DETERMINISTIC BLACKJACK DECISION ENGINE RESEARCH
PHASE 3.164 — FINAL INDEPENDENT HOLDOUT / BRANCH-CLOSURE TEST
LAYMAN / SME EXPLANATION

This was designed as the final independent test of the current £60 / £70 refinement branch.

Nothing was changed before the run.

Fresh cohorts:
BA / BB / BC / BD
20,000 sessions total.

The same +5 checkpoint, same wagering behaviour, same card-play policy and same three intervention intensities were used.

TWO QUESTIONS WERE DECLARED BEFORE THE RESULTS EXISTED

1. Would the intervention ordering reproduce again?

Expected:
Original Capital Risk should intervene most, prevent the most terminal-zero outcomes, and lose the most £200 outcomes.

£60 should intervene least, prevent fewer terminal-zero outcomes, and lose the fewest £200 outcomes.

£70 should remain between them.

2. Would the <=£60 region still look much more terminal-risk concentrated than the upper £80-100 Capital-Risk region?

RESULT:
ALL FOUR PREDECLARED TESTS PASSED.

THE LIVE POLICY RESULT

REFERENCE
Average final bankroll £84.48
£200 21.69%
£0 39.35%

ORIGINAL CAPITAL RISK
Average final bankroll £85.06
£200 21.13%
£0 37.22%
Activation 7.35%

Against reference:
427 terminal-zero outcomes prevented
113 £200 outcomes lost
+£11,450 net paired bankroll
£14.10 less exposure per session
£2.54 lower maximum drawdown

REFINED £60
Average final bankroll £84.61
£200 21.64%
£0 38.38%
Activation 2.10%

Against reference:
194 terminal-zero outcomes prevented
only 11 £200 outcomes lost
+£2,426.25 net paired bankroll

REFINED £70
Average final bankroll £84.65
£200 21.59%
£0 38.12%
Activation 2.99%

Against reference:
247 terminal-zero outcomes prevented
20 £200 outcomes lost
+£3,225 net paired bankroll

THE ORDERING PASSED AGAIN

Activation:
Original > £70 > £60.

Terminal-zero prevention:
Original > £70 > £60.

£200 cost:
Original > £70 > £60.

That is the same objective ordering seen in the previous prospective populations.

THE REGIONAL HOLDOUT

<=£60 region:
46.30% were terminal-zero prevention cases.
2.63% were lost-£200 cases.

£80-100 region:
17.08% terminal-zero prevention.
12.81% lost-£200.

This is a very strong replication of the regional interpretation.

The low Capital-Risk region is far more dominated by terminal-risk prevention.

The upper Capital-Risk region is substantially more mixed with target opportunity.

WHAT DOES THIS MEAN FOR THE BRANCH?

The branch has now been tested through:

discovery,
fresh validation,
prospective rule specification,
multiple fresh live executions,
pooled precision analysis,
checkpoint instrumentation,
boundary-neighbourhood analysis,
broader-region analysis,
cohort robustness,
and now an independent final holdout.

The holdout did not require us to rescue or reinterpret the hypothesis.

It passed exactly in the direction declared beforehand.

PHASE 3.164 VERDICT

CURRENT £60 / £70 CAPITAL-RISK REFINEMENT BRANCH — SUFFICIENTLY PROVEN FOR CONSOLIDATION.

This does NOT mean blackjack itself has become profitable or that one intervention is universally best.

It means the evidence is now strong that the three intervention intensities genuinely represent different objective trade-offs:

ORIGINAL <£100:
maximum tested capital protection and highest target cost.

<=£60:
selective capital protection and lowest target cost.

<=£70:
intermediate.

The £60 and £70 values should still be understood as practical cut-points through a broader bankroll-risk gradient, not unique mathematical optima.

NEXT

The correct next step is NOT another threshold tweak.

We should consolidate this branch in the cumulative research record and then return to the brainstorm to select the next major unresolved investigation.

If desired, the next numbered phase can be used as a formal BRANCH CONSOLIDATION / HANDOVER phase before beginning the next research branch.
