCENARIO SEED

Primary Evaluation Order
------------------------
1. Deterministic replay under the new experiment seed.
2. Aggregate seated passengers.
3. Improved, worse and equal pair counts.
4. Converted-success cases.
5. Comparison with Experiment 27's aggregate direction.
6. Outside-queue accumulation and entry saturation.
7. Longest dependency-chain length and critical-tree depth.
8. Residual stall-family distribution.

Experimental Hypothesis
-----------------------
If broad geographic cohesion is a stable constructive property, the new
deterministic scenario set should again produce meaningful improvements and
converted-success cases without the catastrophic aggregate losses seen in
Experiments 25 and 26.

A materially different result would show that the observed advantage is highly
manifest-dependent and should be reported as conditional rather than general.

Interpretation Safeguards
-------------------------
- The architecture is identical to Experiment 27.
- Only the deterministic experiment seed changes.
- Standard and cohort executions remain paired within every scenario.
- Numerical entry headway remains active independently for both aisles.
- Rear-boundary locks remain a separate physical limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 27
======================================================================
1. Experiment 27 lost 626 seated passengers overall.
2. Two improvements included one converted-success case.
3. Twelve worse pairs prevented aggregate confirmation.
4. Equal-thirds cohorts remain more stable than local fragmentation.
5. The broad-cohort effect remains scenario-sensitive.
6. Experiment 28 performs an architecture-identical independent-seed test.
7. No new admission mechanism is introduced.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v28 - EXPERIMENT 28 DESIGN
======================================================================


======================================================================
EXPERIMENT 28 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Standard and cohort modes shared each new scenario seed,
manifest, cabin geometry, occupancy, numerical headway, seat assignments,
aisle assignments, movement rules and dependency architecture. Only the
equal-thirds rear-to-middle-to-front admission policy differed.

Observed Results
----------------
- Standard mode seated 8,431 of 8,599 passengers.
- Independent-seed coarse-cohort mode seated 6,812 of 8,599 passengers.
- The aggregate change was a loss of 1,619 seated passengers.
- Coarse cohorts improved two pairs, were worse in 19 and were equal in nine.
- One incomplete standard case converted to complete success.
- Pair 14 improved from 197/228 to 228/228 and converted to success.
- Pair 7 produced the largest regression: 353/353 fell to 152/353.
- The longest observed dependency chain reached 35 passengers.

Principal Findings
------------------
1. The broad-cohort architecture again produced a genuine converted-success
   case, confirming that it can be constructive for selected manifests.
2. Independent replication did not confirm an aggregate advantage.
3. Nineteen worse pairs and a 1,619-passenger aggregate loss show that the
   effect is strongly scenario-dependent.
4. The architecture is more stable than local subdivision and completion-gate
   variants, but it is not generally superior to the standard deterministic
   dataset.
5. Strict rear-to-front geographic progression can still concentrate a
   rearward admission wave and create severe internal dependency chains.
6. The independent-seed result therefore weakens any claim that Experiment 27
   represented a broadly reproducible improvement.
7. Ordinary numerical entry headway remained active throughout.
8. The movement and dependency architecture remained frozen.

Discussion Incorporated Before Experiment 29
---------------------------------------------
The independent-seed result confirms that selected improvements are real but
not reliably generalisable. Because one replication can still be unusually
favourable or unfavourable, Experiment 29 performs one final architecture-
identical replication using a third deterministic experiment seed.

Experiment 28 Conclusion
------------------------
Broad geographic cohorts should be reported as a conditional policy whose
effect depends strongly on manifest composition and scenario geometry. A second
independent-seed replication is justified to strengthen the closure of this
research branch without introducing a new optimisation mechanism.


======================================================================
EXPERIMENT 29 - SECOND INDEPENDENT-SEED COARSE-COHORT REPLICATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen.
Experiment 29 retains the same equal-thirds rear-to-middle-to-front admission
policy used in Experiments 27 and 28. Only the reproducible experiment seed
changes.

Planned Purpose
---------------
Provide a third 30-pair sample so that the coarse-cohort conclusion is not based
on one original dataset and one independent replication.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. SECOND INDEPENDENT-SEED COARSE-COHORT DATASET.

Experimental Seed
-----------------
    2283790101901

Admission Contract
------------------
- Equal front, middle and rear row thirds are retained.
- Rear cohort admission completes before middle cohort admission begins.
- Middle cohort admission completes before front cohort admission begins.
- Seeded relative order is preserved inside each aisle/cohort queue.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- There are no local subregions, completion gates, overlap, micro-batches,
  complexity classes or runtime-adaptive release rules.
- Within every pair, standard and cohort modes share the same new scenario
  seed, manifest, seat assignments, aisle assignments and headway.
- Runtime congestion cannot alter cohort membership or progression.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    THIRD DETERMINISTIC SCENARIO SAMPLE

Primary Evaluation Order
------------------------
1. Deterministic replay under the new experiment seed.
2. Aggregate seated passengers.
3. Improved, worse and equal pair counts.
4. Converted-success cases.
5. Comparison with Experiments 27 and 28.
6. Outside-queue accumulation and entry saturation.
7. Longest dependency-chain length and critical-tree depth.
8. Residual stall-family distribution.

Experimental Hypothesis
-----------------------
If broad geographic cohesion has a stable constructive effect, a third scenario
sample should again produce meaningful improvements without a large aggregate
loss.

If the third sample again performs materially worse overall, the branch can be
closed with stronger evidence that coarse geographic cohorts are conditional
rather than generally beneficial.

Interpretation Safeguards
-------------------------
- The architecture is identical to Experiments 27 and 28.
- Only the deterministic experiment seed changes.
- Standard and cohort executions remain paired within every scenario.
- Numerical entry headway remains active independently for both aisles.
- Rear-boundary locks remain a separate physical limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 28
======================================================================
1. Experiment 28 lost 1,619 seated passengers overall.
2. One converted-success case did not offset 19 worse pairs.
3. The coarse-cohort effect failed independent aggregate replication.
4. Selected benefits remain real but strongly manifest-dependent.
5. Experiment 29 performs one final architecture-identical replication.
6. No new admission mechanism is introduced.
7. The new experiment seed is 2283790101901.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v29 - EXPERIMENT 29 DESIGN
======================================================================


======================================================================
EXPERIMENT 29 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Standard and second independent-seed cohort modes shared
each scenario seed, manifest, cabin geometry, occupancy, numerical headway,
seat assignments, aisle assignments, movement rules and dependency
architecture. Only the equal-thirds rear-to-middle-to-front admission policy
differed.

Observed Results
----------------
- Standard mode seated 7,880 of 7,965 passengers.
- Second independent-seed coarse-cohort mode seated 6,962 of 7,965 passengers.
- The aggregate change was a loss of 918 seated passengers.
- Coarse cohorts improved no pairs, were worse in 12 and were equal in 18.
- No incomplete standard case converted to complete success.
- Pair 27 produced the largest regression: 342/342 fell to 79/342.
- Pair 30 fell from 325/353 to 139/353.
- The longest observed dependency chain reached 34 passengers.

Principal Findings
------------------
1. The third architecture-identical sample produced no improved pair.
2. Twelve regressions caused a 918-passenger aggregate loss.
3. Eighteen equal pairs show that the policy can remain neutral, but neutrality
   does not offset the severe failures in sensitive manifests.
4. The result reinforces Experiment 28: strict rear-to-front coarse cohorts are
   not a generally superior deterministic admission policy.
5. Converted-success cases seen in Experiments 27 and 28 are real but rare and
   manifest-specific.
6. The repeated negative aggregate direction now justifies closure rather than
   further architectural tuning.
7. Ordinary numerical entry headway remained active throughout.
8. The movement and dependency architecture remained frozen.

Three-Sample Pooled Position After Experiments 27-29
----------------------------------------------------
- Standard passengers seated: 24,238 of 24,560.
- Coarse-cohort passengers seated: 21,075 of 24,560.
- Pooled aggregate change: -3,163 passengers.
- Improved pairs: 4 of 90.
- Worse pairs: 43 of 90.
- Equal pairs: 43 of 90.
- Converted-success cases: 2 of 90.

Discussion Incorporated Before Experiment 30
---------------------------------------------
The early Experiment 29 scenarios again appeared stable, but the full run
contained no improved pair and several severe regressions. This confirms why
the report must rely on complete paired datasets rather than early examples.

Experiment 30 therefore does not introduce another optimisation. It performs
one final pre-registered architecture-identical sample and then closes the
branch using a four-seed pooled interpretation.

Experiment 29 Conclusion
------------------------
The second independent replication fails to support coarse geographic cohorts
as a general improvement. Across three samples, benefits are uncommon and
large regressions dominate the aggregate result.


======================================================================
EXPERIMENT 30 - POOLED FOUR-SEED COARSE-COHORT CLOSURE
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen.
Experiment 30 retains the identical equal-thirds rear-to-middle-to-front
admission policy from Experiments 27-29. Only the reproducible experiment seed
changes for the fourth and final sample.

Planned Purpose
---------------
Run one final 30-pair architecture-identical sample and then close the coarse-
cohort branch with pooled evidence across four deterministic experiment seeds.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. POOLED-CLOSURE COARSE-COHORT DATASET.

Experimental Seed
-----------------
    2383790101901

Admission Contract
------------------
- Equal front, middle and rear row thirds are retained.
- Rear cohort admission completes before middle cohort admission begins.
- Middle cohort admission completes before front cohort admission begins.
- Seeded relative order is preserved inside each aisle/cohort queue.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- No local subregions, completion gates, overlap, micro-batches, complexity
  classes or runtime-adaptive release rules are introduced.
- Standard and cohort executions share the same new scenario seed, manifest,
  seat assignments, aisle assignments and headway within every pair.
- Runtime congestion cannot alter cohort membership or progression.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    FOURTH DETERMINISTIC SCENARIO SAMPLE

Primary Evaluation Order
------------------------
1. Deterministic replay under the final experiment seed.
2. Aggregate seated passengers.
3. Improved, worse and equal pair counts.
4. Converted-success cases.
5. Four-seed pooled aggregate direction.
6. Outside-queue accumulation and entry saturation.
7. Longest dependency-chain length and critical-tree depth.
8. Final branch-closure interpretation.

Experimental Hypothesis
-----------------------
If broad geographic cohesion has a stable constructive effect, the fourth
sample should materially improve the pooled direction.

If the sample again performs worse overall, the branch will close with strong
evidence that strict coarse geographic cohorts are conditional and generally
inferior to the standard deterministic dataset under this frozen architecture.

Interpretation Safeguards
-------------------------
- The architecture is identical to Experiments 27-29.
- Only the deterministic experiment seed changes.
- The final interpretation will use all four complete samples.
- Numerical entry headway remains active independently for both aisles.
- Rear-boundary locks remain a separate physical limitation.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 29
======================================================================
1. Experiment 29 lost 918 seated passengers overall.
2. It produced zero improved and zero converted-success pairs.
3. Three pooled samples show a -3,163 passenger aggregate change.
4. Only four of 90 pairs improved, while 43 became worse.
5. Broad-cohort benefits are rare and manifest-dependent.
6. Experiment 30 supplies the fourth and final identical sample.
7. No new admission mechanism is introduced.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v30 - EXPERIMENT 30 DESIGN
======================================================================


======================================================================
EXPERIMENT 30 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Standard and pooled-closure cohort modes shared each
scenario seed, manifest, cabin geometry, occupancy, numerical headway, seat
assignments, aisle assignments, movement rules and dependency architecture.
Only the equal-thirds rear-to-middle-to-front admission policy differed.

Observed Results
----------------
- Standard mode seated 7,578 of 7,739 passengers.
- Pooled-closure coarse-cohort mode seated 6,661 of 7,739 passengers.
- The aggregate change was a loss of 917 seated passengers.
- Coarse cohorts improved one pair, were worse in 14 and were equal in 15.
- One incomplete standard case converted to complete success.
- Pair 27 improved from 200/259 to 259/259 and converted to success.
- Pair 22 produced the largest regression: 306/306 fell to 132/306.
- The longest observed dependency chain reached 32 passengers.

Four-Sample Pooled Position After Experiments 27-30
----------------------------------------------------
- Standard passengers seated: 31,816 of 32,299.
- Coarse-cohort passengers seated: 27,736 of 32,299.
- Pooled aggregate change: -4,080 passengers.
- Improved pairs: 5 of 120.
- Worse pairs: 57 of 120.
- Equal pairs: 58 of 120.
- Converted-success cases: 3 of 120.

Principal Findings
------------------
1. A fourth architecture-identical sample again produced a negative aggregate
   result.
2. The strong Pair 27 conversion confirms that coarse cohorts can solve a
   specific manifest.
3. Fifty-seven pooled regressions and a 4,080-passenger loss outweigh the five
   improved pairs.
4. Strict rear-to-front coarse cohorts are therefore conditional rather than
   generally beneficial under the frozen architecture.
5. The repeated independent samples make further seed replication unnecessary.
6. Severe failures continue to involve entry saturation, long internal chains
   and unresolved rearward seat events.
7. Ordinary numerical entry headway remained active throughout.
8. The movement and dependency architecture remained frozen.

Discussion Incorporated Before Experiment 31
---------------------------------------------
The earlier discussion correctly identified the value of the isolated converted
successes, but the full four-seed evidence now closes the branch decisively.
The next experiment should not tune cohort boundaries or repeat another seed.

Experiment 31 therefore returns to the standard deterministic queue and targets
one repeatedly observed physical mechanism only: blocker-requiring seats in the
final two cabin rows, where insufficient rear yield space often terminates the
dependency chain.

Experiment 30 Conclusion
------------------------
Broad geographic cohorts are not retained as the new reference policy. Their
benefits are real but rare, while the repeated aggregate direction is strongly
negative. The research now moves from global admission grouping to a narrow,
precomputed rear-boundary protection test.


======================================================================
EXPERIMENT 31 - REAR-BOUNDARY RISK-FIRST ADMISSION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original deterministic people-movement backbone remains frozen.
Experiment 31 changes only the precomputed order of a small passenger class
before entry. Once admitted, every passenger follows the same one-tile
synchronous movement, row limit, blocker-yield, seat-event, middle-bank,
dependency-region lifecycle and hierarchy rules.

Planned Purpose
---------------
Test whether passengers whose assigned seats are most exposed to the physical
rear boundary can be admitted before ordinary passengers without recreating the
global geographic waves rejected in Experiments 27-30.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. REAR-BOUNDARY RISK-FIRST DATASET.

Experimental Seed
-----------------
    2483790101901

Protected Class
---------------
A passenger belongs to the protected class only when both conditions are true:

1. The assigned row is one of the final two cabin rows.
2. The assigned seat requires at least one seated blocker to yield.

Admission Contract
------------------
- Standard independent left/right aisle queues are restored.
- A seeded shuffle first establishes reproducible relative order.
- Protected rear-boundary passengers are stably moved to the front of their
  serving-aisle queue.
- Relative order remains unchanged inside the protected class.
- Relative order remains unchanged inside the ordinary class.
- All ordinary passengers follow immediately after the protected class.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- Entry tiles must be empty before insertion.
- No broad geographic cohorts or local regions are used.
- No runtime congestion evidence changes the order.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    PRECOMPUTED FINAL-TWO-ROW BLOCKER-RISK PRIORITY

Primary Evaluation Order
------------------------
1. Deterministic replay and protected-class membership.
2. Rear-boundary-lock frequency.
3. Aggregate seated passengers.
4. Improved, worse and equal pair counts.
5. Converted-success cases.
6. Entry saturation and outside-queue accumulation.
7. Longest dependency-chain length and critical-tree depth.
8. Linked row-event and internal-chain residual families.

Experimental Hypothesis
-----------------------
If repeated rear-boundary locks arise because blocker-requiring final-row seat
events occur after the rear aisle has already become constrained, admitting
that small risk class first should reduce those locks without imposing a global
rear-to-front wave.

A negative result would show that even narrow pre-entry prioritisation disrupts
useful seeded relationships or merely moves congestion earlier in the run.

Interpretation Safeguards
-------------------------
- Protection is based only on fixed seat geometry.
- The class is calculated before execution.
- No runtime optimisation or threshold selection is introduced.
- The remaining manifest keeps its seeded relative order.
- Numerical entry headway remains active independently for both aisles.
- The controlled dataset is not an airline boarding recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 30
======================================================================
1. Four coarse-cohort samples lost 4,080 passengers in aggregate.
2. Only five of 120 pairs improved.
3. Three converted-success cases were real but rare.
4. The broad-cohort branch is now closed.
5. Experiment 31 restores standard independent aisle queues.
6. Only final-two-row blocker-risk passengers receive precomputed priority.
7. No global geographic wave or runtime adaptation is introduced.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v31 - EXPERIMENT 31 DESIGN
======================================================================


======================================================================
EXPERIMENT 31 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic verification run executed all 30 paired scenarios
(60 total executions). Standard and rear-boundary risk-first modes shared each
scenario seed, manifest, cabin geometry, occupancy, numerical headway, seat
assignments, aisle assignments, movement rules and dependency architecture.
Only the precomputed queue position of blocker-requiring final-two-row
passengers differed.

Observed Results
----------------
- Standard mode seated 7,545 of 7,619 passengers.
- Rear-boundary risk-first mode seated 5,663 of 7,619
  passengers.
- The aggregate change was a loss of 1,882
  seated passengers.
- Risk-first admission improved 1 pair, was worse in
  24 and was equal in 5.
- 1 incomplete standard case converted to complete success.
- Pair 12 produced the only improvement:
  350/353 to
  353/353.
- Pair 21 produced the largest regression:
  324/324 to
  109/324.
- The longest observed dependency chain reached 36
  passengers.

Principal Findings
------------------
1. Global prioritisation of the rear-boundary risk class failed strongly.
2. Twenty-four of 30 pairs regressed, while only one improved.
3. The policy created early entry saturation and long internal dependency
   chains.
4. Rear-boundary locks remained present despite admitting the nominally risky
   passengers first.
5. The protected class is therefore not the initiating cause of congestion.
6. Rear-boundary passengers are more accurately interpreted as terminal victims
   of dependency structures that have already formed.
7. The result separates the visible end-state symptom from the upstream cause.
8. Ordinary numerical entry headway and the frozen movement architecture
   remained active.

Discussion Incorporated Before Experiment 32
---------------------------------------------
The discussion following Experiment 31 identified the central interpretation:
rear-boundary locks are often the final visible manifestation of an earlier
dependency chain rather than a problem solved by moving all rear-risk
passengers to the front.

The next test should therefore avoid another priority block. Experiment 32
retains the standard seeded queue as closely as possible and changes only
consecutive clustering of the same risk class.

Experiment 31 Conclusion
------------------------
Rear-boundary risk-first admission is rejected. It amplifies entry saturation
and internal chains while failing to remove the terminal rear-boundary lock.
The research proceeds to a minimal stable-dispersion test rather than another
global priority policy.


======================================================================
EXPERIMENT 32 - REAR-BOUNDARY RISK DISPERSION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 32
changes only the precomputed order in which a very small risk class enters.
Once admitted, every passenger follows the same one-tile movement, blocker
yield, seat-event, middle-bank, dependency-region lifecycle and hierarchy
rules.

Planned Purpose
---------------
Determine whether preventing consecutive admission of blocker-requiring
final-two-row passengers can reduce rear-boundary chain formation without
creating the large priority block that failed in Experiment 31.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. REAR-BOUNDARY RISK-DISPERSION DATASET.

Experimental Seed
-----------------
    2583790101901

Protected Class
---------------
A passenger belongs to the protected class only when both conditions are true:

1. The assigned row is one of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Admission Contract
------------------
- Independent left/right aisle queues are retained.
- A seeded shuffle establishes the initial reproducible order.
- The initial order is preserved unless two protected passengers would be
  emitted consecutively.
- A consecutive protected passenger is deferred until after the next ordinary
  passenger.
- Only one deferred protected passenger is reinserted after each ordinary
  passenger.
- Relative order among deferred protected passengers is preserved.
- Relative order among ordinary passengers is preserved.
- Remaining deferred protected passengers are appended only if no ordinary
  passengers remain.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- No broad cohort, completion gate, runtime scan or adaptive threshold is used.
- No passenger changes seat, serving aisle or destination.

Independent Variable
--------------------
    STABLE DISPERSION OF CONSECUTIVE REAR-BOUNDARY RISK PASSENGERS

Primary Evaluation Order
------------------------
1. Deterministic replay and protected-class membership.
2. Aggregate seated passengers.
3. Improved, worse and equal pair counts.
4. Converted-success cases.
5. Rear-boundary-lock frequency.
6. Entry saturation and outside-queue accumulation.
7. Longest dependency-chain length and critical-tree depth.
8. Residual stall-family comparison.

Experimental Hypothesis
-----------------------
If Experiment 31 failed primarily because it concentrated the risk class at the
front of the queue, stable dispersion should avoid that early saturation while
still preventing multiple rear-risk seat events from arriving together.

If the result remains negative, then even small pre-entry manipulation of this
class is unlikely to address the upstream dependency mechanism.

Interpretation Safeguards
-------------------------
- The rule is calculated before execution from fixed seat geometry.
- No runtime congestion evidence alters the order.
- Standard seeded order is changed only when consecutive risk passengers occur.
- Numerical entry headway remains independently active for both aisles.
- Rear-boundary lock remains a physical terminal condition, not an airline
  recommendation.
- The movement and dependency architecture remains frozen.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 31
======================================================================
1. Experiment 31 lost 1,882 passengers overall.
2. Twenty-four of 30 pairs regressed.
3. Rear-risk priority created entry saturation rather than preventing it.
4. Rear-boundary locks are usually terminal symptoms of earlier chains.
5. Experiment 32 removes the priority block.
6. Only consecutive protected passengers are deterministically dispersed.
7. Standard seeded order is otherwise preserved.
8. Normal entry headway and the frozen movement architecture remain active.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v32 - EXPERIMENT 32 DESIGN
======================================================================


======================================================================
EXPERIMENT 32 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete deterministic run executed all 30 paired scenarios (60 total
executions). Standard and rear-boundary risk-dispersion modes shared each
scenario seed, manifest, cabin geometry, occupancy, numerical headway, seat
assignments, aisle assignments, movement rules and dependency architecture.
Only consecutive clustering of the protected final-two-row risk class differed.

Observed Results
----------------
- Standard mode seated 8,362 of 8,433 passengers.
- Risk-dispersion mode seated 8,388 of 8,433
  passengers.
- Aggregate change: +26 seated passengers.
- Improved pairs: 1.
- Worse pairs: 0.
- Equal pairs: 29.
- Converted-success cases: 1.
- Pair 22 improved from
  256/282 to
  282/282.
- The longest observed dependency chain reached 12
  passengers.

Principal Findings
------------------
1. The stable dispersion rule produced no regression in any paired scenario.
2. Twenty-nine pairs remained exactly equal.
3. Pair 22 gained 26 seated passengers and converted to complete success.
4. The aggregate gain is entirely attributable to that one converted manifest.
5. The result supports minimal deterministic perturbation over broad queue
   restructuring.
6. Preserving seeded order while separating consecutive risk passengers can
   remove a specific dependency configuration without disturbing most cases.
7. The result is promising but not yet general because only one pair changed.
8. Ordinary numerical entry headway and the frozen movement architecture
   remained active throughout.

Discussion Incorporated Before Experiment 33
---------------------------------------------
The discussion following Experiment 32 distinguished two levels of conclusion.

Within the defined cabin model, the engine now provides a strong and increasingly
comprehensive deterministic research platform. It can compare many manifests,
cabin layouts, occupancies, headways and admission policies while exposing
dependency chains and reproducible failure families.

It is not yet universally conclusive for live airline operations. Operational
use would require calibrated datasets for luggage, walking speeds, families,
assisted passengers, compliance, crew intervention, aircraft-specific
obstructions and other real boarding effects, followed by validation against
observed operations.

Experiment 33 therefore does not expand operational claims. It performs the
scientifically necessary next step: an independent-seed replication of the
unchanged Experiment 32 policy.

Experiment 32 Conclusion
------------------------
Rear-boundary risk dispersion is retained as a promising low-disturbance
strategy. The zero-regression result is stronger than earlier mixed policies,
but the single improved manifest requires independent replication before
robustness can be claimed.


======================================================================
EXPERIMENT 33 - INDEPENDENT-SEED RISK-DISPERSION REPLICATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 33
changes only the experiment seed. The protected class, stable deferral rule,
ordinary entry headway, passenger movement, blocker-yield, seat-event,
middle-bank and dependency architecture are identical to Experiment 32.

Planned Purpose
---------------
Determine whether Experiment 32's zero-regression and converted-success result
persists under a new independent deterministic sample.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. REPLICATED REAR-BOUNDARY RISK-DISPERSION DATASET.

Experimental Seed
-----------------
    2683790101901

Protected Class
---------------
A passenger belongs to the protected class only when both conditions are true:

1. The assigned row is one of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Replication Contract
--------------------
- The Experiment 32 protected-class definition is unchanged.
- The Experiment 32 stable-dispersion algorithm is unchanged.
- A seeded shuffle establishes the new reproducible sample.
- A second consecutive protected passenger is deferred until after the next
  ordinary passenger.
- Relative order among deferred protected passengers is preserved.
- Relative order among ordinary passengers is preserved.
- Remaining deferred protected passengers are appended only when no ordinary
  passengers remain.
- Independent left/right entry queues are retained.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- No runtime adaptation, cohort, gate or threshold is introduced.
- No passenger changes seat, serving aisle or destination.
- Only the deterministic experiment seed changes.

Independent Variable
--------------------
    INDEPENDENT DETERMINISTIC SAMPLE

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Rear-boundary-lock frequency.
6. Entry saturation and outside-queue accumulation.
7. Longest dependency-chain length.
8. Agreement or disagreement with Experiment 32.

Experimental Hypothesis
-----------------------
If the stable dispersion rule is generally low risk, most pairs should remain
equal and regressions should remain absent or rare. Repeated converted successes
would strengthen the case for the policy.

A neutral replication would show that the rule is safe but conditionally
active. A negative replication would show that the Experiment 32 result was
sample-specific.

Operational Interpretation Boundary
-----------------------------------
The code is increasingly conclusive for deterministic dependency behaviour
inside its defined model and can support research exploration by aviation or
operations specialists. It does not by itself establish universal live-airline
performance. Operational conclusions require real-world parameter calibration,
domain datasets and empirical validation.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 32
======================================================================
1. Experiment 32 gained 26 passengers overall.
2. One pair improved and converted to complete success.
3. No pair regressed.
4. Twenty-nine pairs remained exactly equal.
5. Minimal stable perturbation is more promising than broad restructuring.
6. The positive result is still based on one changed manifest.
7. Experiment 33 repeats the identical policy under a new seed.
8. Operational claims remain bounded by the current deterministic cabin model.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v33 - EXPERIMENT 33 DESIGN
======================================================================


======================================================================
EXPERIMENT 33 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The complete independent-seed replication executed all 30 paired scenarios
(60 total executions). Standard and replicated risk-dispersion modes shared
each scenario seed, manifest, cabin geometry, occupancy, numerical headway,
seat assignments, aisle assignments, movement rules and dependency
architecture. Only the stable dispersion of consecutive protected passengers
differed.

Observed Results
----------------
- Standard mode seated 7,965 of 8,103 passengers.
- Replicated risk-dispersion mode seated 7,975 of
  8,103 passengers.
- Aggregate change: +10 seated passengers.
- Improved pairs: 1.
- Worse pairs: 0.
- Equal pairs: 29.
- Converted-success cases: 0.
- Pair 23 improved from
  270/288 to
  280/288.
- The longest observed dependency chain reached 28
  passengers.

Principal Findings
------------------
1. Experiment 33 reproduced the Experiment 32 structural pattern exactly:
   one improved pair, no regressions and 29 equal pairs.
2. The aggregate gain was smaller than Experiment 32 but remained positive.
3. Pair 23 improved by ten passengers without reaching complete success.
4. No standard success was damaged.
5. No stall family worsened into a lower seated outcome.
6. The policy again behaved as a highly conservative deterministic
   perturbation.
7. The absence of regression across two independent seeds is now stronger
   evidence than the isolated Experiment 32 result.
8. Because both gains came from one manifest per seed, the policy remains
   conditionally active rather than generally transformative.

Comparison with Experiment 32
-----------------------------
Experiment 32:
- Standard: 8,362/8,433.
- Risk dispersion: 8,388/8,433.
- Aggregate change: +26.
- Improved/worse/equal: 1/0/29.
- Converted successes: 1.

Experiment 33:
- Standard: 7,965/8,103.
- Risk dispersion: 7,975/8,103.
- Aggregate change: +10.
- Improved/worse/equal: 1/0/29.
- Converted successes: 0.

Pooled Experiments 32-33:
- Standard: 16,327/16,536.
- Risk dispersion: 16,363/16,536.
- Aggregate change: +36.
- Improved pairs: 2/60.
- Worse pairs: 0/60.
- Equal pairs: 58/60.
- Converted-success cases: 1/60.

Scientific Interpretation
-------------------------
The independent replication supports the low-risk interpretation of the policy.
The algorithm does not broadly improve all manifests; instead, it leaves most
scenarios untouched and occasionally removes a specific dependency
configuration.

This is materially different from earlier broad admission strategies, which
often gained passengers in some scenarios while producing substantial
regressions elsewhere. The present rule has so far changed only two of 60
paired scenarios and both changes were beneficial.

Discussion Incorporated
------------------------
The code is increasingly conclusive for deterministic dependency behaviour
inside its defined model and can be used by aviation or operations researchers
to explore many boarding scenarios. It should not yet be described as fully
conclusive for live airline operations without calibrated operational inputs
and empirical validation.

Experiment 33 Conclusion
------------------------
Experiment 33 successfully replicates the zero-regression character of
Experiment 32. The policy is retained as a promising, conditionally active,
low-disturbance intervention. One further independent-seed replication is
appropriate before formally accepting or closing this branch.


======================================================================
EXPERIMENT 34 - SECOND INDEPENDENT-SEED RISK-DISPERSION REPLICATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 34
changes only the experiment seed. The protected class, stable deferral rule,
ordinary entry headway, passenger movement, blocker-yield, seat-event,
middle-bank and dependency architecture are identical to Experiments 32 and 33.

