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=3,584TENNIS_WTA N=909TENNIS_ATP N=873MMA_MIXED_MARTIAL_ARTS N=388NFL_PRESEASON N=145LALIGA N=116NCAAF N=95SERIEA N=87EPL N=86LIGUE1 N=84UCL N=61BUNDESLIGA N=53NBA N=12NCAAB N=11NHL N=9AMERICANFOOTBALL_NFL N=2

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=nfl_preseason

Overall: N = 145 · mean predicted 55.6% · actual win rate 58.6% · gap +3.0pp · Brier 0.243 · log-loss 0.678
Calibration by predicted-probability bucket.
Predicted-prob bucket N Mean predicted Actual win rate Gap (actual − predicted) Brier
<50% 11 49.5% 36.4% -13.2pp 0.249
50-55% 67 52.1% 61.2% +9.1pp 0.248
55-60% 40 57.4% 52.5% -4.9pp 0.252
60-65% 19 61.7% 63.2% +1.5pp 0.238
65-70% 5 67.1% 100.0% +32.9pp 0.108
70-75% 3 73.4% 66.7% -6.7pp 0.224