
================ AUTOMATIC SHUFFLE ROBUSTNESS | SESSION 20260916_184907 ================
ENGINE VERSION: integrated-post-session-v3 | paired_replays_per_method 5000 | seed_base 202609050001
Captured-card reservoir: 59 cards
HNL_SAME_STATE_CORPUS | blackjack_hnl_same_state_corpus_20260916_184907.tsv | states 550 | A/B/C1/C2/C3/C4/D/E | Frozen + Casual individual paths | pair IDs provenance only
Comparable-action definition: same replay hand, branch, decision ordinal, player-rank state and dealer upcard in both independently consumed paths.
A | Observed-card permutation: shuffle only the captured session-card reservoir; no cards added.
  Frozen median 100.00 | Casual median 100.00 | Frozen mean 98.70 | Casual mean 98.30
  £200 reach Frozen 7.32% | Casual 3.66% | Mean exposure Frozen 169.59 | Casual 155.67 | Mean maxDD Frozen 53.11 | Casual 48.28
  30-hand completion Frozen 0 | Casual 0 | Both 0
  Mean paired final-bankroll difference Frozen-Casual (both completed): N/A
  Comparable action decisions total: 60215 | Different actions total: 3835 | Mean Comparable Action Differences: 0.7670
B | Full six-deck RNG: generate and uniformly shuffle a fresh 312-card six-deck shoe.
  Frozen median 80.00 | Casual median 92.50 | Frozen mean 99.52 | Casual mean 96.91
  £200 reach Frozen 30.12% | Casual 20.12% | Mean exposure Frozen 450.70 | Casual 413.83 | Mean maxDD Frozen 92.90 | Casual 83.79
  30-hand completion Frozen 3348 | Casual 3519 | Both 3150
  Mean paired final-bankroll difference Frozen-Casual (both completed): 7.58
  Comparable action decisions total: 137976 | Different actions total: 5740 | Mean Comparable Action Differences: 1.1480
C1 | Penetration 50%: captured reservoir; reshuffle only remaining cards between hands at 50%.
  Frozen median 100.00 | Casual median 100.00 | Frozen mean 99.16 | Casual mean 98.56
  £200 reach Frozen 7.30% | Casual 3.46% | Mean exposure Frozen 169.51 | Casual 155.32 | Mean maxDD Frozen 52.88 | Casual 48.01
  30-hand completion Frozen 0 | Casual 0 | Both 0
  Mean paired final-bankroll difference Frozen-Casual (both completed): N/A
  Comparable action decisions total: 60290 | Different actions total: 3825 | Mean Comparable Action Differences: 0.7650
C2 | Penetration 65%: captured reservoir; reshuffle only remaining cards between hands at 65%.
  Frozen median 100.00 | Casual median 100.00 | Frozen mean 98.37 | Casual mean 97.66
  £200 reach Frozen 6.68% | Casual 2.42% | Mean exposure Frozen 169.56 | Casual 155.28 | Mean maxDD Frozen 53.48 | Casual 48.32
  30-hand completion Frozen 0 | Casual 0 | Both 0
  Mean paired final-bankroll difference Frozen-Casual (both completed): N/A
  Comparable action decisions total: 60393 | Different actions total: 3717 | Mean Comparable Action Differences: 0.7434
C3 | Penetration 75%: captured reservoir; reshuffle only remaining cards between hands at 75%.
  Frozen median 100.00 | Casual median 100.00 | Frozen mean 98.47 | Casual mean 97.27
  £200 reach Frozen 6.72% | Casual 2.66% | Mean exposure Frozen 169.63 | Casual 155.02 | Mean maxDD Frozen 53.07 | Casual 48.24
  30-hand completion Frozen 0 | Casual 0 | Both 0
  Mean paired final-bankroll difference Frozen-Casual (both completed): N/A
  Comparable action decisions total: 61279 | Different actions total: 3847 | Mean Comparable Action Differences: 0.7694
C4 | Penetration 80%: captured reservoir; reshuffle only remaining cards between hands at 80%.
