A few days ago, Larry Brown declared that for basketball, advanced stats “don’t work”. But when you read what he has to say, you get the sense that Brown thinks that some advanced stats don’t work. In other words, I think there are some “advanced stats” that Brown might like.
To illustrate, consider this quote from the article:
…if you rebound, if you defend, you share the ball, you have a better opportunity to win…
What explains wins in the NBA for Brown? Rebounds, shooting efficiency, and defense.
Here are Dean Oliver’s “Four Factors” (an advanced stat):
- Shooting efficiency from the field
- Free throws (which is really a part of shooting efficiency in general)
Here are the factors that drive Wins Produced (another advanced stat):
- Getting possession of the ball (rebounds and opponent’s turnovers)
- Keeping possession of the ball (offensive rebounds and avoiding turnovers)
- Turning possessions into points (shooting efficiency)
Brown often talks about “playing the right way”; it appears that much of Brown’s “playing the right way” is captured in Oliver’s Four Factors and in Wins Produced. So Brown should agree with some advanced stats. Of course, in the article Brown doesn’t say which “advanced stats” he was shown, but my sense is that he was being shown something other than the Four Factors or Wins Produced. What might he have been shown instead?
In the past, I have heard coaches comment on being shown “line-up analysis”, which shows how a team does with certain players on the floor. For example, read what Memphis Grizzlies’ coach Lionel Hollins had to say in January:
We had a guy a few years ago that was sending me emails about different lineup combinations, and he was saying, ‘this lineup should be on the court a lot more because they’re the most effective.’ So, then you coach that lineup and keep them on the floor for 40 minutes. I’m going to stay with the lineups that I have on the floor.
Or perhaps Brown was shown plus-minus analysis, which is supposed to show the impact of individuals on team outcomes (even if it doesn’t work very well). But lineup and plus-minus analysis are “black box” stats — they aren’t very useful for decision-making because they don’t tell a coach why a certain player is “good” or “bad”.
Going back to the Hollins quote, Hollins also sung the praises of Rudy Gay, who Memphis traded away to the Toronto Raptors about twenty days later. Given the success of the Grizzlies since that trade — something that we expected, even though it was contrary to popular opinion at the time — it’s clear that Hollins was incorrect in his assessment of Gay. Could better stats have led Hollins to see this before the trade?
But presenting a “black box” stat – such as line-up analysis or plus-minus – is simply not going to convince someone not trained in stats (of course it shouldn’t convince someone who is trained in stats either!). A better approach would be to try to explain why teams win games. Again, coaches seem to know that teams win because of rebounds, turnovers, and shooting efficiency. Showing that Gay doesn’t help a team in these areas might have convinced Hollins that all would be okay if Gay left town.
And the same approach might also work with Brown. Again, Brown already knows why his teams are good or bad. If he was shown stats that allowed him to connect what he knows about winning basketball teams to individual players, he might be more enthusiastic about “advanced stats”.
– DJ (with some help from Devin)