Coaches tend to absorb the blame when a team loses and get little of the credit when it wins. A common view is that coaches are more or less interchangeable, and that the real drivers of a season are the players’ skill, chemistry, and health. Christopher R. Berry and Anthony Fowler of the University of Chicago set out to test that view with data, in a paper presented at the 2019 MIT Sloan Sports Analytics Conference.
The method
The difficulty with measuring a coach is that good coaches tend to end up with good players, so a simple comparison of win rates says little. Berry and Fowler built a method they call Randomization Inference for Leader Effects, or RIFLE. Instead of estimating the skill of each individual coach, it asks a narrower question: do coaches matter in the aggregate? The method shuffles coaching assignments many times to see how much a coaching change would be expected to move a team’s results if coaches had no effect at all, and compares that to what actually happened.
What they found
The study covered the NBA, MLB, NHL, NFL, college football, and college basketball. Across all of them, coaches were found to have a measurable effect on outcomes, though the size differed by sport and by statistic.
Basketball stood out. In the NBA, coaches explained about 30 percent of the variation in points scored and points allowed, which carried through to point margin and to wins. The paper’s table for NBA coaches from 1947 to 2017 lists an r-squared of .396 for points scored and .415 for victories, with p-values at or below .001 for every outcome. College basketball coaches from 1938 to 2017 showed a similar pattern, and one extra result: coaches also affected the total number of points scored in a game, which the authors attribute to a preference for fast or slow pace.
What to make of it
The paper does not rank coaches or say who is best. Its claim is more modest and more defensible: coaching, as a category, is not noise. The team-building factors people usually point to still matter, but a share of a team’s results is attributable to who is standing on the sideline. The practical takeaway from the post was simple. If we hold coaches accountable when a team plays badly, the same numbers say they have earned some of the praise when it plays well.