Estimating the Effect of Pitch Selection and Plate Discipline Using the G-Computation Algorithm

• Activity Number: 184
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Sports
• Author(s): Laura Boehm Vock* and David Vock
• Companies: St. Olaf College and University of Minnesota
• Keywords: Causal inference ; G-computation ; Pitchf/x ; baseball ; counterfactuals

Abstract

Offensive performance in baseball, measured by batting average, on-base percentage, and slugging percentage, depends upon a number of correlated factors: the pitches the batter faces, the batter's choice to swing, and the batter's hitting ability. Using the g-computation algorithm within a causal inference framework, we estimate the causal effects of these factors on offensive performance of baseball players and performance under unobserved scenarios. In most medical applications, a causal framework is used to estimate the effect of one course of treatment at a time; the novelty of this application is that we estimate the effects of all three "treatments": pitch characteristics, swing choice, and hitting ability. We examine data collected using the PITCHf/x system over the 2013 season for Starlin Castro and Andrew McCutchen. We are able to identify the improvement Castro would expect to see in offensive performance were he to improve to McCutchen's level on any of the three factors above: pitch characteristics, swing choice, and hitting ability. Castro's notoriously poor plate discipline is found to be only part of the reason for his inferiority to McCutchen.