Putting Skill as Nearly Indistinguishable from Noise: An Empirical Bayes Analysis of PGA Tour Performance

TL;DR

This paper uses an empirical Bayes approach to analyze PGA Tour performance, revealing that while tee-to-green skills vary significantly among players, putting skill is nearly indistinguishable from noise and less reliably measurable.

Contribution

It introduces an empirical Bayes method to estimate golfer skill components and assesses their significance, highlighting the limited reliability of putting performance as a skill measure.

Findings

Tee-to-green skill varies significantly among players.

Putting skill is nearly indistinguishable from noise.

The method controls false discovery rate in skill estimation.

Abstract

We revisit a foundational question in golf analytics: how important are the core components of performance--driving, approach play, and putting--in explaining success on the PGA Tour? Building on Mark Broadie's strokes gained analyses, we use an empirical Bayes approach to estimate latent golfer skill and assess statistical significance using a multiple testing procedure that controls the false discovery rate. While tee-to-green skill shows clear and substantial differences across players, putting skill is both less variable and far less reliably estimable. Indeed, putting performance appears nearly indistinguishable from noise.