Nylon Calculus: A quick look at deflections

The NBA stats page releases the totals and per game calculations for Hustle stats, a somewhat helpful collection of statistics that have long been tracked by scouts and coaching staffs. Totals and per game values aren’t incredibly useful but it’s a simple calculation to convert that over to a rate statistic, which, with sufficient sample size, will give us a better idea of who is good at getting deflections and who is not. For this, I converted over to per forty minutes for deflections.

The leaders through November 16 (of players with a minimum of 200 minutes played) on a per 40 minute basis are below, with Thabo Sefolosha holding a significant lead.

The top ten leaders so far generally line up with a reputation as a good defender, including guys like Chris Paul, Kawhi Leonard, and Draymond Green. Though that’s not universally true (hi Rondo). Like steals, we aren’t measuring the number of gambles a player might make to rack up those deflections, so just like tracking points scored without taking into account shots used it is not a complete picture of defense.

On the other side of the spectrum here are the twenty players with the lowest rate of deflections. There is a noticeable position bias for centers and big men to record relatively fewer assists than wings and guards, which comes out the same here.

I also found that, unsurprisingly, deflections are fairly well correlated with steals, with a R^2 of .61 for players with at least 200 minutes played, a number we can expect to increase modestly as the season wears on.

Deflections are also more common than steals with the average player having of 2.5 deflections per 40 minutes, but only 1.3 steals per 40. So, even if the information is mostly similar to the established steals metric, potentially, that indicates a more quickly stabilizing statistic in deflections compared to the more rare steal.