Predicting Major League Baseball Playoff Outcomes through Multiple Linear Regression
Date of Award
Econometrics, Linear Probability Model, Sabermetrics, MLB Playoffs
Statistics and Probability
The Major League Baseball playoff system involves eight teams competing for the title of World Series Champion. Each year the National League and American League send their top four teams to compete for this honor. Like many other sports, Major League Baseball evaluates individual teams and players based on a variety of statistics. These statistics can be used to compare, rank, and evaluate the skill levels of every entity involved. Most notably, statistics can also be used to project the outcomes of future events through multiple linear regression analysis. Using past playoff data to develop linear probability models, the ability to predict future playoff match-ups is made possible. It is this type of analysis that this paper will discuss and evaluate. By looking at various metric evaluation tools such as sabermetrics and betting lines, I will begin to develop the framework of playoff predictions. Finally, producing linear probability models of my own I will show the likelihood of developing an accurate playoff outcome prediction tool.
Firsick, Zachary, "Predicting Major League Baseball Playoff Outcomes through Multiple Linear Regression" (2013). Honors Thesis. 207.