Planned Purpose
---------------
Test whether the two-seed zero-regression pattern survives a third independent
deterministic sample.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. SECOND REPLICATED REAR-BOUNDARY RISK-DISPERSION DATASET.

Experimental Seed
-----------------
    2783790101901

Replication Contract
--------------------
- The protected class remains blocker-requiring passengers in the final two
  cabin rows.
- The stable-dispersion algorithm is unchanged.
- A second consecutive protected passenger is deferred until after the next
  ordinary passenger.
- Relative order among deferred protected passengers is preserved.
- Relative order among ordinary passengers is preserved.
- Independent left/right entry queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation, cohort, gate or threshold is introduced.
- No passenger changes seat, serving aisle or destination.
- Only the deterministic experiment seed changes.

Independent Variable
--------------------
    THIRD INDEPENDENT DETERMINISTIC SAMPLE

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Rear-boundary-lock frequency.
6. Entry saturation and outside-queue accumulation.
7. Longest dependency-chain length.
8. Agreement or disagreement with Experiments 32 and 33.

Experimental Hypothesis
-----------------------
If the stable dispersion rule is robustly low risk, Experiment 34 should again
produce mostly equal pairs with no or very few regressions.

A third zero-regression result would support formal acceptance of the branch as
a conditionally beneficial, low-disturbance policy. A negative result would
show that the first two samples were insufficient.

Operational Interpretation Boundary
-----------------------------------
The engine is increasingly conclusive for deterministic dependency behaviour
inside its defined cabin model. Live operational claims still require
airline-specific datasets, behavioural calibration and empirical validation.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 33
======================================================================
1. Experiments 32 and 33 both produced one improvement and no regressions.
2. Across 60 pairs, two improved, none worsened and 58 were equal.
3. The pooled aggregate gain is +36 passengers.
4. One scenario converted to complete success.
5. The policy is low-disturbance and conditionally active.
6. Broad restructuring remains less promising than minimal stable perturbation.
7. Experiment 34 provides a third independent deterministic sample.
8. Operational conclusions remain bounded by the current cabin model.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v34 - EXPERIMENT 34 DESIGN
======================================================================


======================================================================
EXPERIMENT 34 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The second independent-seed replication executed all 30 paired scenarios
(60 total executions). Standard and risk-dispersion modes shared each scenario
seed, manifest, cabin geometry, occupancy, numerical headway, seat assignments,
aisle assignments, movement rules and dependency architecture. Only stable
dispersion of consecutive blocker-requiring final-two-row passengers differed.

Observed Results
----------------
- Standard mode seated 8,277 of 8,356 passengers.
- Risk-dispersion mode seated 8,310 of 8,356
  passengers.
- Aggregate change: +33 seated passengers.
- Improved pairs: 3.
- Worse pairs: 0.
- Equal pairs: 27.
- Converted-success cases: 3.
- Longest observed dependency chain: 24 passengers.

Changed Pairs
-------------
- Pair 01: 340/342 -> 342/342 (+2); converted success=true.
- Pair 18: 264/288 -> 288/288 (+24); converted success=true.
- Pair 19: 335/342 -> 342/342 (+7); converted success=true.

Principal Findings
------------------
1. Experiment 34 is the strongest result in the risk-dispersion branch.
2. Three paired scenarios improved and all three converted to complete success.
3. No paired scenario regressed.
4. Twenty-seven pairs remained exactly equal.
5. Pair 01 directly removed a rear-boundary lock.
6. Pair 18 produced the largest gain, adding 24 seated passengers.
7. Pair 19 added seven passengers and also converted to complete success.
8. The policy again remained highly conservative: it altered only scenarios
   where its deterministic queue perturbation changed a critical dependency.
9. The result independently confirms the zero-regression pattern from
   Experiments 32 and 33.

Three-Experiment Pooled Evidence
--------------------------------
Experiments 32-34 combined:
- Standard: 24,604/24,892.
- Risk dispersion: 24,673/24,892.
- Aggregate improvement: +69 passengers.
- Improved pairs: 5/90.
- Worse pairs: 0/90.
- Equal pairs: 85/90.
- Converted-success cases: 4/90.

Scientific Verdict
------------------
Rear-boundary risk dispersion is provisionally accepted as a robust,
conditionally beneficial and low-disturbance admission policy within the frozen
deterministic cabin model.

The policy is not broadly transformative. Its strength is that it preserves
most manifests exactly while occasionally preventing a specific deterministic
failure. Across three independent seeds, no regression has been observed.

Discussion Incorporated
------------------------
The project is now entering a final boundary-and-closure stage. Reaching a
well-finalised position by Experiment 40 remains realistic provided subsequent
experiments test the limits of the accepted rule rather than repeating identical
replications indefinitely.

A suitable closing sequence is:
- boundary testing of the protected geographic scope;
- confirmation of any retained boundary;
- a final holdout or pooled stress sample;
- synthesis of accepted and rejected policy families;
- Experiment 40 as the formal deterministic research conclusion.

The model is increasingly conclusive for deterministic dependency behaviour
inside its defined cabin assumptions. Live-airline operational claims still
require calibrated luggage, walking-speed, family, accessibility, compliance,
crew and aircraft-specific datasets, followed by empirical validation.

Experiment 34 Conclusion
------------------------
The three-seed replication objective is complete. The final-two-row dispersion
policy is retained as the accepted reference candidate. Experiment 35 begins
boundary testing rather than another identical independent-seed replication.


======================================================================
EXPERIMENT 35 - LAST-ROW-ONLY RISK-DISPERSION BOUNDARY TEST
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 35
changes only the pre-run protected-class boundary. Passenger movement,
blocker-yield behaviour, seat events, middle-bank arbitration, dependency
diagnostics, ordinary entry headway and deterministic replay remain unchanged.

Planned Purpose
---------------
Determine whether the accepted final-two-row protected scope is wider than
necessary.

Experiment 35 narrows protection to blocker-requiring passengers assigned to
the final cabin row only. It deliberately reuses the Experiment 34 seed so that
the strongest prior sample becomes a matched boundary comparison.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. LAST-ROW-ONLY RISK-DISPERSION DATASET.

Experimental Seed
-----------------
    2783790101901

Protected Class
---------------
A passenger belongs to the Experiment 35 protected class only when both
conditions are true:

1. The passenger is assigned to the final cabin row.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active independently for both
  aisles.
- No runtime adaptation is introduced.

Independent Variable
--------------------
    PROTECTED GEOGRAPHIC SCOPE:
    FINAL ROW ONLY versus NO DISPERSION

Historical Matched Comparator
-----------------------------
Because Experiment 35 reuses the Experiment 34 seed, its results can also be
compared directly with the Experiment 34 final-two-row policy:

- Experiment 34 final-two-row result: 8,310/8,356.
- Experiment 34 standard result: 8,277/8,356.
- Experiment 34 converted three scenarios to complete success.

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Preservation or loss of Experiment 34's three conversions.
6. Rear-boundary-lock frequency.
7. Dependency-chain and stall-family changes.
8. Evidence that the penultimate row is or is not necessary.

Decision Rule
-------------
- If the last-row-only rule preserves the Experiment 34 gains without
  regression, adopt the narrower boundary.
- If gains disappear without regression, retain the final-two-row scope because
  the penultimate row contributes materially.
- If regressions appear, reject the narrower scope and retain the accepted
  final-two-row definition.

Experimental Hypothesis
-----------------------
The final row is the most geometrically constrained location, but Experiment 34
may have benefited from interactions involving both the final and penultimate
rows. The narrower rule may therefore preserve some rear-boundary conversions
while losing others.

Operational Interpretation Boundary
-----------------------------------
This remains a deterministic model-boundary experiment, not a live-airline
policy recommendation. Operational use requires domain calibration and
empirical validation.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 34
======================================================================
1. The final-two-row risk-dispersion policy has passed three independent seeds.
2. Across 90 pairs, five improved, none worsened and 85 were equal.
3. Four incomplete scenarios converted to complete success.
4. The pooled aggregate improvement is +69 passengers.
5. Experiment 34 supplied the strongest single replication.
6. The policy is provisionally accepted within the frozen model.
7. Experiment 35 tests whether the final row alone is sufficient.
8. A finalised position by Experiment 40 remains realistic.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v35 - EXPERIMENT 35 DESIGN
======================================================================


======================================================================
EXPERIMENT 35 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The matched last-row-only boundary test executed all 30 paired scenarios
(60 total executions). Both modes shared the Experiment 34 seed, manifests,
cabin geometry, occupancy, numerical headway, seat assignments, aisle
assignments, movement rules and dependency architecture. Only the protected
geographic scope differed.

Observed Results
----------------
- Standard mode seated 8,277 of 8,356 passengers.
- Last-row-only dispersion seated 8,277 of
  8,356 passengers.
- Aggregate change: +0 passengers.
- Improved pairs: 0.
- Worse pairs: 0.
- Equal pairs: 30.
- Converted-success cases: 0.
- Longest observed dependency chain: 24 passengers.

Principal Findings
------------------
1. Every paired result was identical to the standard control.
2. The three Experiment 34 converted successes disappeared completely.
3. The last-row-only rule produced neither benefit nor regression.
4. Pair 01 remained a rear-boundary lock at 340/342 rather than converting to
   342/342 as it did under the final-two-row rule.
5. The final row alone is therefore insufficient to reproduce the accepted
   policy's effect.
6. The penultimate row is materially involved in the successful mechanism.
7. Experiment 35 does not weaken Experiment 34. It defines the lower boundary
   and explains why the broader final-two-row scope is necessary.

Matched Boundary Comparison
---------------------------
Experiment 34, final-two-row dispersion:
- Standard: 8,277/8,356.
- Tested: 8,310/8,356.
- Aggregate change: +33.
- Improved/worse/equal: 3/0/27.
- Converted successes: 3.

Experiment 35, final-row-only dispersion:
- Standard: 8,277/8,356.
- Tested: 8,277/8,356.
- Aggregate change: 0.
- Improved/worse/equal: 0/0/30.
- Converted successes: 0.

Scientific Verdict
------------------
The last-row-only variant is rejected as ineffective.

The accepted reference remains stable dispersion of blocker-requiring
passengers across the final two cabin rows. Experiment 35 establishes that the
penultimate row cannot simply be removed from the protected class.

Discussion Incorporated
------------------------
This result strengthens the publication because it identifies a meaningful
boundary rather than merely repeating a positive policy result. The research
now distinguishes between:

- an effective final-two-row interaction;
- an ineffective final-row-only simplification; and
- the unresolved question of whether the penultimate row alone carries all or
  only part of the effect.

The project remains on course for a well-finalised position by Experiment 40.
The remaining experiments should complete component isolation, confirmation,
holdout testing and formal synthesis rather than reopen broad strategy families.

Experiment 35 Conclusion
------------------------
The final row is not independently sufficient. The penultimate row is essential
to the accepted policy, either as the active component or through interaction
with the final row. Experiment 36 isolates that remaining component.


======================================================================
EXPERIMENT 36 - PENULTIMATE-ROW-ONLY RISK-DISPERSION COMPONENT TEST
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 36
changes only the pre-run protected-class boundary. Passenger movement,
blocker-yield behaviour, seat events, middle-bank arbitration, dependency
diagnostics, numerical entry headway and deterministic replay remain unchanged.

Planned Purpose
---------------
Determine whether the penultimate row alone carries the Experiment 34 benefit,
or whether the successful mechanism requires the combined final-two-row scope.

Experiment 36 reuses the Experiment 34 and 35 seed, creating a direct matched
component comparison.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. PENULTIMATE-ROW-ONLY RISK-DISPERSION DATASET.

Experimental Seed
-----------------
    2783790101901

Protected Class
---------------
A passenger belongs to the Experiment 36 protected class only when both
conditions are true:

1. The passenger is assigned to the penultimate cabin row.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variable
--------------------
    PROTECTED GEOGRAPHIC SCOPE:
    PENULTIMATE ROW ONLY versus NO DISPERSION

Matched Four-Way Framework
--------------------------
The shared seed now supports direct comparison among:

1. Standard control.
2. Experiment 34 final-two-row dispersion.
3. Experiment 35 final-row-only dispersion.
4. Experiment 36 penultimate-row-only dispersion.

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Recovery of any Experiment 34 conversions.
6. Rear-boundary-lock frequency.
7. Dependency-chain and stall-family changes.
8. Evidence for isolated versus joint row interaction.

Decision Rule
-------------
- Full recovery of Experiment 34 gains with no regression means the penultimate
  row is the active component.
- No effect means both rows must be included together.
- Partial recovery means the penultimate row contributes strongly but does not
  fully explain the accepted policy.
- Any regression rejects the isolated variant and retains the final-two-row
  policy.

Experimental Hypothesis
-----------------------
Because Experiment 35 removed all benefit, the penultimate row is expected to
be important. Experiment 36 will determine whether it is independently
sufficient or whether the final and penultimate rows must act as a combined
protected region.

Operational Interpretation Boundary
-----------------------------------
This remains a deterministic component-ablation experiment. Operational use
still requires airline-specific calibration and empirical validation.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 35
======================================================================
1. Final-two-row risk dispersion remains the accepted reference candidate.
2. The final-row-only simplification produced 30 equal pairs and no benefit.
3. The penultimate row is materially involved.
4. Experiment 36 isolates the penultimate row.
5. The movement and dependency architecture remains frozen.
6. The branch is now in boundary-definition and closure testing.
7. A finalised research position by Experiment 40 remains realistic.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v36 - EXPERIMENT 36 DESIGN
======================================================================


======================================================================
EXPERIMENT 36 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The matched penultimate-row-only component test executed all 30 paired
scenarios (60 total executions). Both modes shared the Experiment 34-35 seed,
manifests, cabin geometry, occupancy, numerical headway, seat assignments,
aisle assignments, movement rules and dependency architecture. Only the
protected geographic scope differed.

Observed Results
----------------
- Standard mode seated 8,277 of 8,356 passengers.
- Penultimate-row-only dispersion seated 8,284 of
  8,356 passengers.
- Aggregate change: +7 passengers.
- Improved pairs: 1.
- Worse pairs: 0.
- Equal pairs: 29.
- Converted-success cases: 1.
- Longest observed dependency chain: 24 passengers.

Changed Pair
------------
- Pair 19: 335/342 -> 342/342 (+7); converted success=true.

Principal Findings
------------------
1. The penultimate row alone produced one deterministic improvement.
2. Pair 19 converted from 335/342 to 342/342.
3. No paired scenario regressed.
4. Twenty-nine pairs remained equal.
5. The isolated penultimate-row rule recovered seven of the 33 passengers
   gained by the full final-two-row policy.
6. It recovered one of Experiment 34's three converted successes.
7. The final row alone produced no effect in Experiment 35.
8. The full final-two-row region therefore contains both an independent
   penultimate-row contribution and an additional interaction contribution.

Matched Component Comparison
----------------------------
Experiment 34, final-two-row dispersion:
- Aggregate change: +33.
- Improved/worse/equal: 3/0/27.
- Converted successes: 3.

Experiment 35, final-row-only dispersion:
- Aggregate change: 0.
- Improved/worse/equal: 0/0/30.
- Converted successes: 0.

Experiment 36, penultimate-row-only dispersion:
- Aggregate change: +7.
- Improved/worse/equal: 1/0/29.
- Converted successes: 1.

Scientific Verdict
------------------
The penultimate-row-only variant is conditionally beneficial but insufficient
to replace the accepted final-two-row policy.

The evidence supports a mixed mechanism:

1. an independently useful penultimate-row component; and
2. an additional effect that appears only when the final and penultimate rows
   are treated as one protected geographic region.

Discussion Incorporated
------------------------
The programme will not be forced to terminate at Experiment 40. Experiment 40
is now only a possible checkpoint. Testing will continue beyond it whenever
replication, mechanism isolation, adverse-case analysis or holdout evidence is
needed before SME handover.

The handover objective is not simply working code. It is a research package
that is as fully demonstrated as practical, with:

- proven deterministic replay;
- replicated benefits;
- explicit boundary conditions;
- documented rejected variants;
- no hidden regression pattern;
- clear model limitations; and
- operational instructions added only after the scientific evidence is mature.

Experiment 36 Conclusion
------------------------
The penultimate row is independently capable of preventing one rear-boundary
failure, but it explains only part of the full final-two-row benefit.
Experiment 37 therefore tests the cross-row interaction mechanism directly.


======================================================================
EXPERIMENT 37 - CROSS-ROW-ONLY RISK-DISPERSION INTERACTION TEST
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 37
changes only the pre-run condition that triggers stable admission deferral.
Passenger movement, blocker-yield behaviour, seat events, middle-bank
arbitration, dependency diagnostics, numerical entry headway and deterministic
replay remain unchanged.

Planned Purpose
---------------
Determine whether the additional Experiment 34 gains arise specifically from
preventing adjacency between blocker-requiring passengers assigned to different
rows within the final-two-row region.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. CROSS-ROW-ONLY RISK-DISPERSION DATASET.

Experimental Seed
-----------------
    2783790101901

Protected Class
---------------
A passenger belongs to the Experiment 37 protected class when:

1. The passenger is assigned to either of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Selective Cross-Row Rule
------------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- Intervention occurs only when two consecutive protected passengers belong to
  different rows within the final-two-row region.
- The later cross-row protected passenger is deferred until after the next
  ordinary passenger.
- Consecutive protected passengers from the same row remain unchanged.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred passengers is preserved.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variable
--------------------
    CROSS-ROW PROTECTED ADJACENCY DISPERSION:
    ENABLED versus STANDARD SEEDED ORDER

Matched Mechanism Framework
---------------------------
The shared seed supports direct comparison among:

1. Standard control.
2. Full final-two-row dispersion.
3. Final-row-only dispersion.
4. Penultimate-row-only dispersion.
5. Cross-row-only dispersion.

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Recovery of Experiment 34 pairs 1, 18 and 19.
6. Separation of penultimate-only and cross-row effects.
7. Rear-boundary-lock frequency.
8. Dependency-chain and stall-family changes.

Decision Rule
-------------
- Full recovery of Experiment 34 means cross-row adjacency is the principal
  mechanism.
- Recovery of pairs 1 and 18 but not pair 19 indicates separable cross-row and
  penultimate-row components.
- Partial recovery means cross-row adjacency contributes but does not fully
  explain the accepted policy.
- No effect means broader dispersion among all final-two-row protected
  passengers is required.
- Any regression rejects the selective variant and retains the full
  final-two-row policy.

Experimental Hypothesis
-----------------------
Experiment 36 recovered pair 19 only. Experiment 37 may therefore recover the
remaining Experiment 34 conversions if pairs 1 and 18 were caused by direct
final-row/penultimate-row adjacency. A different outcome will refine the
mechanism further rather than invalidate the accepted policy.

Operational Interpretation Boundary
-----------------------------------
This remains a deterministic mechanism-isolation experiment. SME handover and
operational instructions will be prepared only after sufficient replication,
holdout testing, failure analysis and model-boundary documentation.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 36
======================================================================
1. Final-two-row risk dispersion remains the accepted reference candidate.
2. Final-row-only dispersion is ineffective.
3. Penultimate-row-only dispersion is beneficial in one matched scenario.
4. The full policy has an additional unexplained +26 passenger effect.
5. Experiment 37 tests whether cross-row adjacency explains that effect.
6. The architecture remains frozen.
7. Experiment 40 is a checkpoint, not a mandatory stopping point.
8. Research continues until the package is suitably mature for SME handover.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v37 - EXPERIMENT 37 DESIGN
======================================================================


======================================================================
EXPERIMENT 37 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The matched cross-row-only interaction test executed all 30 paired scenarios
(60 total executions). Both modes shared the Experiment 34-36 seed, manifests,
cabin geometry, occupancy, numerical headway, seat assignments, aisle
assignments, movement rules and dependency architecture. Only the selective
cross-row deferral trigger differed.

Observed Results
----------------
- Standard mode seated 8,277 of 8,356 passengers.
- Cross-row-only dispersion seated 8,303 of
  8,356 passengers.
- Aggregate change: +26 passengers.
- Improved pairs: 2.
- Worse pairs: 0.
- Equal pairs: 28.
- Converted-success cases: 2.
- Longest observed dependency chain: 24 passengers.

Changed Pairs
-------------
- Pair 01: 340/342 -> 342/342 (+2); converted success=true.
- Pair 18: 264/288 -> 288/288 (+24); converted success=true.

Principal Findings
------------------
1. Cross-row-only dispersion converted Pair 01 from 340/342 to 342/342.
2. Cross-row-only dispersion converted Pair 18 from 264/288 to 288/288.
3. Aggregate improvement was +26 passengers.
4. No paired scenario regressed.
5. Twenty-eight pairs remained equal.
6. Experiment 36 independently recovered Pair 19 and +7 passengers.
7. The isolated Experiment 36 and 37 gains sum exactly to the complete
   Experiment 34 gain: +7 + +26 = +33.
8. The three converted successes also partition exactly:
   - penultimate-row-only mechanism: Pair 19;
   - cross-row mechanism: Pairs 01 and 18.
9. The accepted final-two-row policy is therefore explained by two separable
   deterministic mechanisms rather than an unexplained aggregate heuristic.

Mechanism Decomposition
-----------------------
Component A - penultimate-row stabilisation:
- Experiment 36.
- +7 passengers.
- One converted success.
- Pair 19.

Component B - cross-row adjacency suppression:
- Experiment 37.
- +26 passengers.
- Two converted successes.
- Pairs 01 and 18.

Combined reconstruction:
- +33 passengers.
- Three converted successes.
- Exact match to Experiment 34.

Scientific Verdict
------------------
Experiment 37 confirms that cross-row adjacency is a major causal component of
the accepted final-two-row policy.

The full final-two-row rule remains the preferred reference because it naturally
combines both validated mechanisms in one conservative deterministic policy.

This result is stronger than merely showing that the policy works. It explains
why the matched Experiment 34 result occurred and identifies which submechanism
resolved each converted-success case.

Discussion Incorporated
------------------------
The research programme now moves from mechanism discovery into holdout
validation.

Experiment 40 is not a forced endpoint. Further experiments will be performed
whenever necessary to establish:

- independent-seed generalisation;
- repeatability of benefits;
- absence of hidden regressions;
- behaviour across cabin configurations and occupancy conditions;
- limits of the mechanism decomposition; and
- evidence suitable for later SME review.

Instructions for advantageous operational use will be written only after the
policy has completed sufficient holdout, adverse-case and limitation testing.

Experiment 37 Conclusion
------------------------
The complete Experiment 34 gain has now been exactly decomposed into a
penultimate-row component and a cross-row component. Experiment 38 restores the
full accepted policy and tests it on a new, previously unused deterministic
holdout seed.


======================================================================
EXPERIMENT 38 - FINAL-TWO-ROW RISK-DISPERSION HOLDOUT REPLICATION I
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 38
changes only the deterministic passenger admission ordering. Passenger
movement, blocker-yield behaviour, seat events, middle-bank arbitration,
dependency diagnostics, numerical entry headway and deterministic replay remain
unchanged.

Planned Purpose
---------------
Test whether the accepted final-two-row risk-dispersion policy generalises to a
new deterministic sample that was not used for discovery, replication or
mechanism decomposition.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Holdout Seed
------------
    3783790101901

This seed is independent of the Experiment 34-37 matched mechanism sample.

Protected Class
---------------
A passenger belongs to the Experiment 38 protected class when:

1. The passenger is assigned to either of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variable
--------------------
    COMPLETE FINAL-TWO-ROW RISK DISPERSION:
    ENABLED versus STANDARD SEEDED ORDER

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Rear-boundary-lock frequency.
6. Cabin-configuration distribution.
7. Occupancy and headway distribution.
8. Dependency-chain and stall-family changes.
9. Any evidence of regression.
10. Comparison with the pooled Experiments 32-34 evidence.

Decision Rule
-------------
- Positive holdout results with no regression strengthen generalisation.
- Neutral results remain compatible with a rare but safe intervention.
- Any regression requires detailed case analysis before the policy can remain
  an SME-facing reference.
- One holdout experiment is not sufficient for closure; further independent
  seeds will be added as scientifically required.

Experimental Hypothesis
-----------------------
The accepted policy is expected to remain conservative: most paired scenarios
should remain equal, occasional rear-boundary failures may convert to success,
and regressions should remain absent. The exact frequency of improvements is
not assumed in advance.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It does not yet constitute a live
airline recommendation. SME-facing instructions will follow only after
independent holdout replication, adverse-case testing and limitation
documentation are sufficiently mature.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 37
======================================================================
1. Final-two-row risk dispersion remains the accepted reference policy.
2. Its matched-sample gain has been exactly decomposed.
3. Penultimate-row stabilisation explains +7 passengers.
4. Cross-row adjacency suppression explains +26 passengers.
5. Together they reconstruct the complete +33 result.
6. No regression occurred in Experiments 32-37.
7. Experiment 38 begins independent holdout validation.
8. Experiment 40 remains a checkpoint rather than a termination requirement.
9. SME usage instructions remain deferred until evidence is mature.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v38 - EXPERIMENT 38 DESIGN
======================================================================


======================================================================
EXPERIMENT 38 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The first independent holdout replication executed all 30 paired scenarios
(60 total executions). Both modes shared the new holdout seed, manifests, cabin
geometry, occupancy, numerical headway, seat assignments, aisle assignments,
movement rules and dependency architecture. Only the final-two-row admission
dispersion rule differed.

Observed Results
----------------
- Standard mode seated 7,659 of 7,920 passengers.
- Final-two-row dispersion seated 7,665 of
  7,920 passengers.
- Aggregate change: +6 passengers.
- Improved pairs: 1.
- Worse pairs: 0.
- Equal pairs: 29.
- Converted-success cases: 1.
- Longest observed dependency chain: 27 passengers.

Changed Pair
------------
- Pair 29: 229/235 -> 235/235 (+6); converted success=true.

Principal Findings
------------------
1. The first unseen deterministic holdout produced one converted success.
2. Pair 29 improved from 229/235 to 235/235.
3. Aggregate improvement was +6 passengers.
4. No paired scenario regressed.
5. Twenty-nine pairs remained equal.
6. The policy therefore retained its conservative empirical signature:
   rare intervention, occasional complete recovery and no observed harm.
7. The result extends positive evidence beyond the discovery and
   mechanism-decomposition seed.
8. The holdout does not establish universal performance, but it materially
   strengthens generalisation.

Pooled Positive Evidence
------------------------
Experiments 32-34:
- 90 paired scenarios.
- Aggregate improvement: +69.
- Improved/worse/equal: 5/0/85.

Experiment 38:
- 30 independent holdout pairs.
- Aggregate improvement: +6.
- Improved/worse/equal: 1/0/29.

Combined accepted-policy evidence:
- 120 paired scenarios.
- Aggregate improvement: +75.
- Improved pairs: 6.
- Worse pairs: 0.
- Equal pairs: 114.
- Observed regression rate: 0/120.

Scientific Verdict
------------------
Experiment 38 is a successful first holdout replication.

The magnitude is smaller than Experiment 34, but magnitude equality was not the
test. The important findings are:

- the benefit appeared on an unseen seed;
- the changed case converted fully;
- no regression appeared; and
- the intervention remained sparse rather than broadly disruptive.

This supports continued validation of the accepted final-two-row policy.

Discussion Incorporated
------------------------
The research programme remains evidence-led rather than phase-count-led.
Experiment 40 is only a checkpoint. Additional independent seeds, adverse cases
or boundary tests will be added whenever necessary before SME handover.

Operational instructions will be prepared only after the evidence package
contains sufficient replication, known limitations, rejected variants and
clear guidance on what the code can and cannot support.

Experiment 38 Conclusion
------------------------
The accepted final-two-row policy generalised positively to its first
independent holdout seed, producing +6 passengers and one converted success
without regression. Experiment 39 performs a second independent holdout
replication using the unchanged accepted policy.


======================================================================
EXPERIMENT 39 - FINAL-TWO-ROW RISK-DISPERSION HOLDOUT REPLICATION II
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 39
changes only the deterministic passenger admission ordering. Passenger
movement, blocker-yield behaviour, seat events, middle-bank arbitration,
dependency diagnostics, numerical entry headway and deterministic replay remain
unchanged.

Planned Purpose
---------------
Determine whether the accepted final-two-row policy retains its conservative
positive pattern on a second previously unused deterministic holdout seed.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Holdout Seed
------------
    4783790101901

This seed is independent of the discovery, mechanism-decomposition and first
holdout samples.

Protected Class
---------------
A passenger belongs to the Experiment 39 protected class when:

1. The passenger is assigned to either of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variable
--------------------
    COMPLETE FINAL-TWO-ROW RISK DISPERSION:
    ENABLED versus STANDARD SEEDED ORDER

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Rear-boundary-lock frequency.
6. Cabin-configuration distribution.
7. Occupancy and headway distribution.
8. Dependency-chain and stall-family changes.
9. Any evidence of regression.
10. Comparison with Experiments 32-34 and Holdout I.

Decision Rule
-------------
- A positive second holdout with no regression strengthens generalisation.
- A neutral result remains compatible with a rare, conservative intervention.
- Any regression requires detailed case analysis before SME-facing acceptance.
- Experiment 40 remains a checkpoint, not a compulsory endpoint.

Experimental Hypothesis
-----------------------
Most paired scenarios should remain equal. Occasional rear-boundary failures may
convert to success. The exact number and magnitude of improvements are not
assumed in advance. The central safety question remains whether any regression
appears.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It does not yet constitute a live
airline recommendation. SME-facing instructions will follow only after
independent holdout replication, adverse-case testing and limitation
documentation are sufficiently mature.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 38
======================================================================
1. Final-two-row risk dispersion remains the accepted reference policy.
2. Its matched-sample mechanism has been decomposed exactly.
3. The first independent holdout produced +6 and one converted success.
4. No regression has appeared across 120 accepted-policy pairs.
5. Experiment 39 performs a second independent holdout replication.
6. Amit's movement and dependency architecture remains frozen.
7. Experiment 40 remains a checkpoint rather than a termination requirement.
8. SME usage instructions remain deferred until evidence is mature.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v39 - EXPERIMENT 39 DESIGN
======================================================================


======================================================================
EXPERIMENT 39 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The second independent holdout replication executed all 30 paired scenarios
(60 total executions). Both modes shared the new holdout seed, manifests, cabin
geometry, occupancy, numerical headway, seat assignments, aisle assignments,
movement rules and dependency architecture. Only the final-two-row admission
dispersion rule differed.

