GamePlan Science

Sports analytics fails to paint a full picture of team performance or provide meaningful uncertainty estimates for predictions.

GamePlan Science aims to fill this gap. We build probabilistic models of team performance that estimate true team strength. All of our parameters can be explained to the common fan, we can quantify all of our parameters' uncertainty given the sample size, and we can quantify our uncertainty on game outcomes. This approach allows us to explore several aspects of team strength such as season-by-season roster changes, home-field advantage, injuries, etc.