Transparency
Model Accuracy
Every prediction is scored against the real result using only what the model knew beforehand — a fair, out-of-sample test across 10,850 bouts. The only boxing model that grades itself.
89.2%
On confident picks
when the model commits
76.8%
Every bout
incl. true coin-flips
0.182
Brier score
0.25 = a coin-flip
Calibration — the real test
Raw accuracy is the wrong yardstick in a chaotic sport — what matters is honesty: a trustworthy model wins about as often as it says it will. When we call a fighter a 70% favourite, they should win ~70%.
50–60%
said 55% · won 66%
n=4607
60–70%
said 65% · won 87%
n=4657
70–80%
said 72% · won 97%
n=1586