Observed Results
----------------
- Standard mode seated 8,012 of 8,099 passengers.
- Final-two-row dispersion seated 8,083 of
  8,099 passengers.
- Aggregate change: +71 passengers.
- Improved pairs: 3.
- Worse pairs: 0.
- Equal pairs: 27.
- Converted-success cases: 3.
- Longest observed dependency chain: 22 passengers.

Changed Pairs
-------------
- Pair 01: 228/230 -> 230/230 (+2); converted success=true.
- Pair 18: 191/216 -> 216/216 (+25); converted success=true.
- Pair 25: 309/353 -> 353/353 (+44); converted success=true.

Principal Findings
------------------
1. The second unseen deterministic holdout produced three converted successes.
2. Pair 01 improved from 228/230 to 230/230.
3. Pair 18 improved from 191/216 to 216/216.
4. Pair 25 improved from 309/353 to 353/353.
5. Aggregate improvement was +71 passengers.
6. No paired scenario regressed.
7. Twenty-seven pairs remained equal.
8. The accepted policy therefore generalised positively on a second independent
   seed and retained its conservative no-regression signature.
9. The stronger magnitude compared with Experiment 38 demonstrates that holdout
   benefit is sample-dependent; the policy is not tied to one fixed gain size.

Pooled Accepted-Policy Evidence
-------------------------------
Experiments 32-34:
- 90 paired scenarios.
- Aggregate improvement: +69.
- Improved/worse/equal: 5/0/85.

Experiment 38:
- 30 independent holdout pairs.
- Aggregate improvement: +6.
- Improved/worse/equal: 1/0/29.

Experiment 39:
- 30 independent holdout pairs.
- Aggregate improvement: +71.
- Improved/worse/equal: 3/0/27.

Combined accepted-policy evidence:
- 150 paired scenarios.
- Aggregate improvement: +146.
- Improved pairs: 9.
- Worse pairs: 0.
- Equal pairs: 141.
- Observed regression rate: 0/150.
- Converted-success cases recorded across discovery and holdout stages.

Scientific Verdict
------------------
Experiment 39 is a strong second independent holdout replication.

The important result is not merely the +71 total. The key evidence is that:

- three incomplete standard scenarios converted fully;
- the policy remained unchanged;
- the seed was independent;
- most scenarios remained equal; and
- no regression appeared.

Together with Experiment 38, this materially strengthens the case that the
accepted final-two-row rule generalises beyond its discovery sample.

Discussion Incorporated
------------------------
Experiment 40 is retained as a checkpoint rather than a forced endpoint.
The programme will continue whenever the evidence still requires:

- higher-pressure occupancy testing;
- adverse-case isolation;
- limitation documentation;
- further independent seeds;
- or SME-facing translation.

The next experiment therefore increases environmental pressure while leaving
the accepted policy and Amit's movement architecture unchanged.

Experiment 39 Conclusion
------------------------
The accepted final-two-row policy produced +71 passengers and three converted
successes on its second independent holdout seed, with no regression.
Experiment 40 tests the same policy under deliberately high occupancy.


======================================================================
EXPERIMENT 40 - HIGH-OCCUPANCY FINAL-TWO-ROW CHALLENGE
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 40
changes only the deterministic passenger admission ordering in the tested mode
and the scenario sampling pressure. Passenger movement, blocker-yield behaviour,
seat events, middle-bank arbitration, dependency diagnostics, numerical entry
headway and deterministic replay remain unchanged.

Planned Purpose
---------------
Test whether the accepted final-two-row risk-dispersion policy remains safe and
useful when every paired scenario is drawn from a high-occupancy cabin.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Challenge Seed
--------------
    5783790101901

Occupancy Challenge
-------------------
All scenarios use one of the following occupancy levels:

- 95%
- 98%

The lower 80%, 85% and 90% occupancy bands are intentionally excluded.

This increases expected aisle density, blocker interaction and rear-boundary
exposure without changing the movement engine.

Protected Class
---------------
A passenger belongs to the Experiment 40 protected class when:

1. The passenger is assigned to either of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variables
---------------------
1. FINAL-TWO-ROW RISK DISPERSION:
   ENABLED versus STANDARD SEEDED ORDER.

2. ENVIRONMENTAL PRESSURE:
   ALL SCENARIOS RESTRICTED TO 95% OR 98% OCCUPANCY.

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Regression detection.
6. Rear-boundary-lock frequency.
7. Linked-row-event-chain frequency.
8. Cabin-configuration distribution.
9. Headway distribution.
10. Dependency-chain and stall-family changes.
11. Comparison with mixed-occupancy Experiments 38 and 39.

Decision Rule
-------------
- Positive high-occupancy evidence with no regression supports robustness.
- Neutral evidence with no regression remains compatible with a sparse,
  conservative intervention.
- Any regression requires case-level investigation before SME-facing use.
- Experiment 40 does not terminate the programme automatically.

Experimental Hypothesis
-----------------------
High occupancy should increase the number of difficult rear-cabin states.
The accepted policy may therefore show more opportunities to intervene, but the
exact number and magnitude of improvements are not assumed in advance.

The central safety question remains whether any scenario becomes worse.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It is not a live-airline boarding
instruction. Calibrated operational data, real passenger behaviour and SME
review remain mandatory before practical deployment.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 39
======================================================================
1. Final-two-row risk dispersion remains the accepted reference policy.
2. Its mechanism has been decomposed into two deterministic components.
3. Two independent holdouts have now produced positive results.
4. Experiment 39 delivered +71 and three converted successes.
5. No regression has appeared across 150 accepted-policy pairs.
6. Experiment 40 increases occupancy pressure without changing the policy.
7. Amit's movement and dependency architecture remains frozen.
8. Experiment 40 is a checkpoint, not a forced termination point.
9. SME usage instructions remain deferred until evidence and limitations are
   sufficiently mature.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v40 - EXPERIMENT 40 DESIGN
======================================================================


======================================================================
EXPERIMENT 40 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The high-occupancy challenge executed all 30 paired scenarios (60 total
executions). Every scenario used either 95% or 98% occupancy. Both paired modes
shared the same seed, cabin, passenger manifest, seat assignment, aisle
assignment, numerical headway, movement rules and dependency architecture.
Only the final-two-row admission-dispersion rule differed.

Observed Results
----------------
- Standard mode seated 8,476 of 8,627 passengers.
- Final-two-row dispersion seated 8,583 of
  8,627 passengers.
- Aggregate change: +107 passengers.
- Improved pairs: 3.
- Worse pairs: 0.
- Equal pairs: 27.
- Converted-success cases: 3.
- Longest observed dependency chain: 16 passengers.

Changed Pairs
-------------
- Pair 03: 268/282 -> 282/282 (+14); converted success=true.
- Pair 27: 253/342 -> 342/342 (+89); converted success=true.
- Pair 29: 338/342 -> 342/342 (+4); converted success=true.

Principal Findings
------------------
1. All three changed cases converted from incomplete cabins to complete cabins.
2. Pair 03 improved from 268/282 to 282/282.
3. Pair 27 improved from 253/342 to 342/342.
4. Pair 29 improved from 338/342 to 342/342.
5. Aggregate improvement was +107 passengers.
6. No paired scenario regressed.
7. Twenty-seven pairs remained equal.
8. The accepted policy therefore remained safe and strongly effective under
   deliberately dense 95%-98% occupancy conditions.
9. Pair 27 is the largest single recovery observed in this validation stage,
   recovering 89 passengers.
10. The strongest benefits continued to arise from deterministic rear-boundary
    and linked dependency conditions rather than broad changes to ordinary
    completed cabins.

Pooled Accepted-Policy Evidence
-------------------------------
Experiments 32-34:
- 90 paired scenarios.
- Aggregate improvement: +69.
- Improved/worse/equal: 5/0/85.

Experiment 38:
- 30 independent holdout pairs.
- Aggregate improvement: +6.
- Improved/worse/equal: 1/0/29.

Experiment 39:
- 30 independent holdout pairs.
- Aggregate improvement: +71.
- Improved/worse/equal: 3/0/27.

Experiment 40:
- 30 high-occupancy challenge pairs.
- Aggregate improvement: +107.
- Improved/worse/equal: 3/0/27.

Combined accepted-policy evidence:
- 180 paired scenarios.
- Aggregate improvement: +253.
- Improved pairs: 12.
- Worse pairs: 0.
- Equal pairs: 168.
- Observed regression rate: 0/180.

Scientific Verdict
------------------
Experiment 40 is a successful adverse-condition validation checkpoint.

The result is important because the environmental pressure was increased while
the accepted policy and movement architecture remained unchanged. The policy
continued to show a sparse intervention pattern, but when it acted it converted
three difficult failures completely. No harm was observed in the remaining 27
pairs.

This supports isolating the most demanding 98% occupancy band in Experiment 41.

Discussion Incorporated
------------------------
Experiment 40 is not treated as a forced stopping point. The evidence programme
continues because the next scientifically useful question is whether the
positive no-regression pattern survives when every scenario is held at the
maximum occupancy band currently represented in the model.

SME-facing instructions remain deferred until the evidence package includes
sufficient independent replication, adverse-condition validation, limitation
documentation and clear operational boundaries.

Experiment 40 Conclusion
------------------------
The accepted final-two-row policy produced +107 passengers and three converted
successes under 95%-98% occupancy, with no regression. Experiment 41 isolates
98% occupancy only.


======================================================================
EXPERIMENT 41 - EXTREME-OCCUPANCY FINAL-TWO-ROW CHALLENGE
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 41
changes only the deterministic passenger admission ordering in the tested mode
and restricts scenario sampling to 98% occupancy. Passenger movement,
blocker-yield behaviour, seat events, middle-bank arbitration, dependency
diagnostics, numerical entry headway and deterministic replay remain unchanged.

Planned Purpose
---------------
Test whether the accepted final-two-row risk-dispersion policy remains safe and
useful when every paired scenario is executed at 98% cabin occupancy.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Challenge Seed
--------------
    6783790101901

Occupancy Challenge
-------------------
All scenarios use:

    98% OCCUPANCY

The 95% band used in Experiment 40 is intentionally removed.

This isolates the densest occupancy condition currently represented in the
research model.

Protected Class
---------------
A passenger belongs to the Experiment 41 protected class when:

1. The passenger is assigned to either of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variables
---------------------
1. FINAL-TWO-ROW RISK DISPERSION:
   ENABLED versus STANDARD SEEDED ORDER.

2. OCCUPANCY PRESSURE:
   FIXED AT 98%.

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success cases.
5. Regression detection.
6. Rear-boundary-lock frequency.
7. Linked-row-event-chain frequency.
8. Cabin-configuration distribution.
9. Numerical headway distribution.
10. Dependency-chain and stall-family changes.
11. Comparison with the mixed 95%-98% Experiment 40 challenge.

Decision Rule
-------------
- Positive 98%-occupancy evidence with no regression supports extreme-load
  robustness.
- Neutral evidence with no regression remains compatible with a sparse,
  conservative intervention.
- Any regression requires case-level investigation before SME-facing use.
- Experiment 41 does not terminate the wider evidence programme automatically.

Experimental Hypothesis
-----------------------
Fixing occupancy at 98% should increase exposure to severe rear-cabin
dependency states. The exact number and magnitude of improvements are not
assumed. The primary safety test remains whether any paired scenario becomes
worse.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It is not a live-airline boarding
instruction. Calibrated operational data, real passenger behaviour and SME
review remain mandatory before practical deployment.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 40
======================================================================
1. Final-two-row risk dispersion remains the accepted reference policy.
2. Two ordinary holdouts and one high-occupancy challenge are positive.
3. Experiment 40 delivered +107 and three converted successes.
4. No regression has appeared across 180 accepted-policy pairs.
5. Experiment 41 isolates the 98% occupancy band.
6. Amit's movement and dependency architecture remains frozen.
7. The programme remains evidence-led rather than phase-count-led.
8. SME usage instructions remain deferred until evidence and limitations are
   sufficiently mature.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v41 - EXPERIMENT 41 DESIGN
======================================================================


======================================================================
EXPERIMENT 41 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The extreme-occupancy challenge executed all 30 paired scenarios (60 total
executions). Every scenario was fixed at 98% occupancy. Both paired modes shared
the same seed, cabin, manifest, seat assignment, aisle assignment, numerical
headway, movement rules and dependency architecture. Only the accepted
final-two-row admission-dispersion rule differed.

Observed Results
----------------
- Standard mode seated 8,532 of 8,723 passengers.
- Final-two-row dispersion seated 8,604 of
  8,723 passengers.
- Aggregate change: +72 passengers.
- Improved pairs: 2.
- Worse pairs: 0.
- Equal pairs: 28.
- Converted-success cases: 1.
- Longest observed dependency chain: 28 passengers.

Changed Pairs
-------------
- Pair 27: 234/282 -> 282/282 (+48); converted success=true.
- Pair 29: 208/235 -> 232/235 (+24); converted success=false.

Principal Findings
------------------
1. Pair 27 improved from 234/282 to 282/282, recovering 48 passengers and
   converting the scenario to complete success.
2. Pair 29 improved from 208/235 to 232/235, recovering 24 passengers while
   remaining incomplete.
3. Aggregate improvement was +72 passengers.
4. No paired scenario regressed.
5. Twenty-eight pairs remained equal.
6. The accepted policy therefore retained its conservative no-regression
   signature at the maximum occupancy band currently represented.
7. The residual incomplete cases continued to expose severe rear-boundary locks,
   linked row-event chains and insufficient rear yield space.
8. Experiment 41 demonstrates that the policy can materially reduce an extreme
   stall without necessarily resolving every residual boundary condition.
9. The longest observed dependency chain reached 28 passengers, confirming that
   the 98% challenge generated unusually deep deterministic blocking structures.

Comparison with Experiment 40
-----------------------------
Experiment 40 used a mixed 95%-98% challenge:

- Aggregate improvement: +107.
- Improved/worse/equal: 3/0/27.
- Converted successes: 3.

Experiment 41 used 98% occupancy only:

- Aggregate improvement: +72.
- Improved/worse/equal: 2/0/28.
- Converted successes: 1.

The lower gain does not constitute a negative result. Experiment 41 removed the
less extreme 95% cases and exposed more severe residual constraints. The key
safety result remained unchanged: no regression occurred.

Pooled Accepted-Policy Evidence
-------------------------------
Experiments 32-34:
- 90 paired scenarios.
- Aggregate improvement: +69.
- Improved/worse/equal: 5/0/85.

Experiment 38:
- 30 independent holdout pairs.
- Aggregate improvement: +6.
- Improved/worse/equal: 1/0/29.

Experiment 39:
- 30 independent holdout pairs.
- Aggregate improvement: +71.
- Improved/worse/equal: 3/0/27.

Experiment 40:
- 30 high-occupancy challenge pairs.
- Aggregate improvement: +107.
- Improved/worse/equal: 3/0/27.

Experiment 41:
- 30 extreme-occupancy pairs.
- Aggregate improvement: +72.
- Improved/worse/equal: 2/0/28.

Combined accepted-policy evidence:
- 210 paired scenarios.
- Aggregate improvement: +325.
- Improved pairs: 14.
- Worse pairs: 0.
- Equal pairs: 196.
- Observed regression rate: 0/210.

Scientific Verdict
------------------
Experiment 41 is a successful extreme-load validation experiment.

It confirms that the accepted policy remains beneficial and non-regressive at
98% occupancy, while also revealing an important limitation: some severe
rear-boundary and linked-event stalls remain only partially recoverable.

That limitation should be documented rather than hidden. A conclusive research
programme must identify both the policy's reliable benefit and the boundary of
what it does not solve.

Discussion Incorporated
------------------------
The next approximately twenty experiments are treated as a confidence-building
horizon rather than a promise of a predetermined conclusion. The objective is
to pursue independent replication, adverse-case testing, limitation analysis
and eventual SME preparation until the evidence becomes genuinely conclusive.

Experiment 42 therefore does not introduce another mechanism. It repeats the
98% challenge on a new unseen deterministic seed to test generalisation.

Experiment 41 Conclusion
------------------------
The accepted policy produced +72 passengers at fixed 98% occupancy, including
one complete conversion and one substantial partial recovery, with no
regression. Experiment 42 performs an independent 98% holdout replication.


======================================================================
EXPERIMENT 42 - EXTREME-OCCUPANCY HOLDOUT REPLICATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 42
changes only the deterministic experiment seed. The accepted admission policy,
98% occupancy condition, passenger movement, blocker-yield behaviour, seat
events, middle-bank arbitration, dependency diagnostics and numerical headway
remain unchanged.

Planned Purpose
---------------
Determine whether the Experiment 41 extreme-occupancy result generalises to a
new unseen deterministic sample.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Holdout Seed
------------
    7783790101901

Occupancy Condition
-------------------
All scenarios remain fixed at:

    98% OCCUPANCY

Protected Class
---------------
A passenger belongs to the Experiment 42 protected class when:

1. The passenger is assigned to either of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variable
--------------------
FINAL-TWO-ROW RISK DISPERSION:

    ENABLED versus STANDARD SEEDED ORDER

The occupancy pressure and policy are fixed. The experiment seed is new.

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success and partial-recovery cases.
5. Regression detection.
6. Rear-boundary-lock frequency.
7. Linked-row-event-chain frequency.
8. Cabin-configuration distribution.
9. Numerical headway distribution.
10. Dependency-chain depth.
11. Comparison with Experiment 41.

Decision Rule
-------------
- Positive evidence with no regression supports extreme-occupancy replication.
- Neutral evidence with no regression remains compatible with a sparse policy.
- Any regression requires full case-level isolation.
- Residual incomplete cases must be documented as limitations.
- Experiment 42 does not terminate the wider evidence programme automatically.

Experimental Hypothesis
-----------------------
The accepted policy should continue to produce sparse, sample-dependent benefits
without making any paired scenario worse. The exact gain and number of converted
successes are not assumed in advance.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It is not a live-airline boarding
instruction. Empirical calibration, behavioural data and SME review remain
mandatory before practical deployment.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 41
======================================================================
1. Final-two-row risk dispersion remains the accepted reference policy.
2. Two ordinary holdouts, one high-occupancy challenge and one fixed-98%
   challenge have produced positive evidence.
3. Experiment 41 delivered +72, one converted success and one partial recovery.
4. No regression has appeared across 210 accepted-policy pairs.
5. Experiment 41 also exposed a valid limitation: some extreme rear-boundary
   stalls remain incomplete.
6. Experiment 42 tests independent replication at the same 98% pressure.
7. Amit's movement and dependency architecture remains frozen.
8. The next research horizon is confidence-building rather than forced closure.
9. SME usage instructions remain deferred until evidence and limitations are
   sufficiently mature.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v42 - EXPERIMENT 42 DESIGN
======================================================================


======================================================================
EXPERIMENT 42 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The independent extreme-occupancy holdout executed all 30 paired scenarios
(60 total executions). Every scenario remained fixed at 98% occupancy. Both
paired modes shared the same seed, cabin, passenger manifest, seat assignment,
aisle assignment, numerical headway, movement rules and dependency architecture.
Only the accepted final-two-row admission-dispersion rule differed.

Observed Results
----------------
- Standard mode seated 8,277 of 8,700 passengers.
- Final-two-row dispersion seated 8,320 of
  8,700 passengers.
- Aggregate change: +43 passengers.
- Improved pairs: 3.
- Worse pairs: 0.
- Equal pairs: 27.
- Converted-success cases: 2.
- Longest observed dependency chain: 19 passengers.

Changed Pairs
-------------
- Pair 17: 278/282 -> 282/282 (+4); converted success=true.
- Pair 22: 269/282 -> 282/282 (+13); converted success=true.
- Pair 26: 313/353 -> 339/353 (+26); converted success=false.

Principal Findings
------------------
1. Pair 17 improved from 278/282 to 282/282, recovering four passengers and
   converting the cabin to complete success.
2. Pair 22 improved from 269/282 to 282/282, recovering thirteen passengers and
   converting the cabin to complete success.
3. Pair 26 improved from 313/353 to 339/353, recovering twenty-six passengers
   while remaining incomplete.
4. Aggregate improvement was +43 passengers.
5. No paired scenario regressed.
6. Twenty-seven pairs remained equal.
7. The accepted policy therefore replicated positively on a new unseen
   deterministic sample at fixed 98% occupancy.
8. Two of the three changed pairs became complete cabins.
9. The residual incomplete case demonstrates that the policy can materially
   reduce a severe stall without resolving every downstream dependency.
10. The longest observed dependency chain reached 19 passengers.

Comparison with Experiment 41
-----------------------------
Experiment 41:
- Aggregate improvement: +72.
- Improved/worse/equal: 2/0/28.
- Converted successes: 1.

Experiment 42:
- Aggregate improvement: +43.
- Improved/worse/equal: 3/0/27.
- Converted successes: 2.

The aggregate magnitude varied across seeds, but the directional result
replicated: sparse improvement, complete conversions and zero regression.

Pooled Accepted-Policy Evidence
-------------------------------
Experiments 32-34:
- 90 paired scenarios.
- Aggregate improvement: +69.
- Improved/worse/equal: 5/0/85.

Experiment 38:
- 30 independent holdout pairs.
- Aggregate improvement: +6.
- Improved/worse/equal: 1/0/29.

Experiment 39:
- 30 independent holdout pairs.
- Aggregate improvement: +71.
- Improved/worse/equal: 3/0/27.

Experiment 40:
- 30 high-occupancy challenge pairs.
- Aggregate improvement: +107.
- Improved/worse/equal: 3/0/27.

Experiment 41:
- 30 extreme-occupancy pairs.
- Aggregate improvement: +72.
- Improved/worse/equal: 2/0/28.

Experiment 42:
- 30 independent extreme-occupancy holdout pairs.
- Aggregate improvement: +43.
- Improved/worse/equal: 3/0/27.

Combined accepted-policy evidence:
- 240 paired scenarios.
- Aggregate improvement: +368.
- Improved pairs: 17.
- Worse pairs: 0.
- Equal pairs: 223.
- Observed regression rate: 0/240.

Scientific Verdict
------------------
Experiment 42 is a successful independent replication.

The significance is not merely the +43 gain. The accepted policy reproduced the
same qualitative signature on a new extreme-occupancy sample:

- sparse intervention,
- material recovery when activated,
- complete conversions,
- no detected harm.

This strengthens confidence that the Experiment 41 result was not peculiar to
one deterministic seed.

Discussion Incorporated
------------------------
The programme remains directed toward a genuinely conclusive evidence package
over the next research horizon. Conclusiveness will depend on replication,
adverse-case coverage, explicit limitation analysis and eventual SME review,
not simply reaching a predetermined experiment number.

Experiment 43 therefore repeats the 98% holdout design on another unseen seed
without adding a new mechanism.

Experiment 42 Conclusion
------------------------
The accepted policy produced +43 passengers, three improved pairs and two
converted successes on an independent 98%-occupancy holdout, with no regression.
Experiment 43 performs a second independent replication.


======================================================================
EXPERIMENT 43 - EXTREME-OCCUPANCY HOLDOUT REPLICATION II
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 43
changes only the deterministic experiment seed. The accepted admission policy,
98% occupancy condition, passenger movement, blocker-yield behaviour, seat
events, middle-bank arbitration, dependency diagnostics and numerical headway
remain unchanged.

Planned Purpose
---------------
Determine whether the positive extreme-occupancy result persists across a third
independent 98%-occupancy deterministic sample.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Holdout Seed
------------
    8783790101901

Occupancy Condition
-------------------
All scenarios remain fixed at:

    98% OCCUPANCY

Protected Class
---------------
A passenger belongs to the Experiment 43 protected class when:

1. The passenger is assigned to either of the final two cabin rows.
2. The assigned seat requires at least one blocker to yield.

Stable Dispersion Rule
----------------------
- A seeded shuffle establishes the ordinary deterministic aisle order.
- When two protected passengers would be consecutive, the later protected
  passenger is deferred until after the next ordinary passenger.
- Relative order among ordinary passengers is preserved.
- Relative order among deferred protected passengers is preserved.
- Remaining deferred passengers are appended only when no ordinary passengers
  remain.
- Independent left/right queues remain active.
- Ordinary numerical entry headway remains active.
- No runtime adaptation is introduced.

Independent Variable
--------------------
FINAL-TWO-ROW RISK DISPERSION:

    ENABLED versus STANDARD SEEDED ORDER

Only the deterministic holdout seed changes from Experiment 42.

Primary Evaluation Order
------------------------
1. Deterministic replay.
2. Improved, worse and equal pair counts.
3. Aggregate seated-passenger change.
4. Converted-success and partial-recovery cases.
5. Regression detection.
6. Rear-boundary-lock frequency.
7. Linked-row-event-chain frequency.
8. Moving-passenger-yield-chain frequency.
9. Cabin-configuration distribution.
10. Numerical headway distribution.
11. Dependency-chain depth.
12. Comparison with Experiments 41 and 42.

Decision Rule
-------------
- Positive evidence with no regression strengthens three-seed replication.
- Neutral evidence with no regression remains compatible with a sparse policy.
- Any regression requires full case-level isolation.
- Residual incomplete cases must remain visible as limitations.
- Experiment 43 does not terminate the wider evidence programme automatically.

Experimental Hypothesis
-----------------------
The accepted policy should continue to produce sample-dependent improvements
without making any paired scenario worse. The exact aggregate gain, number of
improved pairs and number of converted successes are not assumed in advance.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It is not a live-airline boarding
instruction. Empirical calibration, behavioural data and SME review remain
mandatory before practical deployment.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 42
======================================================================
1. Final-two-row risk dispersion remains the accepted reference policy.
2. The fixed-98% result has now replicated on an independent seed.
3. Experiment 42 delivered +43, three improvements and two converted successes.
4. No regression has appeared across 240 accepted-policy pairs.
5. Aggregate magnitude varies by seed, while directional safety remains stable.
6. Residual incomplete cases remain valid limitations rather than hidden
   failures.
7. Experiment 43 performs a second independent 98%-occupancy replication.
8. Amit's movement and dependency architecture remains frozen.
9. The research horizon remains confidence-building rather than phase-count-led.
10. SME usage instructions remain deferred until evidence and limitations are
    sufficiently mature.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v43 - EXPERIMENT 43 DESIGN
======================================================================


======================================================================
EXPERIMENT 43 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
The second independent extreme-occupancy holdout executed all 30 paired
scenarios (60 total executions). Every scenario remained fixed at 98%
occupancy. Both paired modes shared the same deterministic seed, cabin,
manifest, seat assignment, aisle assignment, numerical headway, movement
rules and dependency architecture. Only the accepted final-two-row
admission-dispersion rule differed.

Observed Results
----------------
- Standard mode seated 8,550 of 8,841 passengers.
- Final-two-row dispersion seated 8,662 of
  8,841 passengers.
- Aggregate change: +112 passengers.
- Improved pairs: 2.
- Worse pairs: 1.
- Equal pairs: 27.
- Converted-success cases: 0.
- Longest observed dependency chain: 25 passengers.

Changed Pairs
-------------
- Pair 02: 302/353 -> 347/353 (+45); ENTRY_SATURATION_WITH_INTERNAL_CHAIN -> MIXED_OR_UNCLASSIFIED_RESIDUAL.
- Pair 14: 350/353 -> 348/353 (-2); LINKED_ROW_EVENT_CHAIN -> REAR_BOUNDARY_LOCK.
- Pair 22: 277/353 -> 346/353 (+69); ENTRY_SATURATION_WITH_INTERNAL_CHAIN -> MIXED_OR_UNCLASSIFIED_RESIDUAL.

Principal Findings
------------------
1. Pair 02 improved from 302/353 to 347/353, recovering 45 passengers.
2. Pair 22 improved from 277/353 to 346/353, recovering 69 passengers.
3. Pair 14 regressed from 350/353 to 348/353, losing two passengers.
4. Aggregate improvement remained strongly positive at +112 passengers.
5. The result comprised two improvements, one regression and twenty-seven
   equal pairs.
6. No scenario converted to complete success.
7. Pair 14 is the first regression observed under the accepted policy.
8. The regression was small relative to the simultaneous gains, but it is
   scientifically material because it disproves a universal no-regression
   interpretation.
9. Pair 14 changed residual family from LINKED_ROW_EVENT_CHAIN under the
   standard dataset to REAR_BOUNDARY_LOCK under dispersion.
10. The longest observed dependency chain reached 25 passengers.

Pooled Accepted-Policy Evidence
-------------------------------
Before Experiment 43:
- 240 paired scenarios.
- Aggregate improvement: +368.
- Improved/worse/equal: 17/0/223.

Experiment 43:
- 30 paired scenarios.
- Aggregate improvement: +112.
- Improved/worse/equal: 2/1/27.

Combined evidence:
- 270 paired scenarios.
- Aggregate improvement: +480.
- Improved pairs: 19.
- Worse pairs: 1.
- Equal pairs: 250.
- Observed regression rate: 1/270 (approximately 0.37%).

Scientific Verdict
------------------
Experiment 43 is both a strong positive aggregate result and the first clear
boundary result.

The accepted policy remains highly beneficial overall across the tested
deterministic conditions, but it can no longer be described as universally
non-regressive. The correct interpretation is now:

- usually neutral,
- occasionally strongly beneficial,
- rarely adverse under the tested samples,
- with the adverse case small but reproducible until shown otherwise.

This is a stronger scientific position than concealing or dismissing the first
negative case. It defines a specific boundary for investigation.

Discussion Incorporated
------------------------
Our earlier discussion anticipated that a conclusive programme must identify
not only where the policy works but also where it does not. Experiment 43
provides the first concrete example of that boundary.

The next step should therefore not immediately alter the policy. The first
requirement is exact deterministic replay. Only after reproducibility is
confirmed should the regression mechanism be isolated or a guarded refinement
be considered.

Experiment 43 Conclusion
------------------------
The policy delivered +112 passengers overall but introduced its first observed
regression: Pair 14 lost two passengers and changed residual stall family from
LINKED_ROW_EVENT_CHAIN to REAR_BOUNDARY_LOCK.

