Summarizing Our Thoughts on the NBA

Posted on May 24, 2006 by


Malcolm Gladwell wrote a wonderful review of our analysis of the NBA in this week’s New Yorker.  Given Gladwell’s prodigious talent as a writer, it is not surprising that he could tell our story so well. 

Having reviewed some of the reaction to his story, though, I thought it would be useful to summarize what our research indicates about decision-making in the NBA.  Here is a point-by-point summary of the story we are telling about the NBA.

  • Payroll does not explain much of wins in the NBA, MLB, or NFL.  Specifically, payroll only explains 12% of the variation in wins in the NBA.  In baseball explanatory power is 18% while in the NFL it is below 5%. 
  • We think the low explanatory power of payroll in baseball and football can at least partially be explained by the relative inconsistency of performance in these sports.  As we note in our book, across time in baseball and football we see fairly wide variations in player productivity.  After all, who expected the Detroit Tigers to be so good this year?
  • Relative to these sports, though, performance in the NBA is more consistent.  So why is payroll still unable to explain much of wins?
  • We think the answer lies in how players are evaluated in the NBA.  For more than two decades economists have looked at the link between player salary and various performance statistics.  Scoring totals are the only player statistic that consistently explains player pay.  Shooting efficiency, rebounds, steals, and turnovers do not consistently offer much explanatory power.  We updated these studies in our book.  Our story, though, was essentially the same.  Scoring totals are the one statistic that matters most in determining player pay. 
  • How much players are paid is not the only decision economists have examined. Ha Hoang and Dan Rascher published a study in Industrial Relations in 1999.  The Hoang and Rascher study looked at the factors that caused a player to be cut from an NBA roster.  The only player statistic these researchers found to matter was scoring.  All other player statistics did not matter.
  • We have looked at the coaches voting for the All-Rookie team and the factors that impact where a college player is drafted.  What matters most? Again, scoring matters more than factors associated with getting possession of the ball (i.e. rebounds, turnovers, and steals).
  • Wins in the NBA, though, are not just about scoring.  Possession factors have a large impact on the outcomes we observe in the NBA. When you look at all the statistics the NBA tracks you find that with these you can explain 95% of the variation in wins.  And when you look at all these statistics you find that you can create a very accurate estimate of the wins each player produces.

From all this what do we conclude? Conventional wisdom in basketball is incorrect.  Players who only score are not as valuable as people think.  Players who do not score much — like Ben Wallace and Dennis Rodman – have a bigger impact on team wins than people seem to think.

Does this fit what many people believe about the NBA?  No, but as academic research often indicates, what people believe does not always match what the data says.

– DJ