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 · sport=nba

Overall: N = 42 · mean predicted 54.1% · actual win rate 47.6% · gap -6.5pp · Brier 0.268 · log-loss 0.732
Calibration by predicted-probability bucket.
Predicted-prob bucket N Mean predicted Actual win rate Gap (actual − predicted) Brier
<50% 9 49.0% 44.4% -4.6pp 0.259
50-55% 21 52.6% 47.6% -5.0pp 0.252
55-60% 7 56.6% 85.7% +29.1pp 0.209
60-65% 3 63.7% 0.0% -63.7pp 0.406
65-70% 1 68.0% 0.0% -68.0pp 0.462
70-75% 1 72.0% 0.0% -72.0pp 0.518