Win probability
Game Prediction Model
- 20seasons of data2006–2025
- 5,133games trained on4,166 more held back for testing
- 64.7%called correctlyon games the model never saw
- ±2.1points of calibration driftwhen it says 70%, it happens about 70% of the time
23.2%decided by 3 points or fewer
47.5%decided by 7 points or fewer
8points, the median margin
A quarter of NFL games come down to a field goal, which is the ceiling every forecaster runs into. Predicting every game decided by more than a touchdown perfectly, and coin-flipping the rest, would still only reach 76.3%. Across 5,133 games, 2006–2025.
Week 3
16 games · 5 of them land in the coin-flip band, where the model has no usable call
- ATLatGBThu, Sep 24, 8:15 PMGB 71%GB -5.8High confidence
- SEAatWASSun, Sep 27, 1:00 PMSEA 57%SEA -4.1Low confidence
- CINatPITSun, Sep 27, 1:00 PMCIN 52%CIN -1.4Coin flip, no call
- LACatBUFSun, Sep 27, 1:00 PMBUF 71%BUF -6.2High confidence
- NYJatDETSun, Sep 27, 1:00 PMDET 75%DET -9.1High confidence
- HOUatINDSun, Sep 27, 1:00 PMIND 52%HOU -1.0Coin flip, no call
- CARatCLESun, Sep 27, 1:00 PMCAR 53%CAR -1.9Coin flip, no call
- TENatNYGSun, Sep 27, 1:00 PMNYG 67%NYG -3.8High confidence
- NEatJAXSun, Sep 27, 1:00 PMJAX 55%JAX -1.3Low confidence
- KCatMIASun, Sep 27, 1:00 PMKC 73%KC -8.3High confidence
- MINatTBSun, Sep 27, 4:05 PMTB 61%TB -1.3Medium confidence
- ARIatSFSun, Sep 27, 4:05 PMSF 81%SF -11.6High confidence
- LVatNOSun, Sep 27, 4:25 PMNO 67%NO -4.4High confidence
- BALatDALSun, Sep 27, 4:25 PMBAL 59%BAL -1.6Low confidence
- LARatDENSun, Sep 27, 8:20 PMLAR 53%LAR -3.3Coin flip, no call
- PHIatCHIMon, Sep 28, 8:15 PMCHI 54%PHI -0.3Coin flip, no call
Calibrated Logistic Regression (core + market features) · walk-forward by season (train on all prior seasons, test on the next)