DETERMINISTIC BLACKJACK DECISION ENGINE RESEARCH
PHASE 1.13 — PROFIT OCCUPANCY AND RISK-ADJUSTED SESSION ANALYSIS

PLAIN-ENGLISH PACKAGE GUIDE
---------------------------
Phase 1.12 showed the player's balance after 100 hands. Phase 1.13 goes further and asks when the player was actually making a profit during those 100 hands. It records the probability of currently being ahead after each hand, how many hands were spent ahead, when profit first appeared, how long it lasted, and whether uneven wagering improved those outcomes enough to justify the additional money placed at risk.

The experiment is observational. It does not tell a live player when to wager more. It tests fixed simulated wager rules against flat £1 wagering using exactly the same dealt cards.

MAIN FINDING
------------
Temporary profit was common. The flat strategy was ahead at least once in 91.86% of sessions, but only 48.47% of sessions finished ahead after hand 100. Therefore, "I was ahead during the session" and "I finished ahead" are very different outcomes.

The exact fingerprint strategy staked 5.68% more than flat wagering. It reduced the probability of finishing ahead by 0.43 percentage points, slightly reduced time spent ahead, lowered the average ending balance, and increased drawdown. It failed the strict risk-justification test.

The broad fingerprint strategy staked 12.67% more than flat wagering. It increased average time spent ahead by only 0.02 hands, while lowering the probability of finishing ahead, lowering the average ending balance, and increasing drawdown. It also failed the strict risk-justification test.

FILES
-----
DeterministicBlackjackPhase113.java
    Complete Phase 1.13 Java source. Earlier deterministic policy, gameplay,
    fingerprint, replay, wager and session evidence remains preserved.

Phase_1_13_Results.txt
    Complete console results including preserved earlier-phase evidence.

Phase_1_13_Profit_Occupancy_Study.txt
    Main readable evidence file with hand-by-hand probability of being ahead,
    profit occupancy, and risk-adjusted verdicts.

Phase_1_13_Sample_Profit_Journeys.txt
    Lowest, middle and highest exact-model session examples with player-style
    bankroll checkpoints and profit-journey metrics.

Phase_1_13_Replay_Verification.txt
    Independent deterministic rerun verification.

Earlier Phase 1.10–1.12 evidence files are included unchanged for continuity.

FAIR-PLAY BOUNDARY
------------------
The model assumes genuine fair decks, unbiased shuffling and fixed rules. No hidden card, future card, casino system, shuffle seed, remaining-shoe composition, live-play assistance or evasion method enters a wager decision.

PHASE STATUS
------------
PASS — deterministic replay matched exactly. Both uneven-wager models failed the strict risk-justification test, which is a valid and important negative research result.
