Using Conventional and Sabermetric Baseball Statistics for Predicting Major League Baseball Win Percentage
Open Access
Author:
Decesare, Victoria Ellen
Area of Honors:
Statistics
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Andrew John Wiesner, Thesis Supervisor David Russell Hunter, Thesis Honors Advisor
Keywords:
MLB Major League Baseball baseball sabermetrics regression analysis statistics correlation sports statistics
Abstract:
Major League Baseball is dominated by statistical analysis; one cannot watch a baseball game on the television without hearing and seeing a plethora of statistics such as batting average, runs batted in, earned run average, and the list goes on. In addition to these popular stats that most people are familiar with, there are several, more complex baseball statistics – known as “sabermetric” statistics – that have been developed over the past few decades that seek to evaluate players and the game more scientifically and comprehensively. However, with all of these stats available, it is easy to get caught up in the data and overlook the main goal of MLB teams: to win games. With this in mind, the goal of this research is to explore some of the numerous baseball statistics available, both the traditional and modern ones, and observe which ones are truly the best at predicting wins. Encompassing this, is it better to use the more complex methods in analyzing how teams win, or does it hold true that “less is more”? This research seeks to answer these questions and to provide a unique perspective for fans and managers alike when trying to make use of the ever-growing world of baseball data.
