The mainstream media, when talking about baseball player evaluation, often notes the tension between scouts and stats (i.e. Moneyball) Even though this tension is overstated, the baseball community treats these two types of information as independent entities. We address this problem with a Bayesian framework. Our method combines quantitative scouting information (grades) with minor league performance to predict a player’s major league performance, yielding an entire distribution of predicted batting performance outcomes. We compare our predictions to those naive of scouting grades and established projection systems.