Experiment 44 performs an exact deterministic replay of the complete Experiment
43 sample to establish whether the regression is fully reproducible.


======================================================================
EXPERIMENT 44 - REGRESSION REPRODUCIBILITY REPLAY
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement backbone remains frozen. Experiment 44
changes no movement, seat-event, blocker-yield, dependency, arbitration or
admission-policy logic. It reuses the exact Experiment 43 seed and complete
scenario set so that reproducibility can be tested without confounding changes.

Planned Purpose
---------------
Confirm whether the first accepted-policy regression is deterministic and
exactly reproducible.

Replay Seed
-----------
    8783790101901

Occupancy Condition
-------------------
All scenarios remain fixed at:

    98% OCCUPANCY

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Expected Replay Signature
-------------------------
A valid exact replay should reproduce:

- Pair 02: 302/353 -> 347/353 (+45).
- Pair 14: 350/353 -> 348/353 (-2).
- Pair 22: 277/353 -> 346/353 (+69).
- Aggregate improvement: +112.
- Improved/worse/equal: 2/1/27.
- Total seated: 8,550 standard versus 8,662 dispersion.
- Total capacity represented: 8,841 passengers.

Independent Variable
--------------------
No new independent variable is introduced.

Experiment 44 is a reproducibility control. The experiment identity changes,
but the deterministic sample and accepted policy remain identical.

Primary Evaluation Order
------------------------
1. Exact per-pair replay.
2. Exact reproduction of Pair 14 regression.
3. Aggregate total equality with Experiment 43.
4. Residual stall-family equality.
5. Dependency-chain consistency.
6. Build-integrity confirmation.
7. Decision on subsequent regression-isolation design.

Decision Rule
-------------
- Exact reproduction confirms a deterministic policy boundary.
- Any mismatch blocks further interpretation until build integrity is resolved.
- No policy refinement is permitted within this replay experiment.
- A confirmed boundary may justify targeted isolation in Experiment 45.

Experimental Hypothesis
-----------------------
Because the architecture, seed and scenario set are deterministic, Experiment
44 should reproduce all Experiment 43 paired outcomes exactly.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It is not a live-airline boarding
instruction. The result must not be converted into operational advice without
empirical calibration and SME review.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 43
======================================================================
1. Final-two-row risk dispersion remains strongly positive in aggregate.
2. Experiment 43 delivered the largest recent holdout gain: +112 passengers.
3. The first accepted-policy regression occurred in Pair 14 at -2 passengers.
4. The cumulative evidence is now +480 across 270 paired scenarios.
5. The observed regression rate is 1/270, approximately 0.37%.
6. Universal no-regression language is no longer scientifically supportable.
7. The accepted policy should be described as sparse, usually neutral,
   occasionally strongly beneficial and rarely adverse in tested samples.
8. Experiment 44 tests exact deterministic reproducibility before any policy
   refinement.
9. Amit's movement and dependency architecture remains frozen.
10. SME-facing recommendations remain deferred.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v44 - EXPERIMENT 44 DESIGN
======================================================================


======================================================================
EXPERIMENT 44 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 44 executed the complete Experiment 43 sample again using the same
seed, same thirty scenario pairs, same fixed 98% occupancy condition, same
accepted final-two-row dispersion policy and the same frozen movement and
dependency architecture.

Observed Replay
---------------
The replay matched Experiment 43 exactly:

- Standard mode: 8,550/8,841.
- Final-two-row dispersion: 8,662/8,841.
- Aggregate change: +112 passengers.
- Improved pairs: 2.
- Worse pairs: 1.
- Equal pairs: 27.
- Converted successes: 0.
- Longest observed dependency chain: 25 passengers.

Changed Pairs
-------------
- Pair 02: 302/353 -> 347/353 (+45).
- Pair 14: 350/353 -> 348/353 (-2).
- Pair 22: 277/353 -> 346/353 (+69).

Residual-family transitions also replayed exactly:

- Pair 02:
  ENTRY_SATURATION_WITH_INTERNAL_CHAIN ->
  MIXED_OR_UNCLASSIFIED_RESIDUAL.

- Pair 14:
  LINKED_ROW_EVENT_CHAIN ->
  REAR_BOUNDARY_LOCK.

- Pair 22:
  ENTRY_SATURATION_WITH_INTERNAL_CHAIN ->
  MIXED_OR_UNCLASSIFIED_RESIDUAL.

Scientific Verdict
------------------
Experiment 44 confirms complete deterministic reproducibility.

The first accepted-policy regression is not execution noise, random drift,
temporary machine state or build instability. Pair 14 is a genuine deterministic
boundary case within the present model.

Because Experiment 44 is a reproducibility control rather than a new independent
sample, it is not added to the pooled accepted-policy totals.

The independent evidence therefore remains:

- 270 paired scenarios.
- Aggregate improvement: +480 passengers.
- Improved pairs: 19.
- Worse pairs: 1.
- Equal pairs: 250.
- Observed regression rate: 1/270, approximately 0.37%.

Discussion Incorporated
------------------------
The correct next step is not to hide the adverse case or immediately add a
corrective mechanism. The case should first be isolated so that the exact
admission-order trajectory and residual-family transition can be studied without
the other twenty-nine scenarios obscuring the evidence.

Experiment 44 Conclusion
------------------------
The complete Experiment 43 result, including the -2 Pair 14 regression, was
reproduced exactly. Experiment 45 isolates Pair 14 while preserving its original
deterministic scenario configuration.


======================================================================
EXPERIMENT 45 - PAIR 14 REGRESSION ISOLATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original people-movement backbone remains frozen. Experiment 45 changes
only the experiment harness so that the already confirmed Pair 14 scenario is
executed by itself. Passenger movement, seat events, blocker yields, dependency
chains, middle-bank arbitration and final-two-row admission dispersion are not
altered.

Planned Purpose
---------------
Isolate the first confirmed accepted-policy regression and reproduce its exact
paired outcome without unrelated scenarios in the output.

Source Experiment
-----------------
Experiments 43 and 44.

Source Pair
-----------
Pair 14.

Replay Seed
-----------
    8783790101901

Isolation Method
----------------
The deterministic scenario generator advances through source scenarios 1-13
without executing them. Source Scenario 14 is then executed in both paired
modes.

This preserves the exact Scenario 14:

- cabin configuration,
- scenario seed,
- manifest,
- seat assignment,
- aisle assignment,
- numerical entry headway,
- 98% occupancy,
- standard admission order,
- final-two-row dispersion order.

Paired Modes
------------
1. STANDARD DETERMINISTIC DATASET.
2. FINAL-TWO-ROW RISK-DISPERSION DATASET.

Expected Isolation Signature
----------------------------
- Cabin: 3-4-3.
- Occupancy: 98%.
- Numerical headway: 0-1.
- Standard: 350/353.
- Dispersion: 348/353.
- Delta: -2.
- Standard family: LINKED_ROW_EVENT_CHAIN.
- Dispersion family: REAR_BOUNDARY_LOCK.

Independent Variable
--------------------
No new policy variable is introduced.

The experiment is a targeted isolation control. Only the amount of unrelated
batch output is reduced.

Primary Evaluation Order
------------------------
1. Exact isolated outcome reproduction.
2. Exact residual-family reproduction.
3. Comparison of unresolved passengers.
4. Comparison of dependency-chain length and terminal prerequisite.
5. Comparison of rear-yield-space requirements.
6. Identification of admission-order displacement associated with the boundary.
7. Decision on whether a later guarded counterfactual is scientifically justified.

Decision Rule
-------------
- Exact reproduction validates the isolation harness.
- A mismatch blocks further interpretation.
- No corrective policy is introduced in this phase.
- A validated isolated case may support a controlled counterfactual in the next
  experiment.

Experimental Hypothesis
-----------------------
The isolated source scenario should reproduce the confirmed -2 regression and
the LINKED_ROW_EVENT_CHAIN to REAR_BOUNDARY_LOCK transition exactly.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic model validation. It is not live-airline guidance.
Empirical calibration and SME review remain required before practical use.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 44
======================================================================
1. Experiment 44 reproduced Experiment 43 exactly.
2. Pair 14 is a confirmed deterministic boundary case.
3. The independent pooled evidence remains +480 across 270 pairs.
4. The replay control is not double-counted as new policy evidence.
5. Universal no-regression language remains invalid.
6. The accepted policy remains strongly positive overall but not universally safe.
7. Experiment 45 isolates Pair 14 without modifying the policy.
8. Amit's movement and dependency architecture remains frozen.
9. Corrective logic remains deferred until the isolated mechanism is understood.
10. SME-facing recommendations remain deferred.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v45 - EXPERIMENT 45 DESIGN
======================================================================


======================================================================
EXPERIMENT 45 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 45 executed only the previously confirmed Experiment 43/44 Pair 14
scenario. The deterministic generator advanced through source scenarios 1-13
without executing them and then ran source Scenario 14 in both paired modes.

The isolation harness preserved:

- experiment seed 8783790101901,
- scenario seed -457221805872256975,
- 3-4-3 cabin,
- 353 passengers at 98% occupancy,
- numerical headway 0-1,
- the original passenger manifest,
- aisle assignments,
- standard admission order,
- final-two-row dispersion admission order,
- Amit Amlani's frozen movement and dependency architecture.

Observed Results
----------------
Standard deterministic dataset:

- Seated: 350/353.
- Unresolved: 3.
- Dependency length: 2.
- Residual family: LINKED_ROW_EVENT_CHAIN.
- Final prerequisite: YIELD_TILE_OCCUPIED.
- One-tile-earlier hold would preserve space: true.

Final-two-row risk-dispersion dataset:

- Seated: 348/353.
- Unresolved: 5.
- Dependency length: 5.
- Residual family: REAR_BOUNDARY_LOCK.
- Final prerequisite: INSUFFICIENT_REAR_YIELD_SPACE.
- One-tile-earlier hold would preserve space: false.

Paired delta:

    -2 passengers

Standard terminal chain:

    P32 -> P342 -> WAITING_FOR_SEAT_EVENT

Dispersion terminal chain:

    P32 -> P342 -> P54 -> P283 -> P208
    -> WAITING_FOR_SEAT_EVENT

Principal Findings
------------------
1. The isolated case reproduced Experiments 43 and 44 exactly.
2. The regression does not depend on the other twenty-nine batch scenarios.
3. It is not caused by accumulated batch state or execution order.
4. The adverse result belongs to one deterministic Scenario 14 trajectory.
5. The accepted dispersion policy extended the terminal dependency from two
   passengers to five.
6. The residual mechanism changed from a linked row-event chain to a physical
   rear-boundary lock.
7. The standard case retained a possible one-tile-earlier preservation point.
8. The dispersion case did not; its rear boundary required unavailable physical
   space beyond the cabin.
9. The immediate scientific target is therefore the admission-order divergence
   that produced the deeper chain, not a general scheduler modification.

Scientific Verdict
------------------
Experiment 45 validates Pair 14 as an isolated deterministic policy boundary.

The evidence indicates that final-two-row dispersion does not merely delay a
single passenger. In this case it changes the downstream dependency topology,
creating a longer chain ending at an impossible rear-yield requirement.

The correct next step is instrumentation of the pre-run admission queues to
identify the earliest deterministic divergence and the displacement of the
critical passengers.

Evidence Accounting
-------------------
Experiment 45 is an isolation control and is not added to the independent pooled
policy totals.

The independent evidence remains:

- 270 paired scenarios.
- Aggregate improvement: +480 passengers.
- Improved/worse/equal: 19/1/250.
- Observed regression rate: approximately 0.37%.

Discussion Incorporated
------------------------
Our discussion after Experiment 45 established that the regression should be
understood before any correction is attempted. The experiment supports this
approach: the policy changes the dependency architecture rather than producing
a simple timing delay.

Experiment 45 Conclusion
------------------------
The first regression is fully isolated and mechanistically associated with a
dependency-depth increase from two to five and a transition to rear-boundary
lock. Experiment 46 identifies the earliest pre-run admission-order divergence.


======================================================================
EXPERIMENT 46 - PAIR 14 ADMISSION-ORDER DIVERGENCE ANALYSIS
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement backbone remains completely frozen. Experiment 46 adds only
pre-run observation of the two deterministic admission queues. It does not
change passenger movement, seat events, blocker yields, middle-bank arbitration,
dependency analysis or the accepted final-two-row dispersion policy.

Planned Purpose
---------------
Determine the earliest queue position at which the standard and dispersion
admission orders separate and measure the displacement of the five passengers
in the confirmed dispersion terminal chain.

Source Scenario
---------------
Experiment 43/44/45 Pair 14.

Replay Configuration
--------------------
- Experiment seed: 8783790101901.
- Scenario seed: -457221805872256975.
- Cabin: 3-4-3.
- Occupancy: 98%.
- Passengers: 353.
- Numerical headway: 0-1.
- Focused paired executions: 2.

New Instrumentation
-------------------
Before movement begins, the code now prints:

1. Queue length for each aisle in both modes.
2. First queue-position divergence on each aisle.
3. A local window around the first divergence.
4. Passenger identity, seat, aisle and rear-protection status.
5. Standard and dispersion positions of:
   - P32,
   - P342,
   - P54,
   - P283,
   - P208.
6. Signed displacement for each critical passenger.

Independent Variable
--------------------
No new operational or policy variable is introduced.

The only change is observational instrumentation.

Primary Evaluation Order
------------------------
1. Confirm the isolated -2 outcome remains unchanged.
2. Locate the first left-aisle divergence.
3. Locate the first right-aisle divergence.
4. Identify which passenger is first displaced.
5. Measure whether any terminal-chain passenger moves earlier or later.
6. Relate the queue displacement to the five-passenger rear-boundary chain.
7. Decide whether a single narrow queue-order counterfactual is justified.

Decision Rule
-------------
- The paired outcome must remain 350/353 versus 348/353.
- Any movement-result change would invalidate the instrumentation.
- A clear first divergence provides the candidate cause for the next
  counterfactual experiment.
- No correction is introduced in Experiment 46.

Experimental Hypothesis
-----------------------
The left aisle may remain identical, while a narrow right-aisle displacement
among protected final-row passengers is expected to precede the deeper
rear-boundary chain.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic mechanism analysis and not operational airline
guidance. Empirical calibration and SME review remain required.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 45
======================================================================
1. Pair 14 is isolated and reproducible.
2. The regression is intrinsic to one deterministic admission trajectory.
3. Dispersion increases terminal dependency depth from two to five.
4. The adverse outcome ends in an impossible rear-boundary yield requirement.
5. The effect is not a simple one-tick or one-passenger delay.
6. Experiment 46 examines the pre-run queue divergence without changing policy.
7. The independent pooled evidence remains +480 across 270 pairs.
8. Isolation and instrumentation controls are not double-counted.
9. Amit's movement and dependency architecture remains frozen.
10. Corrective logic and SME recommendations remain deferred.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v46 - EXPERIMENT 46 DESIGN
======================================================================


======================================================================
EXPERIMENT 46 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 46 executed the isolated Pair 14 scenario with pre-run admission-order
instrumentation. The added diagnostics did not change the paired outcome.

Observed Outcome
----------------
Standard deterministic dataset:

- Seated: 350/353.
- Residual family: LINKED_ROW_EVENT_CHAIN.
- Dependency length: 2.

Final-two-row risk-dispersion dataset:

- Seated: 348/353.
- Residual family: REAR_BOUNDARY_LOCK.
- Dependency length: 5.

Paired delta:

    -2 passengers

Admission-Order Findings
------------------------
Left aisle:

- Queue length: 178 in both modes.
- First divergence: NONE.

Right aisle:

- Queue length: 175 in both modes.
- Positions 1-156: identical.
- First divergence: position 157.

Standard order:

    157: P283 -> 36J -> protected rear-boundary passenger
    158: P97  -> 12J -> unprotected passenger

Dispersion order:

    157: P97  -> 12J -> unprotected passenger
    158: P283 -> 36J -> protected rear-boundary passenger

All later observed positions immediately returned to the same sequence.

Critical-Chain Passenger Displacement
-------------------------------------
- P32: 0.
- P342: 0.
- P54: 0.
- P283: +1.
- P208: 0.

Principal Finding
-----------------
Only P283 among the five dispersion terminal-chain passengers is displaced, and
the displacement is exactly one admission position later.

The first observed regression can therefore be traced through a narrow
deterministic chain:

    accepted dispersion rule
    -> one local P283/P97 queue swap
    -> altered right-aisle arrival order
    -> dependency depth 2 becomes 5
    -> rear-boundary lock
    -> -2 passengers seated

Scientific Verdict
------------------
Experiment 46 identifies a single minimal candidate cause rather than a broad
difference between the two admission orders.

The result does not yet prove causal sufficiency. The correct next step is to
retain the complete dispersion queue but restore only P283 before P97 and observe
whether the original standard outcome returns.

Evidence Accounting
-------------------
Experiment 46 is an observational mechanism-control experiment. It is not added
to the independent pooled policy totals.

The independent evidence remains:

- 270 paired scenarios.
- Aggregate improvement: +480 passengers.
- Improved/worse/equal: 19/1/250.
- Observed regression rate: approximately 0.37%.

Discussion Incorporated
------------------------
Our discussion established that the regression should be followed from policy
to queue perturbation to dependency topology. Experiment 46 completes the
queue-perturbation stage and supports a single-swap counterfactual rather than a
general policy rewrite.

Experiment 46 Conclusion
------------------------
The earliest and only local admission-order divergence relevant to the confirmed
terminal chain is the P283/P97 swap at right-aisle positions 157-158.


======================================================================
EXPERIMENT 47 - PAIR 14 SINGLE-SWAP COUNTERFACTUAL
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's people-movement, dependency, seat-event, blocker-yield and arbitration
architecture remains frozen. Experiment 47 changes only one pre-run admission
ordering in a third counterfactual execution. No runtime decision logic, rescue
scheduler, adaptive guard or behavioural rule is introduced.

Planned Purpose
---------------
Test whether the single P283/P97 queue swap identified in Experiment 46 is
causally sufficient to produce the Pair 14 regression.

Source Scenario
---------------
Experiment 43/44/45/46 Pair 14.

Replay Configuration
--------------------
- Experiment seed: 8783790101901.
- Scenario seed: -457221805872256975.
- Cabin: 3-4-3.
- Passengers: 353.
- Occupancy: 98%.
- Numerical headway: 0-1.

Three Deterministic Executions
------------------------------
1. STANDARD_DETERMINISTIC_DATASET.
2. FINAL_TWO_ROW_RISK_DISPERSION_DATASET.
3. FINAL_TWO_ROW_SINGLE_SWAP_COUNTERFACTUAL_DATASET.

Counterfactual Definition
-------------------------
The third mode first builds the complete accepted final-two-row dispersion queue.

It then restores only:

    P283 before P97

on the right aisle.

No other queue position is changed.

Expected Queue Integrity
------------------------
- P283 standard position: 157.
- P283 dispersion position: 158.
- P283 counterfactual position: 157.
- P97 standard position: 158.
- P97 dispersion position: 157.
- P97 counterfactual position: 158.
- Counterfactual differences from dispersion: exactly 2 queue positions.
- Counterfactual differences from standard: 0 for this isolated scenario.

Primary Evaluation Order
------------------------
1. Confirm the standard outcome remains 350/353.
2. Confirm the dispersion outcome remains 348/353.
3. Measure the single-swap counterfactual outcome.
4. Compare residual stall family.
5. Compare terminal dependency length.
6. Determine whether the +2 loss is fully reversed.
7. Decide whether the isolated swap is a complete causal explanation.

Decision Rule
-------------
If the counterfactual returns to:

- 350/353,
- LINKED_ROW_EVENT_CHAIN,
- dependency length 2,

then the P283/P97 swap is causally sufficient to explain the observed regression
within Pair 14.

If it remains:

- 348/353,
- REAR_BOUNDARY_LOCK,
- dependency length 5,

the swap is correlated but not sufficient.

Any intermediate result indicates partial causation.

Experimental Hypothesis
-----------------------
Restoring only P283 before P97 will remove the rear-boundary lock and return the
isolated outcome to the standard 350/353 result.

Research Boundary
-----------------
This is a single-case causal counterfactual. It is not a new accepted boarding
policy and must not be generalised to other scenarios without independent tests.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic research evidence, not live-airline guidance.
Empirical calibration and SME review remain required.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 46
======================================================================
1. The Pair 14 regression is reproducible and isolated.
2. The left admission order is completely unchanged.
3. The right order first diverges at position 157.
4. The divergence is a single P283/P97 swap.
5. Only P283 among the terminal-chain passengers is displaced.
6. P283 moves exactly one position later.
7. The resulting dependency depth increases from two to five.
8. Experiment 47 tests causal sufficiency through one isolated restoration.
9. The independent pooled evidence remains +480 across 270 pairs.
10. Amit's movement architecture remains frozen.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v47 - EXPERIMENT 47 DESIGN
======================================================================


======================================================================
EXPERIMENT 47 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 47 executed the isolated Pair 14 scenario in three deterministic
modes:

1. Standard deterministic dataset.
2. Accepted final-two-row risk-dispersion dataset.
3. Final-two-row dispersion with only the P283/P97 order restored.

The official output confirmed that the third queue differed from the dispersion
queue at exactly two queue positions and matched the standard queue for the
isolated scenario.

Queue Integrity
---------------
P283 positions:

- Standard: 157.
- Dispersion: 158.
- Counterfactual: 157.

P97 positions:

- Standard: 158.
- Dispersion: 157.
- Counterfactual: 158.

Counterfactual queue differences:

- Versus dispersion: 2 positions.
- Versus standard: 0 positions.
- Change restricted to P283/P97: true.

Observed Outcomes
-----------------
Standard:

- Seated: 350/353.
- Residual family: LINKED_ROW_EVENT_CHAIN.
- Dependency length: 2.

Original dispersion:

- Seated: 348/353.
- Residual family: REAR_BOUNDARY_LOCK.
- Dependency length: 5.

Single-swap counterfactual:

- Seated: 350/353.
- Residual family: LINKED_ROW_EVENT_CHAIN.
- Dependency length: 2.

Counterfactual deltas:

- Versus dispersion: +2.
- Versus standard: 0.

Principal Finding
-----------------
Restoring only P283 before P97 reverses the complete observed regression.

The counterfactual restores:

- the passenger count,
- the residual family,
- the dependency length,
- the unresolved-passenger structure,
- the standard final snapshot.

Scientific Verdict
------------------
Within isolated Pair 14, the P283/P97 swap is causally sufficient to produce
and eliminate the regression.

The result establishes the complete deterministic causal sequence:

    final-two-row dispersion
    -> P283 delayed one admission position
    -> dependency depth increases from 2 to 5
    -> linked row-event chain becomes rear-boundary lock
    -> two fewer passengers seated

Restoring the single local order reverses every observed downstream effect.

Evidence Boundary
-----------------
Experiment 47 is a single-case causal counterfactual. It is not independent
policy evidence and is not added to the cumulative totals.

The independent pooled evidence remains:

- 270 paired scenarios.
- Aggregate improvement: +480 passengers.
- Improved/worse/equal: 19/1/250.
- Observed regression rate: approximately 0.37%.

Discussion Incorporated
------------------------
Our discussion concluded that this is stronger than correlation: a single local
admission-order change was shown to be sufficient to create and reverse the
entire deterministic regression.

The next question is whether the causal signature can be expressed generically
without passenger IDs and without destroying the wider gains of the accepted
dispersion policy.

Experiment 47 Conclusion
------------------------
The isolated causal mechanism is proved. Experiment 48 tests a generic same-row
protected-adjacency rule across the original thirty-scenario stress batch.


======================================================================
EXPERIMENT 48 - SAME-ROW PROTECTED ADJACENCY VALIDATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency, seat-event, blocker-yield, reservation and
arbitration architecture remains frozen.

Experiment 48 changes only the pre-run construction of one third dataset mode.
No runtime rescue, scheduler intervention, adaptive correction or passenger-ID
exception is introduced.

Planned Purpose
---------------
Generalise the Experiment 47 causal finding into a passenger-independent
admission-order rule and test it across the original thirty deterministic
Experiment 43 scenarios.

Generic Hypothesis
------------------
The original final-two-row policy separates consecutive protected rear-boundary
passengers.

Experiment 47 showed that separating P208 and P283, who both target row 36,
caused the Pair 14 regression.

Experiment 48 therefore tests:

    preserve consecutive protected passengers when they target the same row;
    otherwise retain the original final-two-row dispersion behaviour.

Three Dataset Modes
-------------------
Each scenario executes in:

1. STANDARD_DETERMINISTIC_DATASET.
2. FINAL_TWO_ROW_RISK_DISPERSION_DATASET.
3. FINAL_TWO_ROW_SAME_ROW_PRESERVATION_DATASET.

Scenario Scope
--------------
- Original Experiment 43 seed: 8783790101901.
- Original thirty deterministic scenario configurations.
- 90 total deterministic executions.
- Same cabins, occupancies, headways, manifests and aisle assignments.

Independent Variable
--------------------
Only the admission-order handling of consecutive protected passengers in the
same final-two-row destination row.

The rule contains:

- no passenger IDs,
- no scenario number,
- no hard-coded queue position,
- no runtime state.

Primary Evaluation Order
------------------------
1. Reproduce the original standard and dispersion totals.
2. Confirm Pair 14 no longer regresses.
3. Compare aggregate refined total with original dispersion.
4. Count refined improved/worse/equal outcomes.
5. Identify any new regressions.
6. Determine whether the original large gains remain.
7. Accept or reject same-row preservation as a general refinement.

Decision Rule
-------------
The refinement is promising only if it:

- removes Pair 14's -2 regression,
- preserves the +112 gain in this thirty-pair batch or remains close to it,
- introduces no new regression,
- retains the two major improvements from Pairs 02 and 22.

If it fixes Pair 14 but loses the broader gains, the causal mechanism is valid
but the proposed generalisation must be rejected.

Experimental Hypothesis
-----------------------
Same-row protected adjacency may remove the known regression while retaining
most of the accepted policy's gains.

Research Boundary
-----------------
Experiment 48 is a first generalisation test, not an accepted operational
policy. A negative result remains scientifically useful because it separates
the proven local cause from an unsafe global rule.

Operational Interpretation Boundary
-----------------------------------
This remains deterministic research evidence. Airline implementation would
require empirical calibration and SME review.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 47
======================================================================
1. Pair 14's regression is causally explained by one local queue swap.
2. Restoring P283 before P97 recovers both passengers.
3. Dependency depth returns from five to two.
4. REAR_BOUNDARY_LOCK returns to LINKED_ROW_EVENT_CHAIN.
5. The result is a complete isolated causal proof.
6. It does not justify a passenger-specific production exception.
7. Experiment 48 tests a generic same-row protected-adjacency rule.
8. The original thirty scenarios are used for validation.
9. Independent pooled evidence remains +480 across 270 pairs.
10. Amit's movement architecture remains frozen.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v48 - EXPERIMENT 48 DESIGN
======================================================================


======================================================================
EXPERIMENT 48 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 48 executed all thirty original scenarios in three deterministic
modes:

1. Standard deterministic dataset.
2. Accepted final-two-row risk-dispersion dataset.
3. Same-row protected-adjacency refinement.

Total executions:

    90

The official run reproduced the expected standard and accepted-dispersion
controls.

Aggregate Results
-----------------
Standard total seated:

    8550

Original final-two-row dispersion total:

    8662

Original dispersion aggregate delta:

    +112

Original improved/worse/equal distribution:

    2 / 1 / 27

Same-row preservation total:

    8511

Same-row preservation aggregate delta versus standard:

    -39

Same-row preservation delta versus accepted dispersion:

    -151

Same-row preservation improved/worse/equal distribution:

    0 / 1 / 29

Pair 14 Result
--------------
The generic refinement repaired the known regression:

- Standard: 350/353.
- Original dispersion: 348/353.
- Same-row preservation: 350/353.

The residual family also returned from:

    REAR_BOUNDARY_LOCK

to:

    LINKED_ROW_EVENT_CHAIN

This confirms that the local causal mechanism identified in Experiments 46-47
was correctly understood.

Loss of Established Improvements
--------------------------------
The refinement removed both major benefits present in the accepted dispersion
dataset.

Pair 02:

- Standard: 302/353.
- Original dispersion: 347/353.
- Same-row preservation: 302/353.
- Lost accepted gain: 45 passengers.

Pair 22:

- Standard: 277/353.
- Original dispersion: 346/353.
- Same-row preservation: 277/353.
- Lost accepted gain: 69 passengers.

New Regression
--------------
Pair 18 introduced a new and substantial failure:

- Standard: 282/282.
- Original dispersion: 282/282.
- Same-row preservation: 243/282.
- New regression: -39 passengers.
- Refined residual family:
  ENTRY_SATURATION_WITH_INTERNAL_CHAIN.

Principal Finding
-----------------
The Experiment 47 local cause is valid, but the proposed same-row preservation
generalisation is unsafe.

The same admission-order relationship that repairs Pair 14 also suppresses the
mechanism responsible for the major improvements in Pairs 02 and 22 and creates
a new large regression in Pair 18.

Scientific Verdict
------------------
Reject FINAL_TWO_ROW_SAME_ROW_PRESERVATION_DATASET as a general refinement.

Retain FINAL_TWO_ROW_RISK_DISPERSION_DATASET as the accepted policy.

Experiment 48 demonstrates the important distinction between:

- proving a local deterministic cause, and
- safely generalising that cause into a global policy.

Discussion Incorporated
------------------------
Our discussion concluded that the negative generalisation result strengthens the
research. It shows that the project does not convert a successful single-case
counterfactual into an untested global rule.

The correct next step is observational rather than corrective:

- revert to the accepted dispersion policy,
- compare the admission-order signatures of the two major improvements and the
  single regression,
- identify why Pair 14 is uniquely sensitive,
- avoid another heuristic until a structural discriminator is found.

Evidence Accounting
-------------------
Experiment 48 is a policy-generalisation rejection experiment.

It does not alter the accepted independent cumulative evidence:

- 270 paired scenarios.
- Aggregate improvement: +480 passengers.
- Improved/worse/equal: 19/1/250.
- Regression rate: approximately 0.37%.

