One of the stories told in Stumbling on Wins is that draft position is a poor predictor of an NBA player’s future per-minute productivity. A player’s college numbers are a better predictor, although the link between college and pro performance is not nearly as good as we would like.
Up until a few weeks ago, though, all we had were those college numbers. But with the preseason now complete, we now have more numbers (thanks to Doug Steele) to use in our evaluation of rookies.
Back in 2008, a model was offered connecting a player’s regular season per-minute Win Score (WSMIN) to his per-minute Win Score in college and the preseason. This model has now been updated. Utilizing 80 observations from 2007, 2008, and 2009; the aforementioned model was re-estimated. The results indicate that 61% of a rookie’s regular season WSMIN is explained by what he did in his last year in college and the prior preseason. This isn’t perfect. But the explanatory power is better than what you see if you consider college and preseason performance alone.
And what do you see when you look at both college and preseason performance? There were 19 rookies who played college basketball and at least 100 minutes in the past preseason (at least, I was able to find data on 19 rookies). The following table presented what these rookies did and then uses this information to predict WSMIN and WP48 [Wins Produced per 48 minutes] for the 2010-11 season [as noted in the past, one can use WSMIN to project WP48].
And here is what we learn from this analysis:
Again, we need to emphasize that these are only projections. And all that is being projected is the first season. Back in 2007, Kevin Durant’s rookie season was projected with these two data points. That projection indicted that Durant would struggle as a rookie. Although Durant did not play well as a rookie (so the projection was correct in Durant’s case), he ultimately did get much, much better. A similar story could play out for any of these players (then again, it might not).
All we can see is that using data from college and the preseason does tell us more than we knew with just data from the NCAA. But these two data sets don’t tell us everything (and in a few days, we will start to see what we haven’t seen so far).
- DJ
