Interpretable Pre-Release Baseball Pitch Type Anticipation from Broadcast 3D Kinematics

TL;DR

This study demonstrates that a machine learning pipeline can classify eight different baseball pitch types from monocular 3D pose sequences with high accuracy, revealing key biomechanical features and the limits of kinematic prediction.

Contribution

The paper introduces a large-scale benchmark for pitch type classification from 3D pose data and identifies the biomechanical features most predictive of pitch type.

Findings

Achieved 80.4% accuracy on 119,561 pitches using body kinematics.

Upper-body mechanics contribute 64.9% of predictive signal.

Pose-based classification cannot distinguish grip variants, indicating a ceiling near 80%."

Abstract

How much can a pitcher's body reveal about the upcoming pitch? We study this question at scale by classifying eight pitch types from monocular 3D pose sequences, without access to ball-flight data. Our pipeline chains a diffusion-based 3D pose backbone with automatic pitching-event detection, groundtruth-validated biomechanical feature extraction, and gradient-boosted classification over 229 kinematic features. Evaluated on 119,561 professional pitches, the largest such benchmark to date, we achieve 80.4\% accuracy using body kinematics alone. A systematic importance analysis reveals that upper-body mechanics contribute 64.9\% of the predictive signal versus 35.1\% for the lower body, with wrist position (14.8\%) and trunk lateral tilt emerging as the most informative joint group and biomechanical feature, respectively. We further show that grip-defined variants (four-seam vs.\ two-seam fastball) are not separable from pose, establishing an empirical ceiling near 80\% and delineating where kinematic information ends and ball-flight information begins.