Experiment 48 Conclusion
------------------------
Same-row protected adjacency repairs the known local regression but destroys the
wider benefit profile. It is rejected as a general policy.


======================================================================
EXPERIMENT 49 - ADMISSION-PERTURBATION SIGNATURE ANALYSIS
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement engine, dependency propagation, scheduler, seat-event mechanics,
blocker yields, reservations, arbitration and deterministic replay remain
frozen.

Experiment 49 introduces no new boarding policy.

It observes only the pre-movement differences between:

- STANDARD_DETERMINISTIC_DATASET, and
- FINAL_TWO_ROW_RISK_DISPERSION_DATASET.

Planned Purpose
---------------
Identify a structural admission-order signature that distinguishes:

- Pair 02: +45 improvement.
- Pair 14: -2 regression.
- Pair 22: +69 improvement.
- Twenty-seven equal pairs.

The experiment seeks an explanation for why the same accepted deterministic
transformation produces large gains in two scenarios but a small regression in
one.

Scenario Scope
--------------
- Experiment seed: 8783790101901.
- Original thirty deterministic scenarios.
- Two modes per scenario.
- Total executions: 60.

Pre-Movement Measurements
-------------------------
For each aisle and scenario:

1. Number of differing queue positions.
2. First divergence position.
3. Number of protected rear-boundary passengers moved later.
4. Number moved earlier.
5. Maximum protected delay.
6. Maximum protected advance.
7. Identity and destination of the largest protected displacement.
8. Same-row protected adjacencies broken.
9. Same-row protected adjacencies created.
10. Protected rear-boundary order inversions.

Outcome Correlation
-------------------
Each signature is reported beside:

- passenger delta,
- improved/worse/equal category,
- standard residual family,
- dispersion residual family.

Primary Research Question
-------------------------
Does Pair 14 have a minimal, late and isolated perturbation signature that is
structurally different from the broader, earlier transformations that create
the major improvements in Pairs 02 and 22?

Working Hypothesis
------------------
The regression may be distinguished by:

- only two changed queue positions,
- one protected passenger delayed by one place,
- one broken same-row protected adjacency,
- a very late first divergence.

The two major improvements may instead involve:

- six changed positions,
- three protected passengers delayed,
- maximum delay of two,
- earlier first divergence,
- multiple same-row adjacency breaks.

Decision Rule
-------------
Experiment 49 is successful if it identifies a reproducible structural
difference between the regression and improvements without:

- passenger IDs,
- scenario-specific exceptions,
- runtime intervention,
- movement-engine changes.

No correction will be accepted from this experiment alone.

Evidence Boundary
-----------------
This is mechanism analysis, not a policy test. Its results do not alter the
accepted cumulative totals.

Operational Interpretation Boundary
-----------------------------------
The analysis remains deterministic and representative. Airline deployment would
require empirical calibration and domain-specialist review.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 48
======================================================================
1. Experiment 47 proved the local P283/P97 causal mechanism.
2. Experiment 48 proved that naïve same-row generalisation is unsafe.
3. Pair 14 was repaired, but Pairs 02 and 22 lost +45 and +69 gains.
4. Pair 18 introduced a new -39 regression.
5. Same-row preservation is rejected.
6. Final-two-row risk dispersion remains the accepted dataset.
7. Experiment 49 returns to observational mechanism analysis.
8. It compares queue-perturbation signatures across all thirty pairs.
9. No new policy or runtime intervention is introduced.
10. Amit's movement architecture remains frozen.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v49 - EXPERIMENT 49 DESIGN
======================================================================


======================================================================
EXPERIMENT 49 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 49 executed the original thirty deterministic scenarios in two
unchanged modes:

1. STANDARD_DETERMINISTIC_DATASET.
2. FINAL_TWO_ROW_RISK_DISPERSION_DATASET.

Total executions:

    60

The accepted outcome control was reproduced:

- Standard total seated: 8550.
- Dispersion total seated: 8662.
- Aggregate delta: +112.
- Improved/worse/equal: 2/1/27.

Key Structural Signatures
-------------------------

Pair 02:

- Outcome: +45.
- Differing positions: 6.
- First divergence: 32.
- Protected passengers moved later: 3.
- Maximum protected delay: 2.
- Same-row protected adjacencies broken: 3.
- Protected-order inversions: 0.

Pair 14:

- Outcome: -2.
- Differing positions: 2.
- First divergence: 157.
- Protected passengers moved later: 1.
- Maximum protected delay: 1.
- Same-row protected adjacencies broken: 1.
- Protected-order inversions: 0.

Pair 22:

- Outcome: +69.
- Differing positions: 6.
- First divergence: 44.
- Protected passengers moved later: 3.
- Maximum protected delay: 2.
- Same-row protected adjacencies broken: 2.
- Protected-order inversions: 0.

Principal Finding
-----------------
Pair 14 has a uniquely minimal, late and isolated perturbation signature.

The two large improvements have broader and earlier perturbations:

    Pair 14:
    two changed positions
    -> one protected passenger delayed by one
    -> first divergence at 157
    -> -2 outcome

    Pairs 02/22:
    six changed positions
    -> three protected passengers delayed
    -> maximum delay two
    -> first divergence at 32/44
    -> +45/+69 outcomes

Scientific Interpretation
-------------------------
The causal mechanism is no longer described only by passenger identity.

Experiment 49 identifies a pre-movement structural distinction between:

- minimal late perturbation associated with the sole regression, and
- broad early perturbation associated with the major gains.

No protected-order inversions occurred in the three decisive cases, so inversion
count does not explain their outcome difference.

Discussion Incorporated
------------------------
Our discussion concluded that the result supports a classification experiment
rather than another policy heuristic.

The next step should assign all scenarios to frozen perturbation families before
movement and correlate those families with:

- passenger delta,
- dependency depth,
- residual stall family,
- rear-boundary lock,
- entry saturation,
- completion.

Evidence Boundary
-----------------
Experiment 49 is explanatory only and does not alter the accepted policy or the
independent cumulative evidence.

Experiment 49 Conclusion
------------------------
A structural admission-order discriminator has emerged. Experiment 50 tests
whether it remains coherent when applied to every scenario.


======================================================================
EXPERIMENT 50 - PERTURBATION-FAMILY CLASSIFICATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency propagation, scheduler, seat-event mechanics,
blocker yields, reservation coordination, arbitration and deterministic replay
remain frozen.

Experiment 50 classifies admission-order structure before movement. It does not
modify passenger order beyond the already accepted dispersion dataset.

Planned Purpose
---------------
Transform Experiment 49's structural signatures into a deterministic
classification framework and correlate each class with downstream outcomes.

Frozen Classification Families
-------------------------------

A_NO_PERTURBATION

- No queue positions differ.

B_MINIMAL_LATE

- At most two positions differ.
- No more than one protected passenger moves later.
- Maximum protected delay is one.
- First divergence occurs at position 100 or later.

C_BROAD_EARLY

- At least four positions differ.
- At least two protected passengers move later.
- Maximum protected delay is at least one.
- First divergence occurs by position 60.

D_OTHER_PERTURBATION

- Any non-zero perturbation not matching B or C.

These thresholds are frozen before outcome comparison.

Scenario Scope
--------------
- Original thirty deterministic scenarios.
- Standard and accepted dispersion modes.
- 60 total executions.
- Same seed, manifests, cabin configurations, occupancy and headways.

Outcome Matrix
--------------
For each perturbation family, report:

1. Scenario count.
2. Improved/worse/equal distribution.
3. Aggregate passenger delta.
4. Average standard dependency depth.
5. Average dispersion dependency depth.
6. Dispersion residual-family distribution.

Primary Hypothesis
------------------
B_MINIMAL_LATE will contain Pair 14 and may concentrate regression risk.

C_BROAD_EARLY will contain Pairs 02 and 22 and may concentrate aggregate gains.

A_NO_PERTURBATION should remain outcome-neutral.

D_OTHER_PERTURBATION will test whether intermediate signatures are also
outcome-neutral or whether further subdivision is required.

Decision Rule
-------------
The classification is useful if:

- Pair 14 falls within B_MINIMAL_LATE.
- Pairs 02 and 22 fall within C_BROAD_EARLY.
- B contains no improvements and concentrates the observed regression.
- C contains no regressions and captures the major aggregate gain.
- A remains neutral.

No policy will be changed based on Experiment 50 alone.

Research Boundary
-----------------
This is deterministic classification research, not a production predictor.
Broader datasets and domain-expert review remain necessary.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 49
======================================================================
1. The accepted dispersion control remains +112 over thirty scenarios.
2. Pair 14 is a two-position, very-late perturbation.
3. Pairs 02 and 22 are six-position, earlier perturbations.
4. Pair 14 delays one protected passenger by one place.
5. Pairs 02 and 22 delay three protected passengers by up to two places.
6. Protected-order inversions do not distinguish the decisive cases.
7. A structural classifier is now justified.
8. Experiment 50 freezes classes before movement.
9. No new boarding policy is introduced.
10. Amit's movement architecture remains frozen.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v50 - EXPERIMENT 50 DESIGN
======================================================================


======================================================================
EXPERIMENT 50 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 50 executed the original thirty deterministic scenarios using:

1. STANDARD_DETERMINISTIC_DATASET.
2. FINAL_TWO_ROW_RISK_DISPERSION_DATASET.

Total executions:

    60

The accepted outcome control was reproduced:

- Standard total seated: 8550.
- Dispersion total seated: 8662.
- Aggregate delta: +112.
- Improved/worse/equal: 2/1/27.

Official Perturbation-Family Matrix
-----------------------------------

A_NO_PERTURBATION

- Scenarios: 7.
- Improved/worse/equal: 0/0/7.
- Aggregate delta: 0.
- Average dependency depth:
  standard 1.71, dispersion 1.71.

B_MINIMAL_LATE

- Scenarios: 6.
- Improved/worse/equal: 0/1/5.
- Aggregate delta: -2.
- Average dependency depth:
  standard 2.67, dispersion 3.17.
- The sole regression, Pair 14, was concentrated here.

C_BROAD_EARLY

- Scenarios: 10.
- Improved/worse/equal: 2/0/8.
- Aggregate delta: +114.
- Average dependency depth:
  standard 4.90, dispersion 1.00.
- Both major improvements, Pairs 02 and 22, were concentrated here.

D_OTHER_PERTURBATION

- Scenarios: 7.
- Improved/worse/equal: 0/0/7.
- Aggregate delta: 0.
- Average dependency depth:
  standard 3.71, dispersion 3.71.

Principal Finding
-----------------
The pre-movement classification produced a clean separation within the original
thirty-scenario set:

- all aggregate gains were concentrated in C_BROAD_EARLY;
- the sole regression was concentrated in B_MINIMAL_LATE;
- A_NO_PERTURBATION and D_OTHER_PERTURBATION were outcome-neutral.

Dependency Interpretation
-------------------------
The outcome separation was supported by dependency topology.

B_MINIMAL_LATE increased average dependency depth:

    2.67 -> 3.17

C_BROAD_EARLY sharply reduced average dependency depth:

    4.90 -> 1.00

Therefore the family distinction correlates not only with passenger totals but
also with downstream dependency simplification or expansion.

Discussion Incorporated
------------------------
Our discussion identified Experiment 50 as the strongest classification result
in this sequence. It transformed the Pair 14 observation into a systematic
pre-movement framework.

The next scientific step is not another heuristic. It is prospective validation
on a fresh deterministic scenario set.

The prediction must be printed and frozen before either movement simulation
executes.

Scientific Boundary
-------------------
Experiment 50 validates the classification on the scenario set from which its
thresholds were derived. It is therefore an internal classification result, not
independent predictive validation.

Independent support requires a new deterministic seed and new scenarios.

Experiment 50 Conclusion
------------------------
The classification is coherent and strongly associated with outcome direction
and dependency-depth change in the original scenario set. Prospective holdout
validation is justified.


======================================================================
EXPERIMENT 51 - PROSPECTIVE PERTURBATION VALIDATION
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement engine, dependency propagation, scheduler, seat-event logic,
blocker mechanics, reservation coordination, arbitration and deterministic
replay remain frozen.

Experiment 51 changes neither the accepted boarding policy nor the runtime
architecture.

Fresh Holdout Dataset
---------------------
Experiment seed:

    9783790101901

This seed is different from the Experiment 43-50 seed.

Scenario pairs:

    30

Total executions:

    60

Each pair compares:

- STANDARD_DETERMINISTIC_DATASET.
- FINAL_TWO_ROW_RISK_DISPERSION_DATASET.

Prospective Execution Order
---------------------------
For each pair, the program performs these steps in fixed order:

1. Construct both admission queues.
2. Measure the admission-perturbation signature.
3. Assign the Experiment 50 family.
4. Assign and print the prediction rule.
5. Print:
   PREDICTION FROZEN BEFORE MOVEMENT: true
6. Execute the standard simulation.
7. Execute the dispersion simulation.
8. Compare prediction with observed direction.

The prediction is therefore committed before movement begins.

Frozen Family Definitions
-------------------------

A_NO_PERTURBATION

- No queue positions differ.

B_MINIMAL_LATE

- At most two queue positions differ.
- No more than one protected passenger moves later.
- Maximum protected delay is one.
- First divergence is position 100 or later.

C_BROAD_EARLY

- At least four queue positions differ.
- At least two protected passengers move later.
- Maximum protected delay is at least one.
- First divergence is position 60 or earlier.

D_OTHER_PERTURBATION

- Any non-zero signature not matching B or C.

Frozen Prediction Rules
-----------------------

A_NO_PERTURBATION:

    PREDICT_EQUAL

B_MINIMAL_LATE:

    PREDICT_NON_IMPROVING

The dispersion outcome is predicted to be equal to or below standard.

C_BROAD_EARLY:

    PREDICT_NON_REGRESSING

The dispersion outcome is predicted to be equal to or above standard.

D_OTHER_PERTURBATION:

    PREDICT_EQUAL

Rationale for Asymmetric Rules
------------------------------
Experiment 50 did not show that every B scenario regresses or every C scenario
improves.

It showed:

- B contained one regression and five neutral cases.
- C contained two improvements and eight neutral cases.

Therefore the prospective claims are deliberately conservative:

- B predicts absence of improvement.
- C predicts absence of regression.

This avoids overstating exact directional prediction.

Validation Measures
-------------------
Experiment 51 reports:

1. Family-level scenario count.
2. Prediction-rule compliance.
3. Improved/worse/equal distribution.
4. Aggregate passenger delta.
5. Overall prospective accuracy.
6. Accuracy by prediction rule.

Decision Rule
-------------
The experiment succeeds as prospective evidence only to the extent that the
frozen rules remain accurate on the fresh holdout set.

Any failure must be retained and documented. Thresholds must not be altered
after viewing holdout outcomes.

Evidence Accounting
-------------------
Experiment 51 uses a new seed and may provide new independent paired evidence.

Its results must not be added to cumulative totals until the user's official run
is reviewed.

Operational Boundary
--------------------
This remains deterministic research. It is not a production airline predictor
and requires broader datasets, empirical calibration and airline-domain review.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 50
======================================================================
1. Experiment 50 separated all original gains and the sole regression by family.
2. C_BROAD_EARLY contained +114 and no regression.
3. B_MINIMAL_LATE contained -2 and no improvement.
4. A and D were neutral.
5. Dependency depth fell sharply in C and rose modestly in B.
6. The original-set result is internal rather than independent validation.
7. Experiment 51 uses a fresh deterministic seed.
8. Predictions are printed before movement.
9. Prediction thresholds and rules are frozen.
10. Amit's movement architecture remains unchanged.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v51 - EXPERIMENT 51 DESIGN
======================================================================


======================================================================
EXPERIMENT 51 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 51 used a fresh deterministic holdout seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

The Experiment 50 family definitions and prediction rules were frozen before
movement.

Official Outcome
----------------
- Standard total seated: 7916.
- Dispersion total seated: 7940.
- Aggregate delta: +24.
- Improved/worse/equal: 3/0/27.
- Overall prediction-rule accuracy: 28/30 (93.33%).

Rule Accuracy
-------------
PREDICT_EQUAL:

    20/20

PREDICT_NON_IMPROVING:

    3/5

PREDICT_NON_REGRESSING:

    5/5

Family Findings
---------------

A_NO_PERTURBATION

- Scenarios: 10.
- Prediction compliance: 10/10.
- Improved/worse/equal: 0/0/10.
- Aggregate delta: 0.

B_MINIMAL_LATE

- Scenarios: 5.
- Prediction compliance: 3/5.
- Improved/worse/equal: 2/0/3.
- Aggregate delta: +14.

C_BROAD_EARLY

- Scenarios: 5.
- Prediction compliance: 5/5.
- Improved/worse/equal: 1/0/4.
- Aggregate delta: +10.

D_OTHER_PERTURBATION

- Scenarios: 10.
- Prediction compliance: 10/10.
- Improved/worse/equal: 0/0/10.
- Aggregate delta: 0.

False-Positive Cases
--------------------

Pair 06:

- Family: B_MINIMAL_LATE.
- Predicted: non-improving.
- Standard: 226/235.
- Dispersion: 235/235.
- Delta: +9.
- Differing positions: 2.
- First divergence: 102.
- Same-row adjacencies broken: 0.
- Standard residual family: REAR_BOUNDARY_LOCK.
- Dispersion residual family: COMPLETE_CABIN.

Pair 19:

- Family: B_MINIMAL_LATE.
- Predicted: non-improving.
- Standard: 230/235.
- Dispersion: 235/235.
- Delta: +5.
- Differing positions: 2.
- First divergence: 103.
- Same-row adjacencies broken: 0.
- Standard residual family: REAR_BOUNDARY_LOCK.
- Dispersion residual family: COMPLETE_CABIN.

Principal Finding
-----------------
Experiment 51 partially validated the Experiment 50 framework.

A, C and D generalised perfectly under their frozen rules.

B_MINIMAL_LATE did not fully generalise. Two cases that appeared structurally
similar before movement improved by releasing a standard rear-boundary lock and
completing the cabin.

Discussion Incorporated
------------------------
Our discussion concluded that this failure strengthens rather than weakens the
research narrative.

Experiment 50 established an internal association.

Experiment 51 showed that the association was not universally predictive.

The correct next step is not threshold retuning. Retuning after observing the
holdout would invalidate the prospective test.

Instead, the two false positives should be analysed as explanatory residual
topology transitions.

Scientific Interpretation
-------------------------
B_MINIMAL_LATE is too coarse to determine outcome direction by itself.

The same pre-movement family can lead to:

- no topology change,
- preserved rear-boundary lock,
- or release of a rear-boundary lock.

The actual downstream residual transition appears decisive.

Evidence Accounting
-------------------
Experiment 51 contributes thirty new independent paired scenarios.

Updated independent cumulative evidence:

- Prior paired scenarios: 270.
- Experiment 51 paired scenarios: 30.
- Total paired scenarios: 300.

Prior aggregate passenger gain:

    +480

Experiment 51 aggregate gain:

    +24

Updated aggregate passenger gain:

    +504

Updated outcome counts:

- Improved: 22.
- Worse: 1.
- Equal: 277.

Updated regression rate:

    1 / 300 = 0.33%

Experiment 51 Conclusion
------------------------
The prospective framework achieved 93.33% rule compliance, but
B_MINIMAL_LATE cannot be treated as a reliable non-improving predictor.
Its false positives were associated with complete release of a standard
rear-boundary lock.


======================================================================
EXPERIMENT 52 - B-MINIMAL-LATE ROOT-CAUSE ANALYSIS
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency propagation, scheduler, seat-event logic, blocker
mechanics, middle-bank reservation, arbitration and deterministic replay remain
frozen.

Experiment 52 is an observational replay of the Experiment 51 holdout.

No new dataset, boarding policy or prediction threshold is introduced.

Purpose
-------
Explain why Pairs 06 and 19 improved despite being assigned to
B_MINIMAL_LATE.

Replay Scope
------------
Experiment seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This is the same holdout set as Experiment 51.

Therefore Experiment 52 does not add new independent paired evidence.

Root-Cause Classes
------------------

B1_REAR_BOUNDARY_LOCK_RELEASED

- Standard residual family is REAR_BOUNDARY_LOCK.
- Dispersion residual family is COMPLETE_CABIN.

B2_REAR_BOUNDARY_LOCK_PRESERVED

- Standard and dispersion both end in REAR_BOUNDARY_LOCK.

B3_RESIDUAL_TOPOLOGY_UNCHANGED

- Standard and dispersion retain the same residual family.

B4_OTHER_TOPOLOGY_TRANSITION

- Any other B_MINIMAL_LATE residual transition.

Timing Boundary
---------------
Root-cause classification occurs only after both simulations complete.

It is explanatory and must not be presented as a pre-movement predictor.

Measures
--------
For each class the code reports:

1. Scenario count.
2. Improved/worse/equal distribution.
3. Aggregate passenger delta.
4. Average standard dependency depth.
5. Average dispersion dependency depth.
6. Average same-row adjacency breaks.

Primary Question
----------------
Do the two false-positive improvements form a coherent downstream transition
class distinct from the three neutral B_MINIMAL_LATE cases?

Expected Diagnostic Pattern
---------------------------
The two improvements should appear in:

    B1_REAR_BOUNDARY_LOCK_RELEASED

The neutral cases should appear in preserved or unchanged residual classes.

Decision Rule
-------------
Experiment 52 succeeds as explanation if:

- both false positives share the same root-cause class;
- that class contains all B-family aggregate gain;
- the neutral B cases retain stable or preserved topology;
- dependency depth collapses when the rear-boundary lock is released.

Research Boundary
-----------------
No B-family prediction rule will be repaired in Experiment 52.

Any future predictive refinement would require a new hypothesis and another
fresh holdout set.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 51
======================================================================
1. Experiment 51 produced 93.33% prospective rule compliance.
2. A_NO_PERTURBATION remained perfectly neutral.
3. C_BROAD_EARLY remained non-regressing.
4. D_OTHER_PERTURBATION remained neutral.
5. B_MINIMAL_LATE failed in two cases.
6. Both failures converted REAR_BOUNDARY_LOCK to COMPLETE_CABIN.
7. The B family is too coarse for outcome prediction.
8. Experiment 52 performs explanatory replay only.
9. Experiment 52 adds no independent evidence.
10. Amit's movement architecture remains frozen.
11. Independent cumulative evidence now covers 300 paired scenarios.
12. Aggregate passenger improvement is +504.
13. Updated regression rate is approximately 0.33%.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v52 - EXPERIMENT 52 DESIGN
======================================================================


======================================================================
EXPERIMENT 52 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 52 replayed the Experiment 51 holdout seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This replay added no new independent evidence.

Official B-Minimal-Late Root-Cause Matrix
-----------------------------------------

B1_REAR_BOUNDARY_LOCK_RELEASED

- Scenarios: 2.
- Improved/worse/equal: 2/0/0.
- Aggregate delta: +14.
- Average standard dependency depth: 6.00.
- Average dispersion dependency depth: 0.00.
- Average same-row adjacency breaks: 0.00.

B2_REAR_BOUNDARY_LOCK_PRESERVED

- Scenarios: 1.
- Improved/worse/equal: 0/0/1.
- Aggregate delta: 0.
- Average standard dependency depth: 2.00.
- Average dispersion dependency depth: 2.00.
- Average same-row adjacency breaks: 1.00.

B3_RESIDUAL_TOPOLOGY_UNCHANGED

- Scenarios: 2.
- Improved/worse/equal: 0/0/2.
- Aggregate delta: 0.
- Average standard dependency depth: 5.50.
- Average dispersion dependency depth: 5.50.
- Average same-row adjacency breaks: 0.00.

B4_OTHER_TOPOLOGY_TRANSITION

- Scenarios: 0.

B_MINIMAL_LATE total:

- Scenarios: 5.
- Improved/worse/equal: 2/0/3.
- Aggregate delta: +14.

Principal Finding
-----------------
All B_MINIMAL_LATE gain was concentrated in the two cases where dispersion
released a standard REAR_BOUNDARY_LOCK and completed the cabin.

The two improving cases reduced average dependency depth from:

    6.00 -> 0.00

The three neutral cases preserved their residual topology and retained their
dependency depth.

Discussion Incorporated
------------------------
Our discussion distinguished explanation from prediction.

Experiment 51 demonstrated that B_MINIMAL_LATE was too coarse as a
pre-movement predictor.

Experiment 52 demonstrated that the two apparent prediction failures shared a
single coherent downstream transition:

    REAR_BOUNDARY_LOCK -> COMPLETE_CABIN

The Experiment 50 classification was therefore not meaningless; it was
insufficiently specific to determine outcome direction by itself.

The decisive explanatory variable was the downstream residual-topology
transition.

Scientific Boundary
-------------------
The root-cause class is known only after simulation.

It cannot be presented as a pre-movement prediction and must not be used to
retroactively repair Experiment 51.

Experiment 52 Conclusion
------------------------
The two false-positive B_MINIMAL_LATE cases were deterministic topology-release
events, not unexplained anomalies.

This justifies broadening the same transition analysis to every holdout pair.


======================================================================
EXPERIMENT 53 - RESIDUAL-TOPOLOGY TRANSITION ANALYSIS
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency propagation, scheduler, blocker mechanics,
seat-event logic, middle-bank reservation, arbitration and deterministic replay
remain frozen.

Experiment 53 changes no movement rule, admission policy, prediction threshold
or dataset construction method.

Purpose
-------
Generalise the Experiment 52 explanatory method from B_MINIMAL_LATE to all
thirty Experiment 51 holdout pairs.

Replay Scope
------------
Experiment seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This is the same holdout set used by Experiments 51 and 52.

Experiment 53 therefore contributes no new independent paired evidence.

Transition Definition
---------------------
After both paired simulations complete, the code forms:

    STANDARD_RESIDUAL_FAMILY_TO_DISPERSION_RESIDUAL_FAMILY

Examples:

- REAR_BOUNDARY_LOCK_TO_COMPLETE_CABIN.
- LINKED_ROW_EVENT_CHAIN_TO_COMPLETE_CABIN.
- COMPLETE_CABIN_TO_COMPLETE_CABIN.
- REAR_BOUNDARY_LOCK_TO_REAR_BOUNDARY_LOCK.

Measures
--------
For every observed transition, the code reports:

1. Scenario count.
2. Improved/worse/equal distribution.
3. Aggregate passenger delta.
4. Number converted to COMPLETE_CABIN.
5. Average standard dependency depth.
6. Average dispersion dependency depth.
7. Perturbation-family composition.

It also reports:

- changed-topology scenario count and aggregate delta;
- unchanged-topology scenario count and aggregate delta;
- total conversions to COMPLETE_CABIN.

Primary Question
----------------
Is passenger gain concentrated in cases where the residual topology changes,
particularly when a stalled standard run becomes COMPLETE_CABIN?

Decision Rule
-------------
Experiment 53 provides a strong explanatory result if:

- all passenger gain occurs in changed-topology cases;
- unchanged-topology cases remain neutral;
- conversions to COMPLETE_CABIN contain the observed gains;
- dependency depth collapses in successful transitions.

Interpretation Boundary
-----------------------
Residual transition classes are observed after execution.

They are explanatory evidence, not pre-movement predictors.

A future predictive use would require an independently measurable pre-movement
proxy and a new holdout experiment.

Evidence Accounting
-------------------
Experiment 53 is a replay and does not change:

- Independent paired scenarios: 300.
- Aggregate passenger gain: +504.
- Improved/worse/equal: 22/1/277.
- Regression rate: approximately 0.33%.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 52
======================================================================
1. Experiment 52 isolated all B-family gain in topology-release cases.
2. Two REAR_BOUNDARY_LOCK releases produced the full B-family +14.
3. Their average dependency depth collapsed from 6.00 to 0.00.
4. Neutral B cases preserved residual topology and dependency depth.
5. The finding explains Experiment 51 failures without retuning thresholds.
6. Residual topology is more informative after execution than perturbation
   family alone.
7. Experiment 53 expands transition analysis to all thirty holdout pairs.
8. Experiment 53 is explanatory replay, not independent evidence.
9. Amit's movement architecture remains frozen.
10. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v53 - EXPERIMENT 53 DESIGN
======================================================================


======================================================================
EXPERIMENT 53 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 53 replayed the Experiment 51 holdout seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This replay added no new independent evidence.

Official Residual-Topology Transition Matrix
--------------------------------------------

COMPLETE_CABIN_TO_COMPLETE_CABIN

- Scenarios: 20.
- Improved/worse/equal: 0/0/20.
- Aggregate delta: 0.
- Average dependency depth: 0.00 -> 0.00.

ENTRY_SATURATION_WITH_INTERNAL_CHAIN_TO_ENTRY_SATURATION_WITH_INTERNAL_CHAIN

- Scenarios: 1.
- Improved/worse/equal: 0/0/1.
- Aggregate delta: 0.
- Average dependency depth: 4.00 -> 4.00.

LINKED_ROW_EVENT_CHAIN_TO_COMPLETE_CABIN

- Scenarios: 1.
- Improved/worse/equal: 1/0/0.
- Aggregate delta: +10.
- Converted to complete: 1.
- Average dependency depth: 9.00 -> 0.00.

LINKED_ROW_EVENT_CHAIN_TO_LINKED_ROW_EVENT_CHAIN

- Scenarios: 1.
- Improved/worse/equal: 0/0/1.
- Aggregate delta: 0.
- Average dependency depth: 11.00 -> 11.00.

MIXED_OR_UNCLASSIFIED_RESIDUAL_TO_MIXED_OR_UNCLASSIFIED_RESIDUAL

- Scenarios: 1.
- Improved/worse/equal: 0/0/1.
- Aggregate delta: 0.
- Average dependency depth: 2.00 -> 2.00.

MOVING_PASSENGER_YIELD_CHAIN_TO_MOVING_PASSENGER_YIELD_CHAIN

- Scenarios: 2.
- Improved/worse/equal: 0/0/2.
- Aggregate delta: 0.
- Average dependency depth: 3.00 -> 3.00.

REAR_BOUNDARY_LOCK_TO_COMPLETE_CABIN

- Scenarios: 2.
- Improved/worse/equal: 2/0/0.
- Aggregate delta: +14.
- Converted to complete: 2.
- Average dependency depth: 6.00 -> 0.00.

