DETERMINISTIC BLACKJACK — POST-SESSION PUBLICATION PLATE REQUEST

I am attaching the complete post-session publication handoff ZIP produced after running S.

IMPORTANT:
The ZIP itself is required input to this request. Do not begin the publication process without inspecting the attached ZIP and its SESSION_EVIDENCE contents.

Your task is to produce the final publication-ready Session Scorecards Comparison plate for the session contained in this ZIP.

============================================================

1. FIRST: INSPECT THE COMPLETE HANDOFF
   ============================================================

Before drawing anything:

• Inspect the complete attached ZIP.
• Identify the verified S/post-session manifest and all session evidence.
• Inspect the plates/reference-material area.
• Inspect benchmark_visual.png carefully.
• Inspect the previous accepted publication plates supplied in the ZIP.
• Inspect any specialised component/layout reference images supplied in the plates area, including the compact Hybrid-vs-Frozen comparison + Risk/Journey N/A layout reference where applicable.

Do not start drawing until these materials have been reviewed.

The S-generated verified manifest/evidence is the DATA AUTHORITY.

The accepted previous plates and benchmark_visual.png are the VISUAL / PUBLICATION AUTHORITY.

Never obtain Session N values by reading values from an older publication plate.

DO NOT DESIGN THE PLATE FROM SCRATCH.

Repeat: DO NOT DESIGN A NEW SCORECARD LAYOUT.

This is an inheritance/adaptation task.

The established Session Scorecards Comparison publication family already exists.

Use benchmark_visual.png as the primary benchmark for the expected visual quality, density, hierarchy, typography, borders, colour treatment, icons, spacing and overall publication finish.

The final result must look as though it belongs directly beside the accepted Session 7/9/10/11 publication plates.

If your proposed result looks like a generic dashboard, simplified infographic, wireframe, draft report, newly invented scorecard or substantially different visual language, STOP and return to benchmark_visual.png.

ENHANCE THE CURRENT SESSION PLATE ACCORDING TO benchmark_visual.png ON THE CANVAS.

Do not merely use benchmark_visual.png for inspiration. Treat it as the principal publication-quality visual benchmark.

Determine from the verified S evidence what kind of session occurred:

• observed Frozen;
• observed Hybrid;
• complete 30-hand journey;
• early-terminated observed journey;
• source-complete or source-exhausted counterfactual;
• availability of Frozen, Casual, Stat-Watching, Ultimate AI and other declared side-controller evidence;
• eligibility or non-eligibility of Risk/Journey scores.

Then inherit the closest established publication structure/components from the supplied accepted plates.

SPECIAL CASE — INCOMPLETE / EARLY-TERMINATED OBSERVED HYBRID:
When the current session is an incomplete or early-terminated observed Hybrid journey, Session12.png (when supplied) is the PRIMARY WHOLE-PLATE STRUCTURAL PRECEDENT because it tells the same evidence-state story. Preserve its overall geometry, relative section sizes, running-P/L treatment, source-exhausted counterfactual treatment, lower analysis structure and N/A score treatment. Then apply benchmark_visual.png as a SECOND, SEPARATE editorial transformation pass so the inherited Session-12-like structure gains the polished typography, iconography, border weight, colour treatment, density and publication finish of the benchmark. The intended operation is: verified Session-N evidence -> Session12-like structural plate -> benchmark_visual editorial transformation. Do not skip the middle structural state and do not replace it with a newly designed dashboard.

CRITICAL PROTECTED COMPONENT: if the specialised benchmark_na_score_layout.png applies, the HYBRID vs FROZEN COMPARISON table plus the two stacked RISK SCORE / JOURNEY INTENSITY SCORE N/A boxes form ONE protected component group. Keep that group present, together, readable and in the established lower-left position. The benchmark_visual transformation may restyle this group, but MUST NOT delete it, split it apart, move it to another conceptual section, replace it with generic score boxes, or lose any of its evidence rows. After the editorial transformation, explicitly verify that this protected component still exists unchanged in information content.

Do NOT force Session 12, or any future session, into a 30-hand template when its verified evidence says otherwise.

Do NOT invent missing replay outcomes.

Do NOT turn source-supported partial evidence into completed-session evidence.

Every Session N number must come from the verified S evidence/manifest.

This includes, where available:

• observed bankroll path;
• starting/final bankroll;
• session P/L;
• peak and trough and their hand indexes;
• maximum drawdown;
• hands completed;
• W/L/P;
• exposure;
• average wager;
• wager match / above / below;
• observed wager count (for Hybrid wager/action summary headings);
• action match / different;
• decision counts and provenance;
• cards captured;
• shuffles;
• amendment/correction information;
• formal-start Frozen replay;
• formal-start Casual replay;
• source-exhaustion state;
• Risk Score;
• Journey Intensity Score and components;
• Stat-Watching evidence;
• Ultimate AI evidence;
• £200 reach;
• highest bankroll;
• complete A–E shuffle robustness results;
• all other established publication fields supported by S.

Never substitute an older session's value.

Never infer an unavailable value merely to fill a box.

Never convert N/A into zero.

Never confuse observed Hybrid results with Frozen/Casual reconstruction results.

Before final rendering, internally cross-check the plate against the verified S manifest.

For an observed Hybrid session, preserve the established:

HYBRID vs FROZEN COMPARISON (SESSION N)

component whenever supported by the verified evidence.

It must clearly show the verified values for:

Frozen decisions recorded
Hybrid/Frozen comparable decisions
Same decision
Different decision
Decision alignment
Wager match / above / below
Observed wager count = wager match + above + below. NEVER use the wager-match count alone as the observed-wager count. Example: 0 / 5 / 4 means 9 observed wagers, not 0. The HYBRID WAGER / ACTION SUMMARY heading must therefore say (9 observed wagers) for that evidence state.

