Bayesball: A Bayesian hierarchical model for evaluating fielding in major league baseball

TL;DR

This paper introduces Bayesball, a hierarchical Bayesian model that evaluates individual fielding performance in Major League Baseball using high-resolution spatial data, sharing information across players and seasons.

Contribution

The paper develops a novel Bayesian hierarchical model for assessing fielding performance, integrating spatial data and sharing information among players at each position.

Findings

Bayesball provides detailed fielding evaluations across four MLB seasons.

The model outperforms existing fielding metrics in accuracy.

Sharing information improves individual performance estimates.

Abstract

The use of statistical modeling in baseball has received substantial attention recently in both the media and academic community. We focus on a relatively under-explored topic: the use of statistical models for the analysis of fielding based on high-resolution data consisting of on-field location of batted balls. We combine spatial modeling with a hierarchical Bayesian structure in order to evaluate the performance of individual fielders while sharing information between fielders at each position. We present results across four seasons of MLB data (2002--2005) and compare our approach to other fielding evaluation procedures.