Masterstrokes: Statistics Research in the Golf World


  • Researchers used the location data of millions of shots taken in golf tournaments to create a new stat called ‘strokes gained-putting’
  • The stat is based on a model that accounts for the difficulty of a putting green and the talent of the other golfers on the course

At the conclusion of the Masters Tournament this month, the winning golfer will slip on the famous green jacket, sealing his name in the sport’s history books.

But a McCombs School of Business researcher has already left his mark on the world of professional golf. Since 2011, the PGA Tour has been using a putting statistic developed in part by Doug Fearing, assistant professor in the Information, Risk, and Operations Management department. That stat, known as strokes gained-putting, “has been hailed by players and pundits alike as the most accurate, meaningful way to present player putting efficiency,” according to the tour website.

An earlier putting statistic, known as putts per round, looked at the average number of putts taken by a golfer over 18 holes. That stat, however, could be skewed by a player who initially missed the green, but was subsequently able to chip the ball close to the hole. The implementation of the PGA’s ShotLink system to track individual shot coordinates, followed by professor and student access to that data starting in 2005, meant golf was able to receive a much-needed statistical makeover.

For a 2010 paper, Fearing — at the time a Massachusetts Institute of Technology Ph.D. student — and co-authors Stephen Graves and Jason Acimovic considered the tour’s ShotLink data, which provided precise location data for millions of golf shots taken. The researchers were able to see the sequence of shots each golfer had taken at each tour hole over a timespan of six years. From that sizable data set, Fearing and his fellow researchers focused on the putting green. 

They created their own approach to modeling putting. “We built it as a memory-less statistical process where, if you putt to a certain distance, then it doesn’t matter where you came from, because your expected putts to go will be the same,” Fearing says. Their model takes into account both the difficulty of a given green, as well as the putting talent of the other golfers on that course.

“A few other researchers did work in this area around the same time, and our claim-to-fame is that the official name for the statistic came from our paper,” Fearing says.