Play Whe · Audit

How we test these predictions

Each method is replayed on every past draw using only the draws before it, and scored on whether the drawn mark was in its top five. Chance is 5 in 36. Every draw is independent.

Prediction audit752 draws tested, through #27449
MethodBacktest hit rateChance95% intervalpStatus
Random top 510.6%13.9%8.6% to 13.0% (expected 11.4% to 16.4%)Baseline
History says: overdue14.6%13.9%12.3% to 17.3%p = 0.29Descriptive
Follower11.3%13.9%9.2% to 13.8%p = 0.98 · perm 0.97Tested
Slot14.0%13.9%11.7% to 16.6%p = 0.49 · perm 0.31Tested
Weekday12.0%13.9%9.8% to 14.5%p = 0.94 · perm 0.90Tested
Hot13.8%13.9%11.5% to 16.5%p = 0.53 · perm 0.38Tested
Weighted signals14.0%13.9%11.7% to 16.6%p = 0.49Tested
Model (softmax ensemble, default weights)13.3%13.9%11.1% to 15.9%p = 0.70Tested

No measurable evidence of a slot or weekday bias in this sample (p = 0.19 and 0.94)

Pre-registered method: how the audit works. Re-run after every draw with 1000 shuffles. A method enters a ranked list only when its interval clears 13.9% and holds on unseen draws.

Predictions · Method in plain words