Developing a Measure for Evaluating Coaches, Part II: Exceeding Pythagorean Wins

by Jon Nichols

Today I’m going to take a different approach to evaluating coaches.  One suggested theory of evaluating head coaches is to look at how their teams outperform their expected (Pythagorean) wins.  The thinking goes that great coaches consistently excel in late-game situations and often win more than random chance would allow.  You could also make the argument that great coaches optimize the way in which they use their players, another reason they outperform their expected win totals.

What does the data say?  I rounded up each team’s actual and expected wins over the last seven years and calculated the averages for each coach.  You can find the data here:

As you can see, those theories may be wrong.  The results appear to be random, at least when you factor in common beliefs about who’s a good coach and who’s not. 

Stay tuned, as there is much more research on coaches to come.





Copyright © 2009