REAR_BOUNDARY_LOCK_TO_REAR_BOUNDARY_LOCK

- Scenarios: 2.
- Improved/worse/equal: 0/0/2.
- Aggregate delta: 0.
- Average dependency depth: 2.50 -> 2.50.

Aggregate Finding
-----------------
Changed residual topology:

- Scenarios: 3.
- Aggregate delta: +24.

Unchanged residual topology:

- Scenarios: 27.
- Aggregate delta: 0.

Conversions to COMPLETE_CABIN:

    3

Principal Finding
-----------------
Every passenger gain in the holdout occurred in a topology-changing scenario.

Every topology-changing scenario converted a standard residual stall into
COMPLETE_CABIN.

All twenty-seven unchanged-topology cases were neutral.

Discussion Incorporated
------------------------
Our discussion concluded that residual-topology transition is a more informative
post-run explanatory variable than perturbation family alone.

The result generalises the Experiment 52 B-family explanation:

- two REAR_BOUNDARY_LOCK releases produced +14;
- one LINKED_ROW_EVENT_CHAIN release produced +10;
- all three successful releases collapsed dependency depth to zero.

This does not establish a pre-movement predictor because the transition is
known only after execution.

Scientific Boundary
-------------------
Experiment 53 is explanatory replay.

It must not be counted as new independent paired evidence and must not be used
to retrospectively alter the Experiment 51 prediction rules.

Experiment 53 Conclusion
------------------------
Passenger gain in this holdout is fully concentrated in residual topology
release to COMPLETE_CABIN.

This supports construction of a detailed signature catalogue for the three
successful conversions.


======================================================================
EXPERIMENT 54 - SUCCESSFUL TRANSITION SIGNATURE CATALOGUE
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency propagation, scheduler, blocker mechanics,
seat-event logic, middle-bank reservation, arbitration and deterministic replay
remain frozen.

Experiment 54 changes no admission policy, dataset construction, threshold or
movement rule.

Purpose
-------
Build a deterministic catalogue describing the structural signatures of every
successful topology conversion identified by Experiment 53.

Replay Scope
------------
Experiment seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This is the same holdout used by Experiments 51-53.

Experiment 54 therefore contributes no new independent paired evidence.

Catalogue Fields
----------------
For each conversion to COMPLETE_CABIN, the code records:

1. Scenario number.
2. Residual transition.
3. Perturbation family.
4. Passenger gain.
5. First admission divergence.
6. Number of differing admission positions.
7. Standard and dispersion dependency depth.
8. Dependency-depth reduction.
9. Standard critical-blocker tree size.
10. Standard critical-blocker tree depth.
11. Standard yield-dependency type.
12. Standard critical prerequisite.
13. Standard required blocker count.
14. Unavailable yield tiles.
15. One-empty-tile diagnostic.
16. One-tile-earlier-hold diagnostic.

Summary Measures
----------------
The catalogue reports:

- successful conversion count;
- aggregate passenger gain;
- average dependency-depth change;
- average critical tree size and depth;
- average required blocker count;
- transition frequencies;
- prerequisite frequencies;
- yield-type frequencies;
- perturbation-family composition.

Primary Question
----------------
Do the three successful conversions share a coherent structural signature,
despite arising from different perturbation families and residual families?

Expected Diagnostic Pattern
---------------------------
The successful cases should share:

- non-zero standard dependency depth;
- zero dispersion dependency depth;
- complete cabin conversion;
- a non-trivial critical blocker tree;
- a measurable prerequisite release.

They need not share the same prerequisite or perturbation family.

Decision Rule
-------------
Experiment 54 succeeds as a catalogue if:

- all three successful Experiment 53 conversions are represented;
- the full +24 gain is captured;
- dependency depth collapses to zero in every entry;
- each entry records sufficient structural detail for later comparison;
- no neutral scenario is mislabelled as a successful signature.

Interpretation Boundary
-----------------------
Catalogue signatures are formed after execution.

They explain successful topology release but are not pre-movement predictors.

A later predictive experiment would require a pre-movement proxy and a fresh
holdout seed.

Evidence Accounting
-------------------
Experiment 54 is a replay and does not alter:

- Independent paired scenarios: 300.
- Aggregate passenger gain: +504.
- Improved/worse/equal: 22/1/277.
- Regression rate: approximately 0.33%.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 53
======================================================================
1. Experiment 53 found three topology-changing cases.
2. All three converted a residual stall into COMPLETE_CABIN.
3. Those cases contained the complete holdout gain of +24.
4. Twenty-seven unchanged-topology cases were neutral.
5. Successful conversion collapsed dependency depth to zero.
6. Residual transition is explanatory, not pre-movement predictive.
7. Experiment 54 catalogues the three successful structural signatures.
8. Experiment 54 adds no independent evidence.
9. Amit's movement architecture remains frozen.
10. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v54 - EXPERIMENT 54 DESIGN
======================================================================


======================================================================
EXPERIMENT 54 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 54 replayed the Experiment 51 holdout seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This replay added no new independent evidence.

Official Successful Transition Signatures
-----------------------------------------

Scenario 06

- Transition: REAR_BOUNDARY_LOCK_TO_COMPLETE_CABIN.
- Perturbation family: B_MINIMAL_LATE.
- Passenger delta: +9.
- First divergence: 102.
- Differing positions: 2.
- Dependency depth: 7 -> 0.
- Critical tree size/depth: 7/7.
- Yield type: NO_YIELD_OCCUPANT.
- Critical prerequisite: ANOTHER_ACTIVE_SEAT_EVENT.
- Required blockers: 2.
- Unavailable yield tile: 31.
- One empty tile would start: false.
- One-tile-earlier hold would preserve: false.

Scenario 19

- Transition: REAR_BOUNDARY_LOCK_TO_COMPLETE_CABIN.
- Perturbation family: B_MINIMAL_LATE.
- Passenger delta: +5.
- First divergence: 103.
- Differing positions: 2.
- Dependency depth: 5 -> 0.
- Critical tree size/depth: 5/5.
- Yield type: NO_YIELD_OCCUPANT.
- Critical prerequisite: INSUFFICIENT_REAR_YIELD_SPACE.
- Required blockers: 2.
- Unavailable yield tile: 31.
- One empty tile would start: false.
- One-tile-earlier hold would preserve: false.

Scenario 25

- Transition: LINKED_ROW_EVENT_CHAIN_TO_COMPLETE_CABIN.
- Perturbation family: C_BROAD_EARLY.
- Passenger delta: +10.
- First divergence: 23.
- Differing positions: 6.
- Dependency depth: 9 -> 0.
- Critical tree size/depth: 9/9.
- Yield type: ACTIVE_EVENT_TRANSIENT.
- Critical prerequisite: YIELD_TILE_OCCUPIED.
- Required blockers: 1.
- Unavailable yield tile: 30.
- One empty tile would start: true.
- One-tile-earlier hold would preserve: true.

Aggregate Catalogue
-------------------
- Successful conversions: 3.
- Aggregate passenger gain: +24.
- Average dependency depth: 7.00 -> 0.00.
- Average critical tree size/depth: 7.00/7.00.
- Average required blockers: 1.67.

Transition counts:

- LINKED_ROW_EVENT_CHAIN_TO_COMPLETE_CABIN: 1.
- REAR_BOUNDARY_LOCK_TO_COMPLETE_CABIN: 2.

Prerequisite counts:

- ANOTHER_ACTIVE_SEAT_EVENT: 1.
- INSUFFICIENT_REAR_YIELD_SPACE: 1.
- YIELD_TILE_OCCUPIED: 1.

Yield-type counts:

- ACTIVE_EVENT_TRANSIENT: 1.
- NO_YIELD_OCCUPANT: 2.

Perturbation-family counts:

- B_MINIMAL_LATE: 2.
- C_BROAD_EARLY: 1.

Principal Finding
-----------------
The three successful conversions shared:

- a non-trivial standard dependency chain;
- a critical tree equal in size and depth to that chain;
- complete dependency collapse to zero;
- conversion to COMPLETE_CABIN.

They did not share one universal prerequisite, yield type or perturbation
family.

Discussion Incorporated
------------------------
Our discussion concluded that the next scientific question should not be
whether a single universal successful signature exists.

Instead, the successful signatures should be contrasted against neutral cases
that began in the same residual family but preserved that topology.

This provides a stronger explanatory comparison than inspecting successful
cases alone.

Scientific Boundary
-------------------
Experiment 54 is a post-run catalogue.

Its signatures cannot be interpreted as pre-movement predictors.

Experiment 54 Conclusion
------------------------
The successful cases form a coherent topology-release class through complete
dependency collapse, while retaining multiple local mechanisms.

This justifies matched neutral contrast analysis.


======================================================================
EXPERIMENT 55 - MATCHED NEUTRAL CONTRAST ANALYSIS
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency propagation, scheduler, blocker mechanics,
seat-event logic, middle-bank reservation, arbitration and deterministic replay
remain frozen.

Experiment 55 changes no movement rule, dataset policy, admission threshold or
accepted boarding order.

Purpose
-------
Compare each successful topology-release case with a unique neutral scenario
that began in the same standard residual family but preserved that topology
under dispersion.

Replay Scope
------------
Experiment seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This is the same holdout replay used by Experiments 51-54.

Experiment 55 therefore contributes no new independent paired evidence.

Control Eligibility
-------------------
A matched control must:

1. Begin in the same standard residual family.
2. Retain that residual family under dispersion.
3. Have zero passenger delta.
4. Not already be assigned to another successful case.

Matching Distance
-----------------
Within the same standard residual family, the deterministic score considers:

- perturbation-family agreement;
- first admission divergence;
- differing admission positions;
- standard dependency depth;
- critical blocker tree size;
- required blocker count;
- yield-dependency type;
- critical prerequisite.

Controls are assigned uniquely.

Primary Question
----------------
Which structural differences remain after successful conversions are compared
with preserved-topology controls from the same residual family?

Measures
--------
For each matched contrast, the code reports:

- successful and control scenario numbers;
- match distance;
- successful and control transitions;
- perturbation families;
- passenger delta;
- first divergence;
- differing positions;
- standard dependency depth;
- standard critical tree size;
- required blocker count;
- yield type;
- prerequisite;
- dispersion dependency depth.

Aggregate measures report average successful/control values.

Decision Rule
-------------
Experiment 55 succeeds if:

- all three successful cases receive a unique same-family control;
- the full +24 passenger gain is captured;
- no complete-cabin neutral case is used as a control;
- successful and control structural differences are reported transparently;
- the analysis remains explanatory rather than predictive.

Interpretation Boundary
-----------------------
Control selection uses post-run residual families and diagnostics.

Experiment 55 is matched explanatory analysis, not prospective prediction.

Evidence Accounting
-------------------
Experiment 55 is a replay and does not alter:

- Independent paired scenarios: 300.
- Aggregate passenger gain: +504.
- Improved/worse/equal: 22/1/277.
- Regression rate: approximately 0.33%.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 54
======================================================================
1. Experiment 54 catalogued all three successful conversions.
2. The complete holdout gain of +24 was represented.
3. Dependency depth collapsed from an average 7.00 to 0.00.
4. Critical tree size and depth averaged 7.00/7.00.
5. No single prerequisite or perturbation family explained every success.
6. Experiment 55 introduces unique same-residual-family neutral controls.
7. Experiment 55 remains post-run explanatory replay.
8. Amit's movement architecture remains frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v55 - EXPERIMENT 55 DESIGN
======================================================================


======================================================================
EXPERIMENT 55 - OFFICIAL FINDINGS
======================================================================

Execution Status
----------------
Experiment 55 replayed the Experiment 51 holdout seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This replay added no new independent evidence.

Official Matched Contrasts
--------------------------

Successful Scenario 06 -> Neutral Control 24

- Standard residual family: REAR_BOUNDARY_LOCK.
- Success/control perturbation family:
  B_MINIMAL_LATE / B_MINIMAL_LATE.
- Success/control passenger delta: +9 / 0.
- First divergence: 102 / 114.
- Differing positions: 2 / 2.
- Standard dependency depth: 7 / 2.
- Standard critical tree size: 7 / 2.
- Required blockers: 2 / 2.
- Yield type:
  NO_YIELD_OCCUPANT / NO_YIELD_OCCUPANT.
- Prerequisite:
  ANOTHER_ACTIVE_SEAT_EVENT / INSUFFICIENT_REAR_YIELD_SPACE.
- Dispersion dependency depth: 0 / 2.

Successful Scenario 19 -> Neutral Control 10

- Standard residual family: REAR_BOUNDARY_LOCK.
- Success/control perturbation family:
  B_MINIMAL_LATE / D_OTHER_PERTURBATION.
- Success/control passenger delta: +5 / 0.
- First divergence: 103 / 10.
- Differing positions: 2 / 2.
- Standard dependency depth: 5 / 3.
- Standard critical tree size: 5 / 3.
- Required blockers: 2 / 3.
- Yield type:
  NO_YIELD_OCCUPANT / NO_YIELD_OCCUPANT.
- Prerequisite:
  INSUFFICIENT_REAR_YIELD_SPACE / INSUFFICIENT_REAR_YIELD_SPACE.
- Dispersion dependency depth: 0 / 3.

Successful Scenario 25 -> Neutral Control 22

- Standard residual family: LINKED_ROW_EVENT_CHAIN.
- Success/control perturbation family:
  C_BROAD_EARLY / B_MINIMAL_LATE.
- Success/control passenger delta: +10 / 0.
- First divergence: 23 / 129.
- Differing positions: 6 / 2.
- Standard dependency depth: 9 / 11.
- Standard critical tree size: 9 / 11.
- Required blockers: 1 / 2.
- Yield type:
  ACTIVE_EVENT_TRANSIENT / ACTIVE_EVENT_TRANSIENT.
- Prerequisite:
  YIELD_TILE_OCCUPIED / YIELD_TILE_OCCUPIED.
- Dispersion dependency depth: 0 / 11.

Aggregate Matched-Control Findings
----------------------------------
- Successful cases: 3.
- Matched contrasts: 3.
- Unmatched successes: 0.
- Captured successful gain: +24.
- Average first divergence success/control: 76.00 / 84.33.
- Average differing positions success/control: 3.33 / 2.00.
- Average standard dependency depth success/control: 7.00 / 5.33.
- Average standard critical tree size success/control: 7.00 / 5.33.
- Average required blockers success/control: 1.67 / 2.33.

Principal Finding
-----------------
No single raw structural magnitude universally separated successful and neutral
cases.

The rear-boundary successes had larger standard dependency trees than their
controls, but the linked-row success had a smaller tree than its control.

Required blocker count also failed to separate the groups consistently.

The stable distinction was post-run:

- every successful case collapsed dispersion dependency depth to zero;
- every matched control preserved non-zero dependency depth.

Discussion Incorporated
------------------------
Our discussion identified Experiment 55 as one of the strongest explanatory
experiments in the series because it introduced scientifically appropriate
same-residual-family controls.

This removed several simple alternative explanations:

- larger blocker trees are not universally sufficient;
- deeper dependency chains are not universally sufficient;
- more blockers are not universally sufficient;
- one perturbation family is not universally responsible.

The successful mechanism is therefore better described as dependency-topology
release than as a threshold on one scalar measurement.

Scientific Boundary
-------------------
The controls were selected using post-run residual families and diagnostics.

Experiment 55 is therefore a controlled explanatory validation, not a
prospective prediction experiment.

Experiment 55 Conclusion
------------------------
Matched neutral controls strengthen the conclusion that complete dependency
collapse distinguishes successful topology release from neutral preservation.

This supports a final synthesis of all holdout pathways rather than further
isolated signature refinement.


======================================================================
EXPERIMENT 56 - EXPLANATORY PATHWAY SYNTHESIS
======================================================================

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency propagation, scheduler, blocker mechanics,
seat-event logic, middle-bank reservation, arbitration and deterministic replay
remain frozen.

Experiment 56 changes no admission rule, movement rule, threshold, dataset
construction or accepted boarding policy.

Purpose
-------
Integrate the Experiment 51 holdout into a deterministic explanatory pathway
model covering complete stability, preserved residual topology, successful
topology release, partial improvement and regression.

Replay Scope
------------
Experiment seed:

    9783790101901

Scenario pairs:

    30

Total executions:

    60

This is the same holdout replay used by Experiments 51-55.

Experiment 56 therefore contributes no new independent paired evidence.

Explanatory Pathways
--------------------

P0_UNPERTURBED_COMPLETE_STABILITY

- No admission perturbation.
- Standard and dispersion both complete.
- Passenger delta must be zero.

P1_PERTURBED_COMPLETE_STABILITY

- Admission order changes.
- Standard and dispersion both complete.
- Passenger delta must be zero.

P2_RESIDUAL_TOPOLOGY_PRESERVED

- Standard and dispersion retain the same non-complete residual family.
- Passenger delta must be zero.

P2_TOPOLOGY_CHANGED_WITHOUT_GAIN

- Residual family changes.
- Passenger delta remains zero.

P3_GAIN_WITHOUT_FAMILY_CHANGE

- Passenger count improves.
- Residual family does not change.

P3_PARTIAL_TOPOLOGY_IMPROVEMENT

- Residual family changes.
- Passenger count improves.
- Dispersion does not reach COMPLETE_CABIN.

P4_SUCCESSFUL_TOPOLOGY_RELEASE

- Standard run retains a residual stall.
- Dispersion converts to COMPLETE_CABIN.
- Passenger delta is positive.
- Dispersion dependency depth must equal zero.

P5_REGRESSION

- Dispersion seats fewer passengers than standard.

Primary Questions
-----------------
1. Is all passenger gain concentrated in successful topology release?
2. Are preserved residual-topology cases neutral?
3. Are regressions absent?
4. Does every scenario satisfy the integrity rule of its assigned pathway?

Output
------
For every scenario pair, the code reports:

- pathway;
- perturbation family;
- residual transition;
- passenger delta;
- first divergence;
- differing positions;
- dependency-depth transition;
- critical-tree-size transition;
- standard prerequisite;
- pathway integrity result.

The final synthesis reports pathway counts, aggregate passenger deltas,
perturbation-family composition and four closure checks.

Decision Rule
-------------
Experiment 56 supports explanatory closure if:

- all +24 passenger gain appears in P4;
- all P2 preserved-residual cases have zero aggregate gain;
- no P5 regressions occur;
- all 30 pathway assignments pass integrity validation.

Interpretation Boundary
-----------------------
Pathways are assigned after both runs complete.

Experiment 56 is an explanatory synthesis and closure audit. It does not create
a pre-movement controller.

Evidence Accounting
-------------------
Experiment 56 is a replay and does not alter:

- Independent paired scenarios: 300.
- Aggregate passenger gain: +504.
- Improved/worse/equal: 22/1/277.
- Regression rate: approximately 0.33%.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 55
======================================================================
1. All three successful cases received unique same-residual-family controls.
2. The complete +24 holdout gain was captured.
3. No scalar measure universally separated successful and neutral cases.
4. Successful cases collapsed dependency depth to zero.
5. Neutral controls preserved non-zero dependency depth.
6. Experiment 56 integrates all 30 holdout pairs into explanatory pathways.
7. Experiment 56 is a closure audit, not a prospective controller.
8. Amit's movement architecture remains frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v56 - EXPERIMENT 56 DESIGN
======================================================================

======================================================================
EXPERIMENT 56 OFFICIAL FINDINGS AND SYNTHESIS DISCUSSION
======================================================================
Experiment 56 completed the scientific synthesis of Experiments 51-56.

Official pathway distribution
-----------------------------
P0_UNPERTURBED_COMPLETE_STABILITY: 8 scenarios, aggregate delta +0.
P1_PERTURBED_COMPLETE_STABILITY: 12 scenarios, aggregate delta +0.
P2_RESIDUAL_TOPOLOGY_PRESERVED: 7 scenarios, aggregate delta +0.
P4_SUCCESSFUL_TOPOLOGY_RELEASE: 3 scenarios, aggregate delta +24.

Improved/worse/equal: 3/0/27.
Integrity passes/failures: 30/0.

All four synthesis checks passed:

- all passenger gain was concentrated in successful topology release;
- preserved residual topology was completely neutral;
- no regressions occurred;
- every pathway assignment satisfied its integrity rule.

Discussion incorporated after Experiment 56
-------------------------------------------
Experiment 56 is the explanatory closure point for the Experiment 51 holdout.
Every one of the 30 paired scenarios was assigned to a deterministic pathway;
no unknown or unclassified category was required. The occupied pathways form a
complete partition of observed behaviour: complete-cabin stability, preserved
residual topology and successful topology release.

The result does not establish a prospective controller. Pathways are assigned
after both simulations complete. The correct scientific claim is therefore
that the framework is explanatorily complete for this holdout, not that it can
predict or control unseen scenarios before movement begins.

======================================================================
RESEARCH EXPERIMENT 57 - DETERMINISTIC EVIDENCE SUFFICIENCY AUDIT
======================================================================
Build tag:

    STRESS-EXP57-DETERMINISTIC-EVIDENCE-SUFFICIENCY-AUDIT-001

Purpose
-------
Determine which evidence categories are indispensable for the Experiment 56
pathway explanation and which categories are supporting or optional.

Replay Scope
------------
Experiment seed: 9783790101901.
Scenario pairs: 30.
Total executions: 60.

The Experiment 51 holdout is replayed unchanged. Experiment 57 contributes no
new independent paired evidence.

Method
------
The Experiment 56 pathway is retained as the baseline. One evidence category
is withheld at a time. The code enumerates the pathway assignments still
consistent with the retained evidence.

Classification rule
-------------------
REQUIRED:
Withholding the category leaves at least one scenario with more than one
possible pathway.

SUPPORTING:
All pathway assignments remain unique, but withholding the category removes or
weakens an independent integrity or corroboration check.

OPTIONAL:
All pathway assignments remain unique and no pathway-integrity test is lost.
The evidence may remain diagnostically useful.

Evidence categories audited
---------------------------
- perturbation family;
- first divergence;
- differing positions;
- passenger delta;
- dependency depth;
- critical-tree size;
- critical-tree depth;
- yield-dependency type;
- final prerequisite;
- required blocker count;
- residual stall family;
- residual-topology transition;
- COMPLETE_CABIN conversion;
- pathway-integrity rules.

Primary questions
-----------------
1. Which categories are required to preserve unique pathway assignment?
2. Which categories only corroborate the assigned pathway?
3. Which diagnostics can be removed without changing explanatory assignment?
4. Does every ablation retain the original baseline among its possible paths?

Scientific boundary
-------------------
Experiment 57 is a post-run evidence audit. It does not alter Amit's movement
architecture, admission queues, scheduler, dependency propagation, seat-event
mechanics, blocker behaviour, middle-bank reservation or arbitration.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 56
======================================================================
1. Experiment 56 completely partitioned all 30 holdout pairs.
2. All +24 holdout gain occurred only in P4 topology release.
3. Preserved residual topology remained neutral.
4. No regression pathway occurred.
5. All 30 pathway integrity checks passed.
6. Experiment 57 now tests evidence necessity rather than adding a mechanism.
7. Experiment 57 is replay-only and explanatory.
8. Amit's movement architecture remains frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v57 - EXPERIMENT 57 DESIGN
======================================================================


======================================================================
EXPERIMENT 57 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 57 completed the deterministic evidence sufficiency audit on the
unchanged 30-pair Experiment 51 holdout.

Official evidence classification
--------------------------------
REQUIRED (3 categories)
- DIFFERING_POSITIONS: withholding reduced unique assignment to 10/30 and
  made 20/30 scenarios ambiguous.
- PASSENGER_DELTA: withholding reduced unique assignment to 0/30 and made all
  30 scenarios ambiguous.
- RESIDUAL_STALL_FAMILY: withholding reduced unique assignment to 0/30 and
  made all 30 scenarios ambiguous.

SUPPORTING (4 categories)
- DEPENDENCY_DEPTH.
- RESIDUAL_TOPOLOGY_TRANSITION.
- COMPLETE_CABIN_CONVERSION.
- PATHWAY_INTEGRITY_RULES.

OPTIONAL FOR PATHWAY ASSIGNMENT IN THIS HOLDOUT (7 categories)
- PERTURBATION_FAMILY.
- FIRST_DIVERGENCE.
- CRITICAL_TREE_SIZE.
- CRITICAL_TREE_DEPTH.
- YIELD_DEPENDENCY_TYPE.
- FINAL_PREREQUISITE.
- REQUIRED_BLOCKER_COUNT.

All baseline assignments were retained under every ablation, and the movement
architecture remained frozen.

Discussion incorporated after Experiment 57
-------------------------------------------
Experiment 57 distinguishes evidence needed to identify pathway membership
from evidence used to explain or corroborate the underlying mechanism.
Passenger-delta sign, the existence of an admission-order difference and the
residual-stall-family identity are indispensable for unique classification in
this holdout. Dependency collapse, topology transition, COMPLETE_CABIN
conversion and pathway-integrity rules remain scientifically important because
they validate the causal interpretation, even though they are not separately
required once the three indispensable categories are retained.

The seven optional categories should not be described as scientifically
unimportant. They retain diagnostic value for mechanism inspection and matched
case explanation; they simply do not change pathway membership in this fixed
holdout.

The Experiment 57 instrumentation was non-invasive. It reproduced the
Experiment 56 pathway distribution and +24 holdout gain while changing no
movement or admission behaviour.

======================================================================
RESEARCH EXPERIMENT 58 - DETERMINISTIC EXPLANATORY ROBUSTNESS AUDIT
======================================================================
Build tag:

    STRESS-EXP58-DETERMINISTIC-EXPLANATORY-ROBUSTNESS-AUDIT-001

Purpose
-------
Test whether the explanatory pathway framework remains invariant when the same
retained evidence predicates are evaluated in different deterministic orders.

Replay Scope
------------
Experiment seed: 9783790101901.
Scenario pairs: 30.
Total executions: 60.

The Experiment 51 holdout is replayed unchanged. Experiment 58 contributes no
new independent paired evidence.

Method
------
The Experiment 56 pathway assignment remains the reference. For every pair,
Experiment 58 evaluates the complete pathway predicate set in five different
orders:

1. regression/release-first order;
2. ascending pathway order;
3. reverse grouped order;
4. release-and-preservation-first order;
5. minimal/stability-first order.

Each order inspects the same evidence values. It changes only predicate
inspection sequence.

A pair is STABLE only if all five orders:

- yield exactly one matching pathway;
- reproduce the Experiment 56 baseline pathway;
- reproduce the same pathway-integrity result.

Primary questions
-----------------
1. Are the pathway predicates mutually exclusive for every holdout pair?
2. Are the predicates collectively complete for every holdout pair?
3. Does evaluation order ever change pathway assignment?
4. Does evaluation order ever change integrity outcome?

Decision rule
-------------
The explanatory framework is robust for this holdout only if:

- all 150 order evaluations are unique;
- all 150 reproduce the baseline pathway;
- all 150 reproduce the baseline integrity result;
- all 30 scenarios are classified STABLE;
- no order-specific failure is recorded.

Scientific boundary
-------------------
Experiment 58 is an internal consistency audit of a post-run explanatory model.
It does not establish prospective prediction, causal intervention or
independent generalisation. Amit's movement architecture, admission queues,
scheduler, seat-event logic, dependency propagation and arbitration remain
frozen.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 57
======================================================================
1. Experiment 56 completely partitioned all 30 holdout pairs.
2. Experiment 57 identified three indispensable classification categories.
3. Four categories remain supporting corroboration and integrity evidence.
4. Seven categories remain diagnostically useful but optional for membership.
5. All Experiment 57 baseline assignments were retained.
6. Experiment 58 now tests order invariance, exclusivity and completeness.
7. Experiment 58 is replay-only and explanatory.
8. Amit's movement architecture remains frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v58 - EXPERIMENT 58 DESIGN
======================================================================


======================================================================
EXPERIMENT 58 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 58 completed the deterministic explanatory robustness audit on the
unchanged 30-pair Experiment 51 holdout.

Official robustness result
--------------------------
- Scenario pairs audited: 30.
- Evaluation orders per pair: 5.
- Total pathway evaluations: 150.
- Identical baseline pathways: 150/150.
- Changed pathways: 0.
- Non-unique evaluations: 0.
- Integrity agreements: 150/150.
- Robust scenarios: 30/30.
- Order failures: NONE.
- frameworkRobust=true.

Discussion incorporated after Experiment 58
-------------------------------------------
Experiment 58 demonstrates that pathway membership does not depend on the
sequence in which the full explanatory predicates are inspected. Every one of
the five deterministic evaluation orders returned the same unique pathway and
the same integrity result for every holdout pair.

The finding strengthens the internal consistency of the explanatory model. It
shows that the pathway framework behaves as a stable classification system,
rather than as an order-dependent decision process. The audit also reproduced
the Experiment 56 distribution of eight P0 cases, twelve P1 cases, seven P2
preserved-residual cases and three P4 successful releases.

The result remains explanatory and replay-bound. Evaluation-order invariance
does not establish prospective prediction, causal intervention or independent
generalisation. The instrumentation was non-invasive and Amit's movement
architecture remained frozen.

======================================================================
RESEARCH EXPERIMENT 59 - MINIMAL EXPLANATORY CORE RECONSTRUCTION AUDIT
======================================================================
Build tag:

    STRESS-EXP59-MINIMAL-EXPLANATORY-CORE-RECONSTRUCTION-AUDIT-001

Purpose
-------
Test whether the three evidence categories classified as REQUIRED by
Experiment 57 can, by themselves, reconstruct every Experiment 56 pathway
assignment in the unchanged holdout.

Replay Scope
------------
Experiment seed: 9783790101901.
Scenario pairs: 30.
Total executions: 60.

The Experiment 51 holdout is replayed unchanged. Experiment 59 contributes no
new independent paired evidence.

Minimal explanatory core
------------------------
The reconstruction classifier uses only:

1. DIFFERING_POSITIONS;
2. PASSENGER_DELTA;
3. RESIDUAL_STALL_FAMILY.

The remaining eleven audited categories are deliberately excluded from the
minimal classifier. They remain available only for comparison, integrity
corroboration and mechanism interpretation.

