PHASE 3.32 — NEXT DESIGN NOTE

DESIGN CORRECTION
Current-hand cards and dealer up-card cannot choose the current wager because the wager precedes the deal.

PHASE 3.32 BASIS
1-unit observational reference.
150,000 rounds per cohort.
Only lagged/completed information used for PRE-WAGER candidate mapping.

RESULT
Replicated positive pre-wager families with >=200 observations in EACH cohort: ZERO.

Card counts:
Useful diagnostic variation, but no robust positive replicated trigger.

Previous natural blackjack:
No positive next-hand signal.

Previous-three-hand trends:
No positive replicated signal.

CONCLUSION
Do not increase to 2x yet. The evidence has not earned additional exposure.

PHASE 3.33
Deepen the map with pre-declared lagged combinations while controlling fragmentation and requiring minimum exposure plus independent Holdout replication.

Candidate combinations:
previous outcome + player/dealer card-count band;
previous 2-3 outcomes + bankroll band;
previous completed hand class/up-card + outcome;
blackjack status + outcome;
session-position band + previous-history state.

Do not search indefinitely for a positive result.
If robust signal remains absent, retain that negative conclusion.

Shuffle remains optional after multiplier research and is not required for project closure.