  Frozen median 97.50 | Casual median 100.00 | Frozen mean 97.98 | Casual mean 98.72
  £200 reach Frozen 6.44% | Casual 2.98% | Mean exposure Frozen 169.76 | Casual 155.66 | Mean maxDD Frozen 53.44 | Casual 48.05
  30-hand completion Frozen 0 | Casual 0 | Both 0
  Mean paired final-bankroll difference Frozen-Casual (both completed): N/A
  Comparable action decisions total: 60936 | Different actions total: 3816 | Mean Comparable Action Differences: 0.7632
D | Physical-style proxy: captured reservoir with deterministic riffle/strip/cut-style mixing.
  Frozen median 100.00 | Casual median 100.00 | Frozen mean 99.87 | Casual mean 98.87
  £200 reach Frozen 7.08% | Casual 3.24% | Mean exposure Frozen 170.07 | Casual 155.76 | Mean maxDD Frozen 52.63 | Casual 47.77
  30-hand completion Frozen 0 | Casual 0 | Both 0
  Mean paired final-bankroll difference Frozen-Casual (both completed): N/A
  Comparable action decisions total: 60476 | Different actions total: 3907 | Mean Comparable Action Differences: 0.7814
E | Automatic/continuous proxy: captured reservoir; after each complete hand, used cards return and the full reservoir is remixed.
  Frozen median 77.50 | Casual median 92.50 | Frozen mean 98.39 | Casual mean 96.24
  £200 reach Frozen 30.40% | Casual 20.26% | Mean exposure Frozen 453.32 | Casual 411.41 | Mean maxDD Frozen 95.23 | Casual 84.50
  30-hand completion Frozen 3238 | Casual 3483 | Both 3057
  Mean paired final-bankroll difference Frozen-Casual (both completed): 9.65
  Comparable action decisions total: 168083 | Different actions total: 10458 | Mean Comparable Action Differences: 2.0916
Method E boundary: after each COMPLETE hand, consumed cards return to the captured reservoir and the entire reservoir is remixed; no captured card is added, removed or substituted.
Method D boundary: declared mathematical riffle/strip/cut proxy; not a proprietary casino-shuffler reproduction.

================ HYBRID FORMAL-START COUNTERFACTUAL | SESSION 20260916_184907 ================
PURPOSE: replay the unchanged Frozen and fixed Casual policies from the same formal-session starting captured cardstream. The observed Hybrid journey remains separate and authoritative for what was actually played.
Captured formal chronology available: 59 blackjack-replayable cards (rank known; any U suit remains unresolved). Each policy consumes independently; no card rank or settlement is invented.
Frozen from formal start: outcome N/A | hands 10 | W/L/P 3/6/1 | exposure £180.00 | cards consumed 59 | termination SOURCE_EXHAUSTED | NOT SOURCE-COMPLETE | last source-supported bankroll £25.00
Frozen synthetic-reference profile | RISK 47.316 | JOURNEY 33.557 | Survival 48.095 | Path Exposure 24.864 | Recovery Activity 27.711 | inputs: maxDD £75.00 | giveback £75.00 | cumulative wager £180.00 | absolute movement £165.00 | recovery episodes 1 | successful 0 | recovery ratio 0.000000 | table_exit 0
Casual from formal start: outcome N/A | hands 10 | W/L/P 3/6/1 | exposure £165.00 | cards consumed 59 | termination SOURCE_EXHAUSTED | NOT SOURCE-COMPLETE | last source-supported bankroll £47.50
Casual synthetic-reference profile | RISK 35.233 | JOURNEY 33.156 | Survival 48.095 | Path Exposure 23.662 | Recovery Activity 27.711 | inputs: maxDD £52.50 | giveback £52.50 | cumulative wager £165.00 | absolute movement £157.50 | recovery episodes 1 | successful 0 | recovery ratio 0.000000 | table_exit 0
Publication rule: if a replay is NOT SOURCE-COMPLETE, its outcome is N/A; the displayed bankroll is only the last source-supported bankroll. Risk/Journey remain synthetic-reference profile outputs for the source-supported path, not completed-outcome claims.