Method
------
For every pair, the code:

- obtains the full Experiment 56 baseline pathway;
- reconstructs a pathway from the three-category minimal core;
- checks exact pathway agreement;
- compares full-path and reconstructed-path integrity outcomes;
- reports mismatches, pathway distribution and evidence compression.

Decision rule
-------------
The minimal core is descriptively sufficient for this holdout only if:

- all 30 baseline pathways are reconstructed exactly;
- no scenario is ambiguous or mismatched;
- all 30 integrity outcomes agree;
- the movement architecture remains frozen.

Primary questions
-----------------
1. Can the three REQUIRED categories reproduce all baseline pathways?
2. Does the minimal reconstruction preserve the occupied pathway distribution?
3. Are integrity outcomes retained after reconstruction?
4. How much of the fourteen-category evidence profile can be omitted from the
   membership classifier without changing the result?

Scientific boundary
-------------------
Experiment 59 tests descriptive reconstruction within a fixed replay holdout.
It does not claim that the minimal core is a complete mechanism explanation.
The eleven omitted categories remain scientifically important for causal
interpretation, corroboration, diagnostic inspection and auditability.

The experiment does not establish prospective prediction or independent
generalisation and does not alter Amit's movement architecture.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 58
======================================================================
1. Experiment 56 completely partitioned all 30 holdout pairs.
2. Experiment 57 identified three indispensable membership categories.
3. Experiment 58 confirmed 150/150 order-invariant pathway evaluations.
4. All 30 Experiment 58 scenarios were stable and retained integrity.
5. Experiment 59 now reconstructs the framework from the three-category core.
6. The omitted categories remain mechanism and corroboration evidence.
7. Experiment 59 is replay-only and explanatory.
8. Amit's movement architecture remains frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v59 - EXPERIMENT 59 DESIGN
======================================================================

======================================================================
EXPERIMENT 59 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 59 completed the minimal explanatory core reconstruction audit on
30 unchanged Experiment 51 holdout pairs.

Official reconstruction result
------------------------------
- Scenario pairs audited: 30.
- Exact baseline reconstructions: 30/30.
- Reconstruction mismatches: 0.
- Integrity agreements: 30/30.
- Minimal pathway distribution: P0=8, P1=12, P2=7 and P4=3.
- Evidence retained by the membership classifier: 3/14 (21.43%).
- Evidence omitted from the membership classifier: 11/14 (78.57%).
- minimalCoreSufficient=true.
- movementArchitectureFrozen=true.

Discussion incorporated after Experiment 59
-------------------------------------------
Experiment 59 demonstrates that the established explanatory pathway taxonomy
can be reconstructed from the three REQUIRED Experiment 57 categories while
the complete diagnostic evidence remains preserved for verification,
corroboration and mechanism interpretation.

The result establishes a clean separation between evidence collection, evidence
preservation and pathway reconstruction. The classifier uses only differing
positions, passenger delta and residual stall family, but the simulator still
records dependency chains, critical-blocker trees, prerequisites, topology
transitions and all other diagnostic evidence.

The conclusion is deliberately limited to this deterministic holdout and the
established explanatory taxonomy. It does not show that the three categories
are a universal minimum for future datasets, nor does it establish prospective
prediction or independent generalisation. The eleven omitted categories remain
scientifically important even though they are not required for membership
assignment in this holdout.

======================================================================
RESEARCH EXPERIMENT 60 - MINIMAL EXPLANATORY CORE IRREDUCIBILITY AUDIT
======================================================================
Build tag:

    STRESS-EXP60-MINIMAL-EXPLANATORY-CORE-IRREDUCIBILITY-AUDIT-001

Purpose
-------
Determine whether the three-category minimal explanatory core validated by
Experiment 59 is irreducible within the unchanged holdout.

Replay Scope
------------
Experiment seed: 9783790101901.
Scenario pairs: 30.
Total executions: 60.

The Experiment 51 holdout is replayed unchanged. Experiment 60 contributes no
new independent paired evidence.

Method
------
Experiment 60 first reconfirms the complete three-category core:

1. DIFFERING_POSITIONS;
2. PASSENGER_DELTA;
3. RESIDUAL_STALL_FAMILY.

It then withholds each category separately and enumerates every explanatory
pathway compatible with the two retained categories and the unknown value of
the withheld category.

A category is indispensable within this holdout when its removal creates at
least one non-unique pathway assignment while the established baseline pathway
remains among the compatible candidates.

Primary questions
-----------------
1. Does the full three-category core still reconstruct all 30 pathways?
2. Does withholding DIFFERING_POSITIONS create ambiguity?
3. Does withholding PASSENGER_DELTA create ambiguity?
4. Does withholding RESIDUAL_STALL_FAMILY create ambiguity?
5. Is every baseline pathway retained among the compatible candidates?

Decision rule
-------------
The minimal core is irreducible for this holdout only if:

- the complete core again reconstructs all 30 baseline pathways;
- all 30 integrity outcomes agree;
- every one-category ablation creates ambiguity in at least one scenario;
- every ablation retains the baseline pathway among its candidates;
- the movement architecture remains frozen.

Scientific boundary
-------------------
Irreducibility is a holdout-specific membership result. It does not establish a
universal minimum for new datasets, and it does not reduce the importance of
supporting and optional diagnostic evidence. The audit is post-run,
non-predictive and non-interventional.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 59
======================================================================
1. Experiment 56 established the complete explanatory pathway taxonomy.
2. Experiment 57 identified three REQUIRED membership categories.
3. Experiment 58 confirmed order-invariant pathway evaluation.
4. Experiment 59 reconstructed all 30 pathways from the three-category core.
5. Full diagnostic evidence remains preserved despite classifier compression.
6. Experiment 60 now tests whether any one core category can be removed.
7. Experiment 60 is replay-only and explanatory.
8. Amit's movement architecture remains frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v60 - EXPERIMENT 60 DESIGN
======================================================================

======================================================================
EXPERIMENT 60 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 60 completed the minimal explanatory core irreducibility audit on
30 unchanged Experiment 51 holdout pairs.

Official irreducibility result
------------------------------
- Full-core baseline reconstructions: 30/30.
- Full-core integrity agreements: 30/30.
- Full-core mismatches: 0.
- Withholding DIFFERING_POSITIONS produced 20 ambiguous assignments.
- Withholding PASSENGER_DELTA produced 30 ambiguous assignments.
- Withholding RESIDUAL_STALL_FAMILY produced 30 ambiguous assignments.
- Every established baseline pathway remained among the compatible candidates.
- Categories producing ambiguity when withheld: 3/3.
- minimalCoreIrreducible=true.
- movementArchitectureFrozen=true.

Discussion incorporated after Experiment 60
-------------------------------------------
Experiment 60 closes the explanatory-core validation sequence. Within the fixed
holdout and established taxonomy, none of the three retained membership
categories can be removed without losing unique pathway assignment in at least
one scenario. The result remains a holdout-specific descriptive finding and does
not claim a universal minimum for future datasets.

The wider discussion also identified the next major abstraction layer. Before a
behavioural SME defines walking speed, hesitation, luggage handling, compliance,
group movement or reduced mobility, the simulator can isolate airline policy as
a separate deterministic input layer. Examples include boarding by sections,
back-to-front release, window-middle-aisle ordering, priority release and other
admission-order policies.

This ordering is scientifically preferable because airline policy and human
behaviour are different independent variables. Policy can first be represented
and compared under the present deterministic engine. Human properties and the
best achievable asynchronous execution model can then be introduced later and
measured against the deterministic policy baseline.

The SME-facing abstraction should return operational information rather than
internal implementation detail alone. Relevant feedback includes completion,
boarding ticks or elapsed time, passengers seated, improved/equal/worse paired
outcomes, bottleneck location and duration, dependency regions, critical
passengers, seat-event delay, residual topology and an explanation of why a
policy succeeded or failed.

======================================================================
RESEARCH EXPERIMENT 61 - AIRLINE POLICY ABSTRACTION READINESS AUDIT
======================================================================
Build tag:

    STRESS-EXP61-AIRLINE-POLICY-ABSTRACTION-READINESS-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's original algorithm separates the movement rule from the occupancy data
supplied to it. Experiment 61 extends that same separation to airline operations:
the admission order is treated as a policy-layer input, while the movement,
dependency and diagnostic architecture remains unchanged.

Purpose
-------
Determine whether the established deterministic paired framework can present an
admission-order transformation as an airline-policy abstraction and return a
clear operational feedback contract suitable for later SME review.

Policy abstraction used in this audit
-------------------------------------
Policy A:

    REFERENCE DETERMINISTIC ADMISSION ORDER

Policy B:

    TWO-ROW RISK-DISPERSION ADMISSION ORDER

Policy B is not claimed to be an airline-approved policy. It is used as the
existing deterministic admission-order transformation through which the policy
abstraction and feedback interface can be validated.

Replay Scope
------------
Experiment seed: 9783790101901.
Scenario pairs: 30.
Total executions: 60.

The Experiment 51 holdout is replayed unchanged. Experiment 61 contributes no
new independent paired evidence.

Method
------
For every scenario pair, Experiment 61 records:

- policy identity;
- exact admission perturbation signature;
- cabin completion and passengers seated;
- simulation ticks;
- improved, equal or worse paired outcome;
- residual-family transition;
- explanatory pathway;
- dependency-region and intervention evidence;
- critical-blocker and prerequisite evidence when incomplete.

The batch summary then presents a policy-level SME feedback contract while
explicitly confirming that human behaviour and asynchronous execution remain
deferred.

Primary questions
-----------------
1. Can admission ordering be represented as a separate deterministic policy
   layer without changing movement?
2. Can paired policy outcomes be translated into operationally understandable
   feedback?
3. Can the complete explanatory diagnostics remain available beneath that
   abstraction?
4. Can policy effects remain separated from future human-behaviour assumptions?
5. Can the deterministic replay identity serve as a baseline for later
   asynchronous and stochastic comparison?

SME feedback contract
---------------------
The policy abstraction should provide:

1. Policy identity and exact admission transformation.
2. Cabin completion, passengers seated and elapsed simulation ticks.
3. Improved, equal and worse paired outcomes.
4. Residual bottleneck family and topology transition.
5. Dependency regions, confirmation, competition and intervention evidence.
6. Critical blocker, prerequisite and yield-space explanation when incomplete.
7. Separation of policy effect from later human-behaviour assumptions.
8. Replay identity for later asynchronous or stochastic comparison.

Scientific boundary
-------------------
Experiment 61 does not validate a real airline boarding policy and does not
introduce passengers with behavioural properties. It does not implement
asynchronous ticks. Airline operations SMEs remain responsible for defining
valid policy, and behavioural SMEs remain responsible for defining human
properties and distributions.

The experiment validates only that the architecture can isolate, compare and
explain a policy-layer admission transformation before those later variables are
introduced.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 60
======================================================================
1. Experiments 56-60 completed the explanatory pathway validation sequence.
2. Experiment 60 confirmed the three-category core is irreducible in the fixed
   holdout.
3. The next abstraction layer is airline policy, not human behaviour.
4. Policy and human behaviour must remain separate independent variables.
5. Experiment 61 treats admission order as a deterministic policy-layer input.
6. The SME-facing output reports operational outcomes and causal diagnostics.
7. Human behaviour and asynchronous execution remain deliberately deferred.
8. Amit's movement architecture remains frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v61 - EXPERIMENT 61 DESIGN
======================================================================

======================================================================
EXPERIMENT 61 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 61 completed the Airline Policy Abstraction Readiness Audit on the
unchanged 30-pair deterministic holdout.

Official result
---------------
- Scenario pairs: 30.
- Total executions: 60.
- Improved/equal/worse under the risk-dispersion policy: 3/27/0.
- Aggregate seated-passenger difference: +24.
- Complete cabins: 20 reference / 23 policy.
- Aggregate ticks: 43,938 reference / 39,546 policy.
- Pathway distribution: P0=8, P1=12, P2=7, P4=3.
- deterministicPolicyLayerIsolated=true.
- pairedOperationalFeedbackAvailable=true.
- explanatoryDiagnosticsPreserved=true.
- policyAbstractionReadyForSMEReview=true.

Discussion incorporated after Experiment 61
-------------------------------------------
The policy layer should precede human-behaviour modelling. Airline policy and
human behaviour are separate independent variables: an airline or operations SME
defines the admission policy, while a behavioural SME later defines properties
such as walking speed, luggage handling, hesitation, compliance, group movement
and reduced mobility.

The simulator's feedback to SMEs should remain operational and understandable:
completion, elapsed ticks or time, passengers seated, bottleneck family,
dependency regions, critical blockers, seat-event delays, paired policy effects
and a causal explanation of why completion improved, remained unchanged or
failed. The full low-level diagnostics remain available beneath that abstraction.

Experiment 61 did not claim that the existing risk-dispersion transformation is
an airline-approved policy. It demonstrated that a deterministic policy can be
isolated, replayed, compared and explained before human properties or
asynchronous execution are introduced.

======================================================================
RESEARCH EXPERIMENT 62 - DETERMINISTIC POLICY PLUG-IN INTERFACE CONFORMANCE AUDIT
======================================================================
Build tag:

    STRESS-EXP62-DETERMINISTIC-POLICY-PLUGIN-INTERFACE-CONFORMANCE-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement rules remain the shared engine. Experiment 62 formalises the
boundary through which different admission policies supply ordered passenger
data to that engine. Policy selection changes input ordering only; it cannot
change aisle movement, row entry, blocker displacement, yield-space mechanics,
reservations, dependency propagation or stall detection.

Purpose
-------
Determine whether multiple deterministic admission policies can be registered,
selected and executed through one common plug-in interface while preserving one
frozen movement engine and one common SME feedback schema.

Registered deterministic plug-ins
----------------------------------
1. REFERENCE_ADMISSION_ORDER
   -> STANDARD_DETERMINISTIC_DATASET

2. TWO_ROW_RISK_DISPERSION_ADMISSION_ORDER
   -> FINAL_TWO_ROW_RISK_DISPERSION_DATASET

These are software-conformance examples. Neither is claimed to be an approved
real-world airline procedure.

Method
------
- Register each policy with a stable identity.
- Map each policy to exactly one deterministic dataset mode.
- Instantiate both policies through the same simulation factory.
- Replay the unchanged 30-pair holdout.
- Confirm all 60 policy routes match their registered dataset mode.
- Confirm both policies return the same operational and explanatory fields.
- Confirm no human behaviour or asynchronous execution is introduced.

Primary questions
-----------------
1. Can policies be exchanged without editing the movement engine?
2. Does every policy have a stable and auditable identity?
3. Does every policy route to exactly one deterministic admission mode?
4. Is the feedback contract identical across policies?
5. Can future SME-defined policies be added at this boundary without changing
   passenger movement mechanics?

Decision rule
-------------
The deterministic policy plug-in interface is ready only if:

- both registered policies are present;
- all 60 executions route through the correct registered policy mapping;
- a common factory is used;
- the common feedback schema is preserved;
- the movement architecture remains frozen;
- human behaviour and asynchronous execution remain deferred.

Scientific boundary
-------------------
Experiment 62 validates a software and research boundary, not airline policy
quality. It contributes no new independent paired evidence and makes no claim
about passenger psychology, realistic elapsed seconds or airline deployment.
Operational policies must later be supplied and reviewed by airline SMEs.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 61
======================================================================
1. Experiment 61 isolated airline policy as a deterministic input layer.
2. Policy remains separate from human behaviour and asynchronous execution.
3. Experiment 62 formalises one interchangeable policy plug-in boundary.
4. Registered policies share one frozen movement engine.
5. Registered policies share one operational and explanatory feedback schema.
6. Future airline-SME policies must enter through this boundary.
7. Amit's movement architecture remains frozen.
8. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v62 - EXPERIMENT 62 DESIGN
======================================================================

======================================================================
EXPERIMENT 62 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 62 completed the Deterministic Policy Plug-in Interface Conformance
Audit on the unchanged 30-scenario holdout.

Official result
---------------
- Registered deterministic policy plug-ins: 2.
- Scenario pairs: 30.
- Total executions: 60.
- Correct policy routes: 60/60.
- Improved/equal/worse: 3/27/0.
- Aggregate seated-passenger difference: +24.
- Complete cabins: 20 reference / 23 risk-dispersion.
- Aggregate ticks: 43,938 reference / 39,546 risk-dispersion.
- Pathways: P0=8, P1=12, P2=7, P4=3.
- deterministicPolicyPluginInterfaceReady=true.

Discussion incorporated after Experiment 62
-------------------------------------------
Experiment 62 established a stable boundary between an airline-policy input and
Amit's frozen movement architecture. An SME should eventually select a named
policy through a front-end control, menu or switch rather than editing movement
code. The policy plug-in then supplies an ordered passenger manifest to the same
movement, dependency, diagnostic and reporting engine.

The eventual SME interface should expose understandable inputs and outputs:

Inputs:
- aircraft configuration;
- named boarding policy;
- occupancy and operating conditions;
- later, a separately supplied human-behaviour profile.

Outputs:
- completion status;
- boarding duration or deterministic ticks;
- improved/equal/worse comparison;
- congestion location and residual family;
- largest dependency chain and critical blocker explanation;
- advanced evidence only when requested.

The research now distinguishes two clocks. The current synchronous deterministic
tick is the scientific clock used for replay and causal evidence. A later
asynchronous operational clock will schedule events in elapsed seconds and will
support calibration against observed wall-clock boarding time. Asynchronicity
must therefore remain a separate later layer rather than being mixed into the
policy catalogue experiments.

The policy layer, human-behaviour layer, asynchronous timing layer and front end
must remain separate responsibilities. This protects reproducibility and lets
future airline SMEs change policies without changing the movement engine.

======================================================================
RESEARCH EXPERIMENT 63 - DETERMINISTIC MULTI-POLICY CATALOGUE AUDIT
======================================================================
Build tag:

    STRESS-EXP63-DETERMINISTIC-MULTI-POLICY-CATALOGUE-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement architecture remains the single shared execution engine.
Experiment 63 changes only which deterministic admission-order plug-in supplies
the passenger queues. No policy may alter aisle movement, row-entry logic,
blocker displacement, yield-space mechanics, reservations, dependency analysis,
scheduler behaviour or stall detection.

Purpose
-------
Expand the two-policy interface from Experiment 62 into a small auditable policy
catalogue and prove that several independently selectable deterministic policies
can coexist without modifying the movement engine.

Registered policy catalogue
---------------------------
1. REFERENCE_ADMISSION_ORDER
   -> STANDARD_DETERMINISTIC_DATASET

2. TWO_ROW_RISK_DISPERSION_ADMISSION_ORDER
   -> FINAL_TWO_ROW_RISK_DISPERSION_DATASET

3. TWO_ROW_SAME_ROW_PRESERVATION_ADMISSION_ORDER
   -> FINAL_TWO_ROW_SAME_ROW_PRESERVATION_DATASET

These remain deterministic research transformations. They are not presented as
approved airline boarding procedures.

Method
------
Each of the 30 unchanged holdout scenarios is executed once through every
registered policy, producing 90 total executions.

For each scenario:
- one shared cabin, occupancy, headway and seed are generated;
- all three policies are instantiated through the common factory;
- the reference queue is compared separately with both alternatives;
- pre-movement perturbation signatures are recorded;
- all policies return the same operational and explanatory evidence schema;
- route identity and catalogue selection are audited.

Primary questions
-----------------
1. Can more than two policy identities coexist in the common interface?
2. Can each identity resolve to one unique deterministic dataset route?
3. Can every policy be selected without movement-engine edits?
4. Can both alternative policies be compared against the same reference run?
5. Can the SME selection contract remain independent of behaviour and timing?

SME selection contract
----------------------
The eventual SME front end should operate conceptually as:

    SELECT AIRCRAFT
          -> SELECT POLICY
          -> RUN THE FROZEN ENGINE
          -> COMPARE OUTCOME
          -> VIEW OPERATIONAL SUMMARY
          -> OPEN ADVANCED EVIDENCE IF REQUIRED

The SME selects a policy identity, not Java movement logic. The catalogue maps
that identity to one deterministic admission transformation. All policies then
use the same engine and return the same report fields.

Decision rule
-------------
The multi-policy catalogue is ready only if:
- exactly three registered policies are present;
- all 90 executions use the correct policy-to-dataset route;
- every registered policy is independently selectable;
- each policy has a unique dataset route;
- one common factory and feedback schema are retained;
- movement remains frozen;
- human behaviour, asynchronous timing and wall-clock calibration remain
  deferred.

Scientific boundary
-------------------
Experiment 63 is a replay and software-conformance experiment. It does not add
new independent paired evidence, validate real airline policy, model passenger
psychology or estimate elapsed boarding minutes. It deliberately prepares the
catalogue layer required before a later SME menu, behaviour plug-ins and an
asynchronous wall-clock calibration programme.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 62
======================================================================
1. Experiment 62 validated one common policy plug-in boundary.
2. Policy selection is now separate from movement implementation.
3. Experiment 63 expands the interface into a three-policy catalogue.
4. The SME will eventually select named policies through a menu or front end.
5. Human behaviour remains a separate future plug-in layer.
6. Asynchronous execution remains a separate operational-clock layer.
7. Wall-clock boarding calibration must follow asynchronous timing, not precede
   it.
8. Amit's deterministic movement and explanatory architecture remain frozen.
9. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v63 - EXPERIMENT 63 DESIGN
======================================================================


======================================================================
EXPERIMENT 63 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 63 completed the Deterministic Multi-Policy Catalogue Audit on the
unchanged 30-scenario holdout.

Official result
---------------
- Registered deterministic policies: 3.
- Holdout scenarios: 30.
- Total executions: 90.
- Correct policy-to-dataset routes: 90/90.
- Successful completions: 66/90.
- Reference complete cabins: 20/30.
- Risk-dispersion complete cabins: 23/30.
- Same-row-preservation complete cabins: 23/30.
- Both alternatives: improved/equal/worse = 3/27/0.
- Both alternatives: aggregate passenger difference = +24.
- deterministicMultiPolicyCatalogueReady=true.

Discussion incorporated after Experiment 63
-------------------------------------------
Experiment 63 confirms that Amit's deterministic movement engine has become a
platform rather than a policy-specific program. Three independently selectable
policy identities now coexist behind one factory, one movement engine and one
feedback schema.

The two alternative policies produced the same headline operational totals but
not the same explanatory pathway distributions. Risk dispersion produced
P0=8, P1=12, P2=7 and P4=3, while same-row preservation produced P0=14, P1=6,
P2=7 and P4=3. This is important scientific evidence: policies that appear equal
in aggregate can still operate through different deterministic mechanisms.

The next layer should therefore not add another policy merely for catalogue
size. It should create a unified comparison record and deterministic ranking
framework that preserves both headline outcomes and explanatory evidence. This
is the beginning of policy decision support rather than policy execution alone.

The future SME should be able to select several policies, run identical cabin
conditions, receive one common comparison table and inspect advanced evidence
when two policies appear equal. Any ranking must be explicitly evidence-first,
reproducible and labelled as a research comparison rather than an airline
recommendation.

Human behaviour, asynchronous operational timing and wall-clock calibration
remain later independent layers. They must not be introduced while the policy
comparison contract is being frozen.

======================================================================
RESEARCH EXPERIMENT 64 - UNIFIED POLICY COMPARISON AND RANKING AUDIT
======================================================================
Build tag:

    STRESS-EXP64-UNIFIED-POLICY-COMPARISON-RANKING-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's movement, dependency, scheduler and diagnostic architecture remains
unchanged. Experiment 64 adds only a comparison and ranking layer above the
three accepted deterministic policy plug-ins. No score or rank may alter a
passenger, queue, aisle, seat event, blocker, reservation or dependency state.

Purpose
-------
Transform the Experiment 63 policy catalogue into a deterministic policy
evaluation framework by producing one common comparison record per policy and
one reproducible ranking across identical holdout conditions.

Method
------
The same 30 holdout scenarios are executed through the same three policies,
producing 90 executions. Each policy receives a unified record containing:

- policy identity and dataset route;
- complete-cabin count;
- aggregate passengers seated;
- aggregate deterministic ticks;
- improved/equal/worse count relative to the reference;
- aggregate passenger difference;
- retained explanatory pathway evidence.

Ranking precedence
------------------
The ranking is deterministic and lexicographic:

1. More complete cabins.
2. More aggregate passengers seated.
3. Fewer worse holdout outcomes.
4. Fewer aggregate deterministic ticks.
5. Stable policy identity as the final tie-break.

This ordering is deliberately transparent. It is not presented as a universal
utility function or airline recommendation. A later SME or airline may assign
different operational priorities through a separately reviewed scoring layer.

Primary questions
-----------------
1. Can every registered policy produce one common comparison record?
2. Can all records be ranked without changing the movement engine?
3. Is the ranking deterministic under exact replay?
4. Are explanatory pathways retained when headline results tie?
5. Can an SME distinguish a research ranking from an operational recommendation?

SME comparison contract
-----------------------
The future front end should operate conceptually as:

    SELECT COMMON CABIN CONDITIONS
          -> SELECT POLICIES TO COMPARE
          -> RUN IDENTICAL DETERMINISTIC SCENARIOS
          -> VIEW COMMON COMPARISON TABLE
          -> VIEW EVIDENCE-FIRST RANKING
          -> OPEN EXPLANATORY PATHWAYS IF REQUIRED

The default view should present clear operational measures. Advanced diagnostics
must remain available for audit, especially where two policies have equal
completion, passenger and tick totals but different causal pathways.

Decision rule
-------------
The unified comparison and ranking layer is ready only if:

- all 90 executions produce conforming policy routes;
- all policies return the same comparison fields;
- ranking is deterministic and uses the published precedence;
- explanatory evidence remains attached to each alternative;
- movement architecture remains frozen;
- human behaviour, asynchronous timing and wall-clock calibration remain
  deferred.

Scientific boundary
-------------------
Experiment 64 does not validate an airline boarding procedure, infer passenger
psychology or predict minutes. It ranks deterministic research policies only on
the accepted holdout and published comparison precedence. The rank is evidence,
not an operational recommendation.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 63
======================================================================
1. Experiment 63 completed the deterministic multi-policy catalogue.
2. The frozen engine now supports policy selection rather than policy-specific
   execution.
3. Equal aggregate outcomes may still have different explanatory pathways.
4. Experiment 64 introduces a common comparison record and reproducible ranking.
5. Ranking remains separate from movement and is not an airline recommendation.
6. The SME layer will later expose comparison first and advanced evidence second.
7. Human behaviour remains a separate future plug-in layer.
8. Asynchronous timing remains a separate operational-clock layer.
9. Wall-clock calibration follows asynchronous timing and observed data.
10. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v64 - EXPERIMENT 64 DESIGN
======================================================================


======================================================================
EXPERIMENT 64 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 64 completed the Unified Policy Comparison and Ranking Audit on the
unchanged 30-scenario, three-policy holdout.

Official result
---------------
- Registered deterministic policies: 3.
- Holdout scenarios: 30.
- Total policy executions: 90.
- Conforming policy routes: 90/90.
- Successful completions: 66/90.
- Reference: 20/30 complete cabins, 7,916 aggregate seated, 43,938 ticks.
- Risk dispersion: 23/30 complete cabins, 7,940 aggregate seated, 39,546 ticks.
- Same-row preservation: 23/30 complete cabins, 7,940 aggregate seated,
  39,546 ticks.
- Both alternatives: improved/equal/worse = 3/27/0.
- Both alternatives: aggregate passenger difference = +24.
- Deterministic ranking:
  1. TWO_ROW_RISK_DISPERSION_ADMISSION_ORDER.
  2. TWO_ROW_SAME_ROW_PRESERVATION_ADMISSION_ORDER.
  3. REFERENCE_ADMISSION_ORDER.
- commonComparisonRecords=true.
- allPolicyRoutesConform=true.
- deterministicTieBreakApplied=true.
- noAlternativeWorse=true.
- explanatoryEvidenceRetained=true.
- movementArchitectureFrozen=true.
- unifiedPolicyComparisonAndRankingReady=true.

Discussion incorporated after Experiment 64
-------------------------------------------
Experiment 64 is an architectural validation rather than another policy-
performance experiment. The deterministic simulator has now changed from a
policy execution engine into a policy evaluation engine. Every registered
policy passes through one common contract: deterministic admission generation,
frozen movement execution, dependency evidence production and operational
reporting.

The comparison layer is reusable. A future policy can enter the accepted plug-in
boundary and automatically receive the same comparison record without movement-
engine edits. This separates the scientific execution layer from the operational
interpretation layer and marks a major transformation in the research.

The two alternative policies were operationally identical under the published
headline measures, but they retained different explanatory pathway
Distributions. Risk dispersion produced P0=8, P1=12, P2=7 and P4=3; same-row
preservation produced P0=14, P1=6, P2=7 and P4=3. The framework therefore does
not collapse equal totals into an unsupported claim of structural identity.
Operational equality does not imply causal or structural equality.

The stable policy-identity tie-break produced a reproducible total ordering, but
that ordering must not be mistaken for evidence that risk dispersion is
operationally superior to same-row preservation. Their accepted headline
measures are equal. The tie-break exists to make reports reproducible, while the
explanatory evidence remains visible for audit.

Architectural reflection - from execution to decision support
-------------------------------------------------------------
The completed architecture now contains two clearly separated layers:

Scientific layer:
- deterministic movement;
- dependency and scheduler evidence;
- critical-blocker and pathway analysis;
- policy execution through a common plug-in contract;
- unified comparison records and reproducible ranking.

Operational interpretation layer:
- explainable recommendation;
- confidence and limitation statements;
- alternative-policy summaries;
- later behaviour profiles;
- later asynchronous timing and wall-clock calibration;
- eventual SME-facing interface.

Experiment 64 therefore establishes the platform on which the remaining work
can be built. The movement engine is no longer the component under development.
The next experiments should wrap its accepted evidence in progressively more
operational forms without weakening deterministic replay.

======================================================================
RESEARCH EXPERIMENT 65 - DETERMINISTIC OPERATIONAL RECOMMENDATION AUDIT
======================================================================
Build tag:

    STRESS-EXP65-DETERMINISTIC-OPERATIONAL-RECOMMENDATION-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement algorithm remains the frozen scientific
engine. Experiment 65 consumes only the accepted policy comparison records and
explanatory evidence produced above that engine. A recommendation may describe
or rank an outcome, but it cannot alter admission queues after construction,
passenger movement, row-entry logic, blocker displacement, yield-space
mechanics, reservations, dependency propagation, scheduler decisions or stall
detection.

