Ensemble confirmation
Your existing strategy formula is the core engine and runs untouched. The statistical layer — weighted frequency, EMA momentum, gap, Z-score, Bayesian, pair and triplet association, Markov transitions, Monte Carlo and entropy — is an additional validation layer. It never replaces a formula candidate; it only confirms, supports or flags a conflict.
Starts at 70 / 30. Change it only when the out-of-sample table below shows a better ratio.
Loading history…
Step 2 — existing formula board
No formula output — enable at least one strategy and import draws.
Step 8 — banker verdict
—
NO BANKER
The formula produced no banker for this slot.
- Statistical rank of formula banker
- —
- Statistical challenger (not applied)
- 36
Step 5 — candidate classification
| No. | Formula | Statistical | Agreement | Ensemble | Class | Strategies |
|---|---|---|---|---|---|---|
| 36 | — | 56% · #1 | 44% | 17% | STATISTICAL ONLY | statistical layer only |
| 06 | — | 55% · #2 | 45% | 17% | STATISTICAL ONLY | statistical layer only |
| 08 | — | 55% · #3 | 45% | 17% | STATISTICAL ONLY | statistical layer only |
| 05 | — | 55% · #4 | 45% | 17% | STATISTICAL ONLY | statistical layer only |
| 11 | — | 55% · #5 | 45% | 16% | STATISTICAL ONLY | statistical layer only |
| 31 | — | 55% · #6 | 45% | 16% | STATISTICAL ONLY | statistical layer only |
| 15 | — | 55% · #7 | 45% | 16% | STATISTICAL ONLY | statistical layer only |
| 26 | — | 55% · #8 | 45% | 16% | STATISTICAL ONLY | statistical layer only |
| 30 | — | 55% · #9 | 45% | 16% | STATISTICAL ONLY | statistical layer only |
| 19 | — | 54% · #10 | 46% | 16% | STATISTICAL ONLY | statistical layer only |
| 37 | — | 54% · #11 | 46% | 16% | STATISTICAL ONLY | statistical layer only |
| 16 | — | 54% · #12 | 46% | 16% | STATISTICAL ONLY | statistical layer only |
| 07 | — | 53% · #13 | 47% | 16% | STATISTICAL ONLY | statistical layer only |
| 23 | — | 53% · #14 | 47% | 16% | STATISTICAL ONLY | statistical layer only |
| 29 | — | 53% · #15 | 47% | 16% | STATISTICAL ONLY | statistical layer only |
Formula candidates are always listed first — a statistical-only number never displaces one, it is only shown as a watch item.
Step 9 — pairs (formula first)
No pairs yet.
Step 9 — triplets
No triplets yet.
Step 3 — statistical model detail
| No. | W. freq | EMA | Gap | Z | Bayes | Markov | Monte Carlo |
|---|---|---|---|---|---|---|---|
| 36 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 13.2% |
| 06 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 13.1% |
| 08 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 13.0% |
| 05 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 13.0% |
| 11 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 13.0% |
| 31 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 13.0% |
| 15 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 12.9% |
| 26 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 12.9% |
| 30 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 12.9% |
| 19 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 12.8% |
| 37 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 12.8% |
| 16 | 50% | 50% | —/8.2 | 0.00 | 12.2% | 12.2% | 12.8% |
Sample 0 draws · entropy 0.000 of 5.615 (0% of maximum randomness).
Step 10 — component scoreboard (out-of-sample)
Needs at least 20 stored draws and one active strategy.
Core engine: your strategy formula. Statistical models: validation and enhancement. Backtesting: performance judge. Actual draws: feedback. The original strategy is never overwritten by a differing statistical ranking.