================ END HYBRID FORMAL-START COUNTERFACTUAL | SESSION 20260916_184907 ================

================ STAT-WATCHING CASUAL | SESSION 20260916_184907 ================
EXPERIMENT STATUS: EXPLORATORY COUNTERFACTUAL / DETERMINISTIC / QUARANTINED FROM FROZEN-vs-CASUAL SCORECARD
REPLAY BASIS: FORMAL_START
VISIBLE-EVIDENCE RULE: wager and exit decisions use only completed prior-hand evidence visible to the player; no future cards, shoe composition, or unrevealed dealer hole card is consulted. A hole-card hunch may affect a current play decision only after the dealer up-card is visible and only from previously revealed hole-card history.
POLICY: ordinary wager £15; rare evidence-triggered 2x/3x/4x presses; rolling last-five-hand table impression; deterministic walk-away rule from Hand 8 onward. Existing Casual and Frozen policies are unchanged.
STAY POLICY: the stay-at-table counterfactual keeps the SAME Stat-Watching Casual personality, wager logic, hunch logic and play policy; only the behavioural walk-away trigger is disabled. It never switches to Frozen style.
Chosen behavioural path: hands 9 | bankroll £47.50 | W/L/P 3/5/1 | exposure £165.00 | termination BEHAVIOURAL_EXIT | chosen exit Hand 9 at £47.50
Stay-at-table counterfactual: hands 10 | bankroll £32.50 | W/L/P 3/6/1 | exposure £180.00 | termination SOURCE_EXHAUSTED | NOT SOURCE-COMPLETE
Exit value versus staying: N/A unless a behavioural exit occurs and the stay-at-table replay is source-complete.
PATH_CHOSEN: H0=100.00;H1=85.00;H2=55.00;H3=70.00;H4=85.00;H5=55.00;H6=77.50;H7=77.50;H8=62.50;H9=47.50
PATH_STAY: H0=100.00;H1=85.00;H2=55.00;H3=70.00;H4=85.00;H5=55.00;H6=77.50;H7=77.50;H8=62.50;H9=47.50;H10=32.50
INFLUENCE | H5 | WAGER | 2x wager (£30.00) • last 4 hands: low cards 50%, 10-value cards 15%, W/L 2/2
INFLUENCE | H9 | EXIT | walked away at £47.50 • last 5 hands: low cards 37%, 10-value cards 41%, W/L 1/3 • drawdown from peak £52.50
Publication boundary: exploratory side experiment only. Do not award established Frozen/Casual scorecard percentiles or alter the main comparison from this block without a later declared research phase.
================ END STAT-WATCHING CASUAL | SESSION 20260916_184907 ================

================ ULTIMATE AI | SESSION 20260916_184907 ================
AI POLICY: ULTIMATE-AI-V1-FROZEN-WEIGHTED-GATE | deterministic | Frozen baseline | repository through Session 11
REPLAY BASIS: FORMAL_START
TEMPORAL BOUNDARY: only repository evidence through Session 11 plus evidence visible before each current decision was available. Session 12 completed outcomes were not in the repository during this replay.
V1 GATE: Frozen wins weak evidence and ties. Action/wager overrides require convergent predeclared support; scorecard/percentile similarity has zero live-decision weight until a predecision-safe mapping exists.
Ultimate AI: hands 10 | bankroll £25.00 | P/L £-75.00 | W/L/P 3/6/1 | exposure £180.00 | termination AI_EXIT | SOURCE-COMPLETE
Frozen same-basis comparator: hands 10 | bankroll N/A | P/L N/A | exposure £180.00 | termination SOURCE_EXHAUSTED | NOT SOURCE-COMPLETE
OVERRIDES: action 0 | wager 0 | exit 1
PATH_AI: H0=100.00;H1=70.00;H2=40.00;H3=25.00;H4=25.00;H5=40.00;H6=55.00;H7=70.00;H8=55.00;H9=40.00;H10=25.00
AI_OVERRIDE | H10 EXIT | bankroll £25.00 | trend -4 | drawdown £75.00
Publication boundary: experimental synthesis controller. Descriptive deterministic comparison only; no claim of casino advantage, card prediction or guaranteed profitability.
================ END ULTIMATE AI | SESSION 20260916_184907 ================
SHUFFLE_ANALYSIS_STATUS | session 20260916_184907 | COMPLETE
================ END AUTOMATIC SHUFFLE ROBUSTNESS | SESSION 20260916_184907 ================