Purpose
-------
Transform the Experiment 64 comparison and ranking framework into a bounded,
explainable operational recommendation layer.

Experiment 64 answered which deterministic research policy ranks first under a
published precedence. Experiment 65 asks whether the simulator can explain why
that policy appears first, assign an honest confidence classification, preserve
alternative policies and state the limitations that prevent the result from
being treated as an approved airline procedure.

Method
------
The unchanged 30 holdout scenarios are replayed through the same three policies,
producing 90 total executions. The accepted common comparison records are
retained. The new recommendation layer then produces:

- one recommended research policy;
- the evidence basis for that recommendation;
- a bounded confidence classification;
- explicit detection of operational ties;
- alternative-policy summaries;
- retained explanatory pathway evidence;
- an SME and scientific-boundary statement.

Recommendation precedence
-------------------------
The recommendation consumes the Experiment 64 deterministic ranking:

1. More complete cabins.
2. More aggregate passengers seated.
3. Fewer worse holdout outcomes.
4. Fewer aggregate deterministic ticks.
5. Stable policy identity only as a reproducibility tie-break.

Confidence contract
-------------------
HIGH confidence may be assigned only when the leading policy has a strict
advantage under the published measures.

MODERATE confidence must be assigned when the leading alternatives are
operationally equal and stable identity alone resolves their order. Different
explanatory pathways must remain visible in this case.

No confidence classification is an airline deployment approval. Confidence is
bounded to deterministic replay, the accepted holdout and the published
precedence.

Expected Experiment 65 interpretation
-------------------------------------
On the current evidence, risk dispersion and same-row preservation are expected
to remain operationally equal. Risk dispersion may appear first because of the
stable policy-identity tie-break, but the recommendation confidence should be
MODERATE rather than HIGH. Same-row preservation must be reported as an
operationally equivalent alternative with a different explanatory pathway
profile.

Primary questions
-----------------
1. Can one explainable recommendation be generated from the common records?
2. Does the recommendation state its evidence basis rather than publishing an
   unexplained rank?
3. Is an operational tie detected and reflected honestly in confidence?
4. Are alternative policies retained rather than hidden?
5. Are explanatory pathway differences preserved when headline measures tie?
6. Does the recommendation remain completely separate from movement execution?
7. Are airline SME review and later calibration stated as mandatory boundaries?

SME recommendation contract
---------------------------
The future front end should operate conceptually as:

    SELECT COMMON CABIN CONDITIONS
          -> SELECT POLICIES
          -> RUN THE FROZEN ENGINE
          -> VIEW RECOMMENDED RESEARCH POLICY
          -> VIEW EVIDENCE AND CONFIDENCE
          -> VIEW OPERATIONALLY EQUIVALENT OR LOWER-RANKED ALTERNATIVES
          -> OPEN ADVANCED EXPLANATORY EVIDENCE IF REQUIRED

The default presentation should be understandable to an operational user, while
advanced deterministic evidence remains available for scientific audit.

Decision rule
-------------
The deterministic operational recommendation layer is ready only if:

- all 90 policy executions use conforming routes;
- all policies retain one common evidence record;
- one recommendation is generated;
- recommendation confidence is assigned according to the published contract;
- operational ties are detected rather than hidden;
- alternative policies are reported;
- explanatory pathway differences remain attached;
- movement and dependency architecture remains frozen;
- human behaviour, asynchronous execution and wall-clock calibration remain
  deferred;
- airline SME review is explicitly required before operational use.

Scientific boundary
-------------------
Experiment 65 creates an explainable research recommendation, not an airline
instruction. It does not model psychology, hesitation, baggage variability,
crew intervention, asynchronous events or elapsed minutes. It does not replace
an airline's objectives or domain judgement. The recommendation is valid only
for this deterministic holdout and published precedence.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 64
======================================================================
1. Experiment 64 completed the common comparison and ranking framework.
2. The simulator has transformed from policy execution into policy evaluation.
3. Equal operational totals can retain different causal pathway signatures.
4. Stable identity makes reports reproducible but does not establish superiority.
5. Experiment 65 adds explainable, bounded recommendation above the frozen core.
6. Operational ties must reduce confidence and preserve equivalent alternatives.
7. Airline SME review remains mandatory before operational use.
8. Human behaviour remains a separate future plug-in layer.
9. Asynchronous timing remains a separate operational-clock layer.
10. Wall-clock calibration follows asynchronous timing and observed data.
11. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v65 - EXPERIMENT 65 DESIGN
======================================================================


======================================================================
EXPERIMENT 65 OFFICIAL FINDINGS AND DISCUSSION
======================================================================
Experiment 65 completed the Deterministic Operational Recommendation Audit on
exactly the same 30-scenario holdout and three accepted deterministic policy
routes.

Official result
---------------
- Registered policies: 3.
- Policy executions: 90/90.
- Conforming policy routes: 90/90.
- Successful completions: 66/90.
- Recommended research policy:
  TWO_ROW_RISK_DISPERSION_ADMISSION_ORDER.
- Recommendation confidence: MODERATE.
- Operational tie detected: true.
- Operationally equivalent alternative retained:
  TWO_ROW_SAME_ROW_PRESERVATION_ADMISSION_ORDER.
- Risk-dispersion pathway distribution:
  P0=8, P1=12, P2=7, P4=3.
- Same-row-preservation pathway distribution:
  P0=14, P1=6, P2=7, P4=3.
- recommendationGenerated=true.
- recommendationConfidenceAssigned=true.
- alternativePoliciesReported=true.
- commonEvidenceContractSatisfied=true.
- explanatoryPathwayDifferenceRetained=true.
- movementArchitectureFrozen=true.
- dependencyEvidencePreserved=true.
- operationalRecommendationReady=true.

Interpretation
--------------
Experiment 65 produced exactly the conservative result required by the
scientific contract. The recommendation layer did not manufacture HIGH
confidence merely because a total ordering existed. Risk dispersion and
same-row preservation were equal on complete cabins, aggregate passengers,
worse cases and deterministic ticks. Stable policy identity resolved report
order only; it did not establish operational superiority.

The MODERATE classification is therefore a positive validation result. It shows
that ranking, recommendation and confidence are separate concepts. The framework
can recommend one research policy for reproducible presentation while retaining
an operationally equivalent alternative and exposing the reason confidence is
bounded.

The experiment also confirms that recommendation consumes evidence rather than
generating or altering it. Movement, dependency propagation, scheduler state,
region management and explanatory pathway assignment remain below the
recommendation layer and unchanged.

Discussion incorporated after Experiment 65
-------------------------------------------
The architecture now supports a further long-term evolution: cumulative or
composable boarding policies. A real airline may not use one isolated policy.
An SME may require accessibility priority, family or group preservation,
congestion dispersion, seat-type preference and other operational rules to
operate together.

That possibility does not permit policies to overwrite one another silently.
Every cumulative approach requires:

- explicit policy modules;
- declared mutation scopes;
- published precedence;
- deterministic conflict resolution;
- evidence provenance identifying which stage affected the final order;
- retention of deferred or overridden objectives in the audit record;
- airline SME review before operational testing.

Precedence is especially important where objectives conflict. For example,
family preservation may be defined above congestion dispersion, so a dispersion
rule cannot split a protected group merely to improve aisle distribution. The
resolution must be configured before movement, reproduced on replay and stated
in the evidence.

The current code still applies one admission-order policy per simulation run.
This is a constraint, not a defect. It protects the accepted evidence while a
safe composition contract is designed. The next experiment should therefore
validate the cumulative-policy scaffold and precedence rules without falsely
claiming that a cumulative operational policy has already been executed.


======================================================================
RESEARCH EXPERIMENT 66 - CUMULATIVE POLICY COMPOSITION AND PRECEDENCE AUDIT
======================================================================
Build tag:

    STRESS-EXP66-CUMULATIVE-POLICY-COMPOSITION-PRECEDENCE-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement engine remains frozen. Experiment 66 does
not alter passenger movement or combine queue transformations inside the engine.
It introduces a composition contract above policy generation so that future SME-
defined modules can be ordered, audited and resolved before one final admission
order is passed to the accepted movement architecture.

Purpose
-------
Validate the deterministic scaffold required before several boarding-policy
objectives may be applied cumulatively.

The experiment must make both the present constraint and future potential
explicit:

    CURRENT CONSTRAINT:
    one admission-order policy executes per movement run.

    EVOLUTION POTENTIAL:
    cumulative policy stacks may be introduced through published precedence,
    deterministic conflict resolution, provenance and airline SME review.

Method
------
The accepted three-policy catalogue and 30-scenario holdout remain unchanged,
producing the same 90 policy executions and Experiment 65 recommendation audit.
A separate non-invasive composition audit then registers five conceptual
modules:

1. Accessibility priority placeholder.
2. Family or group preservation placeholder.
3. Two-row same-row preservation.
4. Two-row risk dispersion.
5. Reference admission order.

Each module declares:

- module identity;
- mutation scope;
- published precedence;
- operational objective;
- active, advisory or deferred status.

The two SME placeholders are deliberately non-executable. They demonstrate the
future interface without pretending that accessibility or family behaviour has
already been modelled.

Published precedence
--------------------
The initial research ordering is:

1. Accessibility priority placeholder - precedence 500.
2. Family/group preservation placeholder - precedence 400.
3. Same-row preservation - precedence 300.
4. Risk dispersion - precedence 200.
5. Reference admission order - precedence 100.

This ordering is a scaffold for audit, not an airline-approved hierarchy. An
airline SME must review and replace it before operational testing.

Conflict contract
-----------------
A conflict exists when two modules claim the same mutation scope. Under the
present architecture, same-row preservation, risk dispersion and reference order
all control ADMISSION_ORDER. They therefore cannot be applied cumulatively by
silent sequential overwrite.

Experiment 66 must:

- detect the overlap;
- assign the highest-precedence module as active controller for that scope;
- retain lower-precedence modules as deferred by precedence;
- state the resolution reason;
- preserve every module in the provenance record;
- report that no cumulative operational policy was executed.

Future implementation contract
------------------------------
A genuinely cumulative policy engine should later transform complete admission
orders through composable, stage-specific rules rather than selecting one whole-
order generator. Every stage must record:

- incoming order identity;
- applied constraint or preference;
- passengers moved or protected;
- conflicts encountered;
- precedence decision;
- outgoing order identity;
- deterministic replay signature.

Primary questions
-----------------
1. Can proposed policy modules declare their mutation scope and precedence?
2. Can overlapping scopes be detected before movement?
3. Can every conflict be resolved through published deterministic precedence?
4. Are deferred modules retained visibly rather than discarded?
5. Are non-implemented SME modules labelled explicitly as placeholders?
6. Does the audit preserve the current one-policy-per-run constraint honestly?
7. Can cumulative potential be documented without changing the frozen engine?

Decision rule
-------------
The cumulative composition scaffold is ready only if:

- all proposed modules are registered;
- precedence order is deterministic;
- overlapping mutation scopes are detected;
- every detected conflict receives one published resolution;
- no module is silently overwritten;
- SME placeholders remain explicitly non-executable;
- evidence provenance is mandatory;
- movement architecture remains frozen;
- cumulativeOperationalPolicyExecuted=false.

Scientific boundary
-------------------
Experiment 66 does not validate a combined airline boarding policy. It does not
model accessibility needs, family membership, passenger psychology, asynchronous
movement or wall-clock minutes. It validates the contract needed to add such
SME-defined policy objectives later without corrupting deterministic replay.

Evidence accounting
-------------------
Independent paired scenarios remain 300.
Aggregate passenger gain remains +504.
Improved/worse/equal remains 22/1/277.
The cumulative composition audit introduces no new operational performance
claim.

======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 65
======================================================================
1. Experiment 65 completed the bounded operational recommendation layer.
2. MODERATE confidence correctly reflects an operational tie.
3. Recommendation now preserves equivalent alternatives and causal differences.
4. The framework may evolve from individual policies to SME-defined policy stacks.
5. Cumulative policies require declared scopes and published precedence.
6. Shared mutation scopes cannot silently overwrite one another.
7. Experiment 66 validates the composition scaffold, not a combined policy.
8. Accessibility and family/group modules remain explicit future placeholders.
9. The frozen movement engine receives only one final deterministic order.
10. Human behaviour, asynchronous timing and wall-clock calibration remain later
    independent layers.
11. Independent cumulative evidence remains 300 paired scenarios and +504.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v66 - EXPERIMENT 66 DESIGN
======================================================================


======================================================================
EXPERIMENT 66 - ADDITIONAL ARCHITECTURAL DISCUSSION
======================================================================
Operational configuration philosophy
-------------------------------------
Experiment 66 established that future policy composition must not be hidden
inside the deterministic movement engine. The same separation principle applies
to operational configuration and temporal modelling.

A runtime menu was considered but deliberately deferred. Menu design is a user-
interface concern, whereas the present programme remains an architectural and
evidential research programme. The more durable interface is therefore a single
clustered configuration contract that can later be populated by a console menu,
desktop application, web interface or airline integration without rewriting the
movement architecture.

All supported runtime values should publish:

- parameter identity;
- valid range or supported values;
- research default;
- current execution status;
- whether the value is operational or reserved for future research.

Temporal modelling direction
----------------------------
The project retains asynchronous ticks as one of its final research layers. This
work was intentionally deferred until deterministic movement, evidence, policy
comparison and recommendation had stabilised.

The asynchronous scheduler should consume validated deterministic movement
events rather than replace or modify the movement engine. Initial movement
metrics may be arbitrary but published and reproducible. For example, forward
aisle movement may use a shorter duration than entering or leaving a seat row.
Such numbers are operational timing assumptions, not measured human kinetics.

A later evolution may permit airline or human-factors SMEs to attach configurable
human behaviour profiles to passengers. Those profiles could alter event timing
while preserving deterministic route, dependency and safety rules. Accessibility,
family requirements and mobility characteristics must not be claimed until they
are defined and reviewed by appropriate SMEs.

Closing roadmap
---------------
The agreed closing sequence is:

- Experiment 67: runtime operational configuration consolidation;
- Experiment 68: asynchronous tick scheduler;
- Experiment 69: concurrent tick and dependency validation;
- Experiment 70: wall-clock calibration and configurable timing assumptions;
- final publication figure: a large contemporary systems diagram connecting
  Experiments 18-70 and the complete architecture.


======================================================================
RESEARCH EXPERIMENT 67 - RUNTIME OPERATIONAL CONFIGURATION CONSOLIDATION
======================================================================
Build tag:

    STRESS-EXP67-RUNTIME-OPERATIONAL-CONFIGURATION-CONSOLIDATION-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement engine remains frozen. Experiment 67 moves
supported operational values into one documented configuration contract above
the engine. Configuration may influence reporting and later timing layers, but
it must not silently modify passenger movement, dependency resolution or the
accepted evidence pathway.

Purpose
-------
Consolidate the supported SME/runtime parameters into one clearly documented
section and verify that configuration is published independently from execution.

Method
------
The accepted three-policy catalogue, 30-scenario holdout, recommendation audit
and Experiment 66 composition audit are retained unchanged. Experiment 67 adds a
single code section containing:

- BOARDING_THRESHOLD_PERCENT;
- POLICY_CATALOGUE_ENABLED;
- RECOMMENDATION_ENGINE_ENABLED;
- PRECEDENCE_PROFILE;
- CUMULATIVE_POLICY_COMPOSITION_ENABLED;
- FUTURE_TIMING_PROFILE;
- FUTURE_HUMAN_BEHAVIOUR_PROFILE.

Each supported parameter documents its valid range or accepted values and its
research default. Reserved future profiles are explicitly non-executable.

Boarding threshold contract
---------------------------
The research default is 85 percent, with a valid configurable range of 1-100.
This value is an operational reporting or decision threshold. It does not replace
the strict completed-cabin definition, which remains every passenger seated.

The separation prevents a configurable threshold from rewriting historical
completion evidence.

Startup and evidence publication
--------------------------------
The executing program prints the complete runtime configuration at startup. A
separate Experiment 67 audit prints the same values, supported ranges, defaults,
reserved status and architectural boundaries into the captured evidence stream.

Future timing preparation
-------------------------
FUTURE_TIMING_PROFILE and FUTURE_HUMAN_BEHAVIOUR_PROFILE are reserved only.
Experiment 67 does not execute asynchronous timing, wall-clock calibration or
human behaviour profiles.

Future experiments may attach published event durations to deterministic
movement events. A later SME-defined profile layer may vary durations between
passengers without changing deterministic movement order or dependency rules.

Primary questions
-----------------
1. Are all supported runtime parameters clustered in one location?
2. Are ranges, supported values and research defaults published?
3. Is the 85 percent threshold configurable without changing strict completion?
4. Is configuration printed at startup and captured in evidence?
5. Are timing and human-behaviour profiles clearly reserved and non-executable?
6. Does the frozen movement and recommendation architecture remain unchanged?
7. Can future interfaces consume this configuration contract without engine
   modification?

Decision rule
-------------
Runtime operational configuration is ready only if:

- runtimeConfigurationClustered=true;
- runtimeParametersPublished=true;
- validInputRangesDocumented=true;
- researchDefaultsPublished=true;
- boardingThresholdSMEConfigurable=true;
- supportedPrecedenceProfile=true;
- configurationPrintedAtStartup=true;
- configurationCapturedInEvidence=true;
- futureTimingProfileReserved=true;
- futureBehaviourProfileReserved=true;
- movementArchitectureFrozen=true;
- recommendationFrameworkPreserved=true;
- asynchronousTimingExecuted=false;
- wallClockCalibrationExecuted=false;
- humanBehaviourProfileExecuted=false;
- runtimeOperationalConfigurationReady=true.

Scientific boundary
-------------------
Experiment 67 is not an operational human-behaviour model. It does not validate
minutes, hours, passenger kinetics, accessibility assumptions or airline policy.
It validates configuration visibility and separation only.

Evidence accounting
-------------------
The independent operational evidence remains unchanged. The configuration audit
creates no additional passenger-performance claim and does not alter the accepted
holdout.


======================================================================
RUNNING CONCLUSIONS AFTER EXPERIMENT 67 DESIGN
======================================================================
1. Cumulative policy composition remains scaffolded, not operationally executed.
2. Runtime menus remain deferred as a separate interface concern.
3. Supported operational parameters are clustered in one configuration contract.
4. The 85 percent threshold is configurable but strict completion remains 100
   percent seated.
5. Configuration is printed at startup and repeated in auditable evidence.
6. Timing and human-behaviour profiles remain explicitly reserved.
7. Asynchronous ticks remain scheduled for Experiment 68.
8. Concurrent movement and dependency validation remain scheduled for Experiment
   69.
9. Wall-clock calibration remains scheduled for Experiment 70.
10. The programme will conclude with a contemporary full-page systems diagram
    connecting Experiments 18-70.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v67 - EXPERIMENT 67 DESIGN
======================================================================


======================================================================
EXPERIMENT 67 - FINDINGS AND DISCUSSION UPDATE
======================================================================
Execution evidence reviewed
---------------------------
The Experiment 67 verification output and captured runtime evidence were reviewed.
The code compiled and completed normally with 90 scenario executions: 66 complete
cabins and 24 incomplete/MAX_TICKS outcomes. This distribution is retained as
movement evidence rather than interpreted as a configuration-layer performance
change.

Findings
--------
The clustered runtime configuration contract passed every intended architectural
check. BOARDING_THRESHOLD_PERCENT remained configurable within 1-100 while strict
completion continued to require every passenger seated. Configuration was printed
at startup and repeated in the evidence audit. The policy catalogue,
recommendation engine and STANDARD precedence profile remained published and
stable.

The reserved boundaries also behaved correctly:

- futureTimingProfileReserved=true;
- futureBehaviourProfileReserved=true;
- asynchronousTimingExecuted=false;
- wallClockCalibrationExecuted=false;
- humanBehaviourProfileExecuted=false;
- movementArchitectureFrozen=true;
- recommendationFrameworkPreserved=true;
- runtimeOperationalConfigurationReady=true.

Interpretation
--------------
Experiment 67 established configuration as a genuine architectural boundary. The
configuration layer can now be populated by future console, desktop, web or
enterprise interfaces without rewriting Amit's deterministic movement engine.
The absence of asynchronous execution in this run is positive baseline evidence:
Experiment 68 can activate timing against a known, configuration-only predecessor.

Discussion incorporated from architectural review
--------------------------------------------------
Configurable variables should remain clustered wherever practical and should
carry inline comments stating their valid ranges or supported values and research
defaults. This requirement is carried directly into Experiment 68. The final
programme roadmap remains Experiments 68-70 followed by one large contemporary
systems diagram connecting Experiments 18-70. The final figure need not repeat a
cabin drawing; its role is to explain the complete research architecture.

Milestone
---------
Experiment 67 is the accepted runtime-configuration baseline. It is not a
boarding-performance milestone and creates no new airline-operational claim.


======================================================================
RESEARCH EXPERIMENT 68 - DETERMINISTIC ASYNCHRONOUS TICK SCHEDULER
======================================================================
Build tag:

    STRESS-EXP68-DETERMINISTIC-ASYNCHRONOUS-TICK-SCHEDULER-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic people-movement engine remains frozen. Experiment 68 adds a
separate timing consumer above the completed movement evidence. It assigns
published operational tick costs to observed movement categories without changing
passenger admission order, movement decisions, dependency resolution, safety
rules or accepted recommendation evidence.

Purpose
-------
Activate the first temporal layer while retaining deterministic replay. The
experiment asks whether movement evidence can be translated into reproducible
operational ticks before any concurrent movement or wall-clock calibration is
introduced.

Clustered configurable timing contract
---------------------------------------
All routine timing variables are placed together in the SME/runtime configuration
area. Inline code comments publish supported values and defaults:

- OPERATIONAL_TIMING_PROFILE: RESEARCH_BASELINE;
- TIMING_FORWARD_AISLE_MOVE_TICKS: 1-100, default 1;
- TIMING_ENTER_ROW_TICKS: 1-100, default 2;
- TIMING_EXIT_ROW_TICKS: 1-100, default 2;
- TIMING_CROSS_SEAT_POSITION_TICKS: 1-100, default 1;
- TIMING_RETURN_TO_SEAT_TICKS: 1-100, default 1;
- TIMING_WAIT_TICKS: 1-100, default 1;
- ASYNCHRONOUS_TIMING_ENABLED: true/false, default true.

The deliberately slower row-entry and row-exit assumptions reflect the agreed
initial research model. They are arbitrary, published and reproducible timing
assumptions, not measured passenger kinetics.

Method
------
The accepted 30-scenario, three-policy holdout is replayed unchanged. After each
movement run, the timing layer consumes existing deterministic evidence:

- observed left/right aisle moves;
- observed left/right seat events;
- observed blocked intervals.

It calculates a deterministic scheduled-tick total using the clustered timing
contract. The calculation does not delay or reorder movement in Experiment 68;
it produces an independent temporal evidence stream. This conservative design
isolates the timing model before Experiment 69 tests safe concurrency and
inter-event dependencies.

Primary questions
-----------------
1. Are all timing variables clustered and documented with supported values?
2. Does the timing profile assign deterministic costs to observed events?
3. Is timing evidence produced for the complete holdout?
4. Are passenger order and dependency rules unchanged?
5. Is concurrent movement still absent?
6. Are wall-clock calibration and human behaviour still deferred?
7. Does the recommendation framework remain preserved?

Decision rule
-------------
Experiment 68 is ready only if:

- timingVariablesClustered=true;
- timingProfilePublished=true;
- timingValuesSupported=true;
- deterministicTickSchedulerEnabled=true;
- timingEvidenceGenerated=true;
- movementOrderingUnchanged=true;
- dependencyRulesUnchanged=true;
- eventDurationsDeterministic=true;
- wallClockCalibrationExecuted=false;
- concurrentMovementExecuted=false;
- humanBehaviourProfileExecuted=false;
- movementArchitectureFrozen=true;
- recommendationFrameworkPreserved=true;
- deterministicAsynchronousTimingReady=true.

Scientific boundary
-------------------
An operational tick is not a second, minute or hour. Experiment 68 does not claim
real boarding duration, human reaction time, mobility behaviour, baggage handling
or airline calibration. It does not execute simultaneous movements. Those issues
remain separated for Experiments 69 and 70 and for later SME-defined profiles.

Next design position
--------------------
Experiment 69 will use the deterministic tick evidence to test which independent
movement events may occupy the same scheduling interval without violating aisle,
row, seat-event or dependency constraints.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v68 - EXPERIMENT 68 DESIGN
======================================================================

======================================================================
EXPERIMENT 68 - FINDINGS AND DISCUSSION UPDATE
======================================================================
Execution evidence reviewed
---------------------------
The Experiment 68 verification output and timing results were reviewed. The code
compiled and completed normally across 90 scenario executions: 66 successful
completions and 24 incomplete/MAX_TICKS outcomes. The deterministic timing audit
reported 368,433 aisle moves, 23,796 seat events, 817,431 blocked intervals and
1,328,640 scheduled operational ticks. The average completed-scenario total was
12,668.73 operational ticks.

Findings
--------
Experiment 68 successfully established reproducible tick metrics while preserving
movement order and dependency rules. All timing variables remained clustered and
published with supported values and defaults. Operational ticks remained explicit
research assumptions rather than seconds or minutes.

Discussion incorporated after result review
--------------------------------------------
The aggregate metrics proved that deterministic movement evidence could be
translated into reproducible timing totals, but they did not provide a human-
readable replay of what each passenger actually did. An airline engineer or
reviewer must be able to trace statements such as:

- Passenger P23 reached the assigned row for seat 40A;
- a seated blocker stood and moved to an aisle yield tile;
- the target passenger crossed the row and occupied the assigned seat;
- the blocker returned to the original seat after the dependency cleared.

This distinction is important: Experiment 68 validated tick metrics; Experiment
69 must publish the underlying event history. The event history is therefore not
retroactively claimed as an Experiment 68 result. It becomes the central evidence
contribution of Experiment 69.

Milestone
---------
Experiment 68 is accepted as the deterministic operational-tick baseline. It is
not yet an operational replay system and makes no wall-clock claim.

======================================================================
RESEARCH EXPERIMENT 69 - CONCURRENT TICK VALIDATION AND OPERATIONAL EVENT LOG
======================================================================
Build tag:

    STRESS-EXP69-CONCURRENT-TICK-VALIDATION-EVENT-LOG-AUDIT-001

Connection to Amit Amlani's Original People-Movement Algorithm
--------------------------------------------------------------
Amit's deterministic movement engine remains frozen. Experiment 69 instruments
actual state transitions already produced by that engine. It does not invent,
reorder or optimise movement. The event log is therefore an auditable publication
of Amit's deterministic decisions rather than a replacement movement model.

Purpose
-------
Introduce a passenger-level operational event log and identify events that share
the same deterministic simulation tick. This creates the evidence required to
validate safe concurrency without yet executing a parallel movement engine.

Clustered configurable event-log contract
-----------------------------------------
All routine event-log variables are clustered beside the timing configuration.
Inline comments publish supported values and defaults:

- OPERATIONAL_EVENT_LOG_ENABLED: true/false, default true;
- OPERATIONAL_EVENT_LOG_SCOPE: TARGET_SCENARIO_ONLY or ALL_SCENARIOS,
  default TARGET_SCENARIO_ONLY;
- OPERATIONAL_EVENT_LOG_SCENARIO: 1-NUMBER_OF_SCENARIOS, default 14;
- OPERATIONAL_EVENT_LOG_DATASET: one of the three published deterministic
  policy datasets, default STANDARD_DETERMINISTIC_DATASET;
- OPERATIONAL_EVENT_LOG_MAX_EVENTS: 100-1,000,000, default 50,000;
- LOG_REPEATED_WAIT_EVENTS: true/false, default false.

Method
------
The accepted 30-scenario, three-policy holdout is replayed unchanged. For the
configured scenario and dataset, the program writes a separate operational event
log. Each record includes:

- deterministic event sequence;
- simulation tick;
- scenario and dataset identity;
- passenger identity and assigned seat;
- serving aisle;
- action;
- origin and destination;
- operational tick cost;
- dependency or release condition;
- plain-language explanation.

Instrumented event types include aircraft entry, forward aisle movement, seat-
event start, blocker stand/yield movement, target seating, blocker reseating and
seat-event completion.

Concurrent-tick validation boundary
-----------------------------------
Experiment 69 groups events that occurred during the same deterministic simulation
tick. This demonstrates that multiple independent events can coexist in one tick
and creates a candidate set for dependency validation. It does not yet execute
threads or reorder events. The flag concurrentMovementExecuted therefore remains
false.

Primary questions
-----------------
1. Is the event-log configuration clustered and fully documented?
2. Does each record identify the passenger and assigned seat?
3. Are blocker stand/yield and reseating events explicitly published?
4. Are event dependencies and operational costs visible?
5. Are same-tick event groups observed without changing deterministic order?
6. Is wall-clock calibration still deferred to Experiment 70?

Decision rule
-------------
Experiment 69 is ready only if:

- eventLogConfigurationClustered=true;
- supportedValuesPublished=true;
- passengerIdentityPublished=true;
- seatDestinationPublished=true;
- blockerYieldEventsPublished=true;
- deterministicEventOrderPreserved=true;
- sameTickGroupingObserved=true;
- concurrentMovementExecuted=false;
- wallClockCalibrationExecuted=false;
- operationalEventLogReady=true.

Scientific boundary
-------------------
The event log is a deterministic replay record. Operational tick costs are not
seconds or minutes. Same-tick grouping is not proof that every grouped event may
be executed by separate threads. Physical aisle, row, seat and dependency
conflicts still require validation before any true concurrent execution claim.

Next design position
--------------------
Experiment 70 will calibrate the validated event/tick evidence to published
wall-clock assumptions and configurable human-behaviour profiles. The frozen
movement engine and Experiment 69 event history will remain unchanged.

======================================================================
END OF DETERMINISTIC STRESS DATASET RESEARCH LOG v69 - EXPERIMENT 69 DESIGN
======================================================================
