Offensive Performance, Omitted Variables, and the Value of Speed in Baseball

Theodore L. Turocy
Department of Economics
Texas A&M University
[back to list of papers]

Abstract

This note considers the problem of estimating the marginal products of offensive events towards a baseball team's objective of scoring runs. Regression techniques on official statistics give a positive marginal product for a stolen base attempt, which is inconsistent with the theory of mixed strategy Nash equilibrium. Augmenting the specification of the production function to include other productive qualities of footspeed restores estimates consistent with equilibrium theory.

Version history

Current version dated December 19, 2003. Available in: [pdf].
A shorter version of this paper appeared in Economics Letters.

Data

The dataset used in the regressions in this paper is available here. The Python script used to generate the data from the output of the Chadwick cwevent program used on the Retrosheet data files is here.

Related work

This paper assumes there is no autocorrelation between the residuals of the model from season to season for a particular team. This is addressed in a note "Quantifying Persistence in Run Model Errors Across Team-Seasons" [pdf] which appears in the November 2003 issue of By the Numbers.