Loading...
Game Outcome Model
Transparent training, accuracy, market, and methodology reporting
Data source
Live API
Model accuracy
63.5%(181/285)
Report date
Mar 23, 2026
2025 report
Model overview
Start with the result, sample size, and strongest signals, then use the report navigation for accuracy evidence, simulated market performance, and methodology.
Historical Simulation Net
Sign in to see probability-derived flat-stake ROI
Exact historical metrics are locked
Model Win Rate
181
out of 285
+9-11 ptsabove industry avg
Top Factor
8-Game Scoring
Predictive strength3.9%
High-Confidence Predictions
97%win rate
33 high-confidence games
Ensemble:-0.7% vs. best component
How this model works
QuantCup trains an ensemble on historical NFL data to predict game winners, then tests it on games the model did not train on.
The report exposes accuracy, errors, market disagreement, component weights, and feature importance so the result can be evaluated instead of treated as a black box.
Review methodologyTraining window
2019–2024
Test sample
285 games · 2025
Ensemble
3 component models
Features used
45 of 62