There has been much debate on which player evaluation in the NBA is “better.” Yesterday one of the participants in this discussion – Malcolm Gladwell – weighed in with some much needed perspective.
In a post entitled “The Perfect and the Good,” Gladwell notes the role algorithms play in decision-making. Gladwell has written about algorithms for movie executives, medical doctors, and of course, coaches and general managers in basketball. As Gladwell notes, “…. no decision rule or algorithm or prediction system is ever perfect. The test of these kinds of decision aids is simply whether–in most cases for most people–they improve the quality of decision-making. They can’t be perfect. But they can be good.”
Algorithms cannot replace thinking. To understand player productivity one must ascertain both how and why a player offered the performance we observed. As we argue in The Wages of Wins and several times in this forum, Wins Produced and Win Score are measures of how productive the player has been. These measures do not tell you why that player was productive.
The problem we note with decision-making in the NBA is that the how question has been difficult for everyone to answer with conventional methods. Studies of player salaries and the coaches voting for the All-Rookie team indicate that scoring is over-valued. We can also see this same bias when we look at metrics like NBA Efficiency and Hollinger’s Player Efficiency Rating. Wins Produced and Win Score values each statistic in terms of their impact on wins. Consequently, the bias we see in other metrics is corrected.
But these metrics alone cannot be relied upon by themselves to make decisions. Knowing how productive the player has been in the past is only part of the story. To make great decisions, one also has to know why. For that question we have done a bit or research that can help. But the why question has not yet been – and may never be — completely answered by analysis of the statistics. And hence the need for a bit of thinking, the crucial element of decision-making no computer can replace.