Model calibration audit

For every bucket of predicted probability, what did the model actually hit? A well-calibrated model has actual win rate ≈ predicted prob. A negative gap means the model is overconfident in that bucket (dangerous); positive gap means it's underpredicting (safe). Rows turn red when |gap| > 5pp AND N ≥ 10.

Calibration cohort by sport (click to filter):
All sports MLB N=3,353TENNIS_WTA N=753TENNIS_ATP N=748MMA_MIXED_MARTIAL_ARTS N=407LIGUE1 N=89NFL_PRESEASON N=65NBA N=42NHL N=33LALIGA N=26SERIEA N=19EPL N=15UCL N=13NCAAB N=11CRICKET_IPL N=6BUNDESLIGA N=3

Combined "all sports" is rarely meaningful — sports differ in market efficiency, signal availability, and base rates. Use the chips to drill into a single sport. N<50 (red) means the calibration is brittle; N≥200 (green) is trustworthy.

Filter: window=90d

Overall: N = 5,583 · mean predicted 57.8% · actual win rate 55.5% · gap -2.3pp · Brier 0.238 · log-loss 0.667
Calibration by predicted-probability bucket.
Predicted-prob bucket N Mean predicted Actual win rate Gap (actual − predicted) Brier
<50% 710 44.5% 41.1% -3.4pp 0.242
50-55% 1732 52.3% 52.8% +0.5pp 0.249
55-60% 1196 57.2% 55.0% -2.1pp 0.248
60-65% 939 61.5% 51.2% -10.3pp 0.260
65-70% 452 67.1% 69.5% +2.3pp 0.212
70-75% 156 72.0% 73.7% +1.7pp 0.194
75-80% 241 76.6% 74.7% -1.9pp 0.191
80-90% 100 85.9% 89.0% +3.1pp 0.099
90%+ 57 94.5% 93.0% -1.6pp 0.065