573 – Analysis of Baseball, Basketball, and Cricket Data
Logistic Regression Based Simulation of Major League Baseball Seasons
Richard Auer
Loyola University Maryland
Claire Marie Reynolds
Loyola University Maryland
A logistic regression model, based on all 2430 Major League Baseball games from the 2010 season, is developed to simulate all of the games of that season and the ensuing playoff games. This simulation is performed 1000 times. The simulation is conducted using a 3000+ line program coded in the package, Matlab. The intent is to observe just how much variability is possible in the final divisional standings and the winner of the World Series given one set of 30 team strengths.