Creating a Hitter's Approach: Analyzing At-Bat Data




Ramos, Dominic C.

Journal Title

Journal ISSN

Volume Title



In this paper, we begin the process of creating an approach for a hitter in baseball dependent on predicting the pitch location. Predicting pitch location has not been introduced in academics up to this point. Most predictions have been done on pitch type or on the decision of the hitter or swing to wait on a pitch. In this study, we find a way to help the hitter create five locations in the hitting-zone to look for pitches to cross the plate and how to maximize the hitter's chance of predicting correctly in different scenarios of a game. Creating an approach dependent on predicting the pitch location was found by using contour maps, the Gaussian mixture model to create five areas for the hitter to set as approaches, and a multinomial logistic regression to explain when to change from one location to another during different situations in a game. The hitter learns to create approaches to handle at-bats when facing a right-handed or left-handed pitcher, when the hitter has strikes on the count, when the game increases in innings played, when outs increase, and when baserunners are on second and third. The future is bright for predicting pitch location. Teams who can implement these strategies will have an advantage over teams who do not.



Baseball, At-bat data, Hitter's approach, Pitch prediction, PitchF/X, Statcast


Ramos, D. C. (2017). <i>Creating a Hitter's approach: Analyzing at-bat data</i> (Unpublished thesis). Texas State University, San Marcos, Texas.


Rights Holder

Rights License

Rights URI