Classifying batted ball outcomes from Division I collegiate baseball players

J Sports Sci. 2024 May 12:1-7. doi: 10.1080/02640414.2024.2351666. Online ahead of print.

Abstract

Modern technology challenges anecdotal beliefs on baseball performance. The study's purpose examines these beliefs by classifying batted ball outcomes. Three categories of independent variables (anthropometry, in-game situation, technique-based), from 1,922 batted ball outcomes produced by 230 players, were used to classify the likelihood of hits during 2021 college baseball games. Anthropometry included player's heights and weights. In-game situation entailed batter side, same side, ahead count, and pitch type. Technique-based variables measured by TrackMan radar included exit speed (ExSp), launch angle (LA), batted ball distance (BBD), and hang time (HT). Binary logistic regression analysis was performed with batted ball outcome as the dependent variable. Independent variables provided a good fit (χ2 (10) = 522.358, p < 0.01) and correctly classified nearly three-fourths of outcomes. Height (β = 0.030, p < 0.05), ExSp (β = 0.023, p < 0.05), LA (β = 0.028, p < 0.01), and BBD (β = 0.067, p < 0.01) each had significant positive associations, yet HT (β = -1.661, p < 0.01) had a significant negative association, with batted ball outcomes. TrackMan provided four significant independent variables. Anthropometry's contribution to batting outcome was modest, while in-game situation's impact was non-significant; results contradict anecdotal beliefs of their importance.

Keywords: Exit speed; batted ball distance; hang time; height; launch angle.