Betting Dashboard
ProModel calibration analysis and betting edge detection. Compare the model's posterior probabilities against market lines to identify value.
How well-calibrated are the model's win probabilities? Points near the diagonal indicate good calibration. Dot size reflects sample count per bin.
Calibration bins are illustrative. Per-bin data will be available in a future update.
Brier Score...
Lower is better. Perfect = 0, coin flip = 0.25
ATS Record...
Spread MAE...
Mean absolute error on point spreads
Betting Edges
Edges are detected by comparing the model's posterior game predictions against market lines. When the model's confidence diverges significantly from the market, it identifies a potential edge on a side or total.
Provide game comparisons to detect edges. Connect market lines (spreads and totals) from your sportsbook to see where the Bayesian model disagrees with the market.
How It Works
- The model generates posterior win probabilities and point spreads for each game.
- Market lines are compared against the model's median spread and total.
- Games where the model's spread differs by more than 2 points from the market are flagged as potential edges.
- Confidence intervals provide context on how certain the model is about each edge.