4D · Backtesting Observatory
A walk-forward test: for each of the last 100 draws, the formulas see only earlier draws, generate 10 tickets, and are scored against what actually came out. The same number of fair random tickets is the baseline.
Formula
Want finer control? Tune the weights in the Formula Lab, then choose “Backtest these settings”.
Replaying 100 4D draws and 1,000 random portfolios…
Based on 5,342 historical draws up to 2026-10-24. Source and provenance
Prize hits
4
2.3 expected by chance
vs chance
1.74×
95% interval 0.47–4.45×
p-value
0.301
1.000 after Bonferroni
Monte Carlo
78th
percentile of 1,000 random runs
Consistent with chance. The difference from random tickets is the size you’d expect from luck alone.
1,000 tickets from 10 engines, seed 1. A hit is any prize-winning ticket.
| 1st prize | 0 | 0.1 |
|---|---|---|
| 2nd prize | 0 | 0.1 |
| 3rd prize | 1 | 0.1 |
| Starter | 2 | 1.0 |
| Consolation | 1 | 1.0 |
Random runs: median 2, 90% between 0 and 5.
96 of 100 draws won nothing. Every miss counts in the totals above.
