OUR OWN MODEL · MLB
Sports Edge — a model you can see inside
Our own MLB win-probability model — built entirely from real game results and the dominant signal, starting-pitcher quality. For every upcoming game you get the model's win-%, and exactly why: each team's Elo rating and each starter's ERA. No black box, no bought feed.
How it works & the honest fine print. Team strength (Elo, home-field derived from real data) + each starting pitcher's innings-regressed ERA, blended by a logistic regression. It's validated out-of-sample at ~61% accuracy (vs 50% for picking the home team) on games it never trained on. This is a model estimate, not a guarantee — MLB has high variance, and we publish the real number, never a fabricated one. The live track record builds forward, game by game.
Training the model & loading today's slate…