Retain the explanatory statement:

“Action and wager divergence retained separately.”

Do not omit this component merely to simplify the design.

Risk Score and Journey Intensity Score remain meaningful publication fields even when the observed journey is not eligible for completed scores.

DO NOT REMOVE THEM.

When the observed Hybrid Risk Score and Journey Intensity Score are N/A because a complete replay/path is required, use the specialised compact layout reference supplied in the ZIP.

First make the HYBRID vs FROZEN COMPARISON table efficient:

• remove excessive horizontal padding;
• keep the same evidence rows;
• bring the numerical value column closer to its labels;
• retain readable publication typography;
• do not achieve compactness by making the text tiny;
• preserve the established blue heading, pale row treatment and explanatory footer.

Then use the recovered horizontal space for TWO COMPACT STACKED BOXES:

RISK SCORE
N/A
(requires complete replay)

and

JOURNEY INTENSITY SCORE
N/A
(requires complete replay)

Retain the established shield and journey/path icons.

Follow the supplied compact N/A-score reference image closely.

This reference is an ESTABLISHED COMPONENT LAYOUT, not permission to create a new design.

When numerical Risk/Journey scores ARE valid, use the established numerical-score presentation from the accepted publication plates instead.

Preserve the established controller/player iconography wherever applicable.

Examples include:

• Human-controlled / observed Hybrid;
• Frozen reference;
• Fixed-policy Casual;
• Stat-Watching Casual;
• Ultimate AI.

Use the same visual vocabulary established by the publication plates and GUI/replay environment.

Do not replace these with arbitrary new icons.

Use the complete verified S-generated A–E robustness results.

Preserve the established dense publication table rather than replacing it with a graph or simplified chart unless an established benchmark plate specifically requires otherwise.

Do NOT independently recalculate robustness values during publication rendering.

Do NOT borrow robustness values from an earlier session.

Preserve source-exhaustion/completion semantics exactly.

Retain the established lower publication-analysis structure where supported, including the established families such as:

KEY FINDINGS (SESSION N)
MOST SIGNIFICANT OBSERVATIONS
ROBUSTNESS INTERPRETATION
OVERALL CONCLUSION (SESSION N)

Populate these from the verified evidence.

Keep statements descriptive and evidence-based.

Do not invent casino-population implications, predictive claims or gambling recommendations.

Create the actual finished visual plate on the canvas.

This request is NOT asking for a textual mock-up of the plate.

Render the complete publication graphic.

Use benchmark_visual.png as the quality target while rendering.

IMPORTANT: benchmark_visual.png controls the FINAL STYLE, not the current-session structure or evidence. First construct/inherit the correct evidence-state structure (for incomplete observed Hybrid, use Session12.png when supplied); only then transform that already-correct plate to the benchmark_visual publication finish. During that transformation, lock all established components and Session-N values in place. Styling is allowed; structural deletion is not.

The plate should have the same high information density, strong section hierarchy, crisp table/grid construction and professional publication finish as the accepted benchmark plates.

The finished PNG should be a genuinely high-resolution publication artifact.

As a practical quality check, established final plates are generally substantial PNG files, commonly around 1.5 MB or greater. Do not deliberately pad a file to reach a size, but a very small/lightweight output should trigger a check that the plate has not been rendered at insufficient resolution or complexity.

Before presenting the result, compare it visually AGAIN against benchmark_visual.png and the closest accepted previous plate.

Ask internally:

• Does this unmistakably belong to the same publication?
• Did I inherit rather than redesign?
• Is the information density comparable?
• Are the typography and borders crisp?
• Are established components present?
• Are N/A Risk/Journey fields still visible where required?
• Is the Hybrid-vs-Frozen table efficiently proportioned?
• For an incomplete observed Hybrid, does the whole plate still follow the Session12-like evidence-state structure after benchmark styling?
• Is the protected Hybrid-vs-Frozen + Risk/Journey N/A component group still present, together and fully populated after the benchmark transformation?
• Is any text clipped, overlapping or corrupted?
• Is the bottom analysis area complete?
• Are all Session N values supported by S?
• Did any value accidentally come from an older plate?

If any answer is unsatisfactory, correct the plate BEFORE presenting it.

Keep track of the process used to reach the final publication-ready state.

In particular record:

• which S evidence was treated as authoritative;
• which accepted plate was selected as the closest structural precedent;
• how benchmark_visual.png was used;
• which specialised component references were inherited;
• whether Risk/Journey were eligible or N/A;
• how source-exhaustion was represented;
• what final integrity/visual checks were performed.

This process is intended to become repeatable for later sessions and usable by a non-technical operator/croupier without requiring them to diagnose plate-design problems manually.

Proceed now from the attached ZIP.

Do not reuse or repair a previously rejected draft plate.

Start the publication-rendering stage again from the verified S handoff.

DO NOT DESIGN FROM SCRATCH.

INHERIT THE ESTABLISHED SCORECARD FAMILY.

USE benchmark_visual.png AS THE PRIMARY VISUAL QUALITY AUTHORITY.

USE THE SPECIALISED COMPONENT REFERENCE IMAGES WHEN THEIR CONDITIONS APPLY.

ENHANCE AND RENDER THE SESSION PLATE ON THE CANVAS TO THE STANDARD OF benchmark_visual.png.

VERIFY EVERY SESSION VALUE AGAINST S BEFORE FINAL DELIVERY.

