Analytics
A deeper read on the model than the headline numbers — which prop types it is actually good at, how that has moved event by event, and how individual players have finished. Back to the headline track record.
Model version v20260502_135254
8 events · 5,656 predictions
Accuracy by Prop Type
ModelBaseline (base rate)Brier score — lower is better
Make Cut0.256vs 0.2496 eventsn=915
Matchup0.255vs 0.2298 eventsn=505
Top 100.082vs 0.0778 eventsn=1,059
Top 200.144vs 0.1418 eventsn=1,059
Top 50.043vs 0.0428 eventsn=1,059
Win0.008vs 0.0078 eventsn=1,059
Brier Score by Event
Make Cut (6 events, n=915)Matchup (8 events, n=505)Top 10 (8 events, n=1,059)Top 20 (8 events, n=1,059)Top 5 (8 events, n=1,059)Win (8 events, n=1,059)
Model v20260502_135254 · Brier score, lower is better. Events with no row for a prop (for example the make-cut prop at a no-cut event) leave a gap rather than a zero.