March 19, 2019
You can’t win them all—but a good coach can help a collegiate or professional sports team rack up points on a regular basis. Those are the findings of a University of Chicago study on the importance of leadership in athletics.
Scholars at the university’s Harris School of Public Policy analyzed hundreds of seasons of data—including wins and losses, and sports scores and statistics— and found that coaches account for 20 % to 30% of the variation in team outcomes.
To reach their findings, Professor Christopher Berry and Associate Professor Anthony Fowler looked at the impact of coaching in Major League Baseball, the National Football League, the National Basketball Association, the National Hockey League, college football, and college basketball.
“Coaches are often credited or blamed for their team’s success or failure, and are compensated as if they are among the most important assets a franchise possesses,” said Berry. “We find that coaches do, in fact, matter—and suggestions that coaches are interchangeable, which has been the dominant view in the sports analytics community, are not true. In every sport we studied, we found that coaches impact variables that contribute to a higher winning percentage.”
The study came up with a number of findings, which Berry and Fowler presented March 1 at the MIT Sloan Sports Analytics Conference in Boston—among them:
- MLB managers affect runs scored, runs allowed, run differential, and victories. They have greater impact on runs allowed than on runs scored.
- NFL coaches affect points allowed and the point margin. They significantly affect the number of fumbles and penalties a team commits.
- Coaches matter more in college football than in the pros. They significantly affect points scored, points allowed, point differential, and victories.
- Coaching is highly significant in both NBA and Division I college basketball outcomes—influencing points scored, points allowed, point differential and victories.
- NHL coaches matter, although they matter much more for goals allowed than for goals scored.
“Although virtually every aspect of player performance has been examined since the recent emergence of sports analytics, we wanted to bring the same level of rigor to coaches as there is for everyone else on the field at a major sporting event,” Fowler said.
The study was conducted with a method called randomization inference for leadership effects, which accounts for player quality and strength of schedule. Berry and Fowler first created the approach to estimate the effects of political leaders on various economic and policy outcomes. The method holds promise for additional research to assess the impact of individual coaches, as well as better understand why and how coaches matter.
Research contact: firstname.lastname@example.org