I have been taking a look at the stats released by the NBA using their new SportVu cameras, including Drives to the basket, Pull Up shots and Catch and Shoot jumpers. All of them have given some interesting insights into the game. This time I am looking at their rebounding stats.
The new wrinkles added are that SportVu tracks how many rebound opportunities a player had, as defined as a rebound within 3.5 feet, what percentage they end up with and how many of those were contested by the opposing team. All interesting stuff.
But the question is how useful is it and in what ways. To kick the tires a bit, I did two things, I aggregated all of the SportVu rebounding data from NBA.com as of Dec, 11th, 2013 by team and then linked it to the NBA.com advanced stat data for both players and teams.
In theory, having a higher percentage of a player's rebounds coming as contested should be valuable indicating that the rebounds grabbed could have ended up in the opponent's hands. On the other hand, it could be that players with better positioning or fearsome reputations dissuade opponents from contesting.
Also, grabbing a higher percentage of rebound opportunities could be valuable, in terms of maximizing total rebounds.
So I tested the aggregated SportVu rebound data for each team against the team rebounding percentages as well as looking at correlations within the SportVu data. Here are the findings:
- A slight negative correlation for the percent of team rebounds contested and the percentage of opportunities grabbed.
- A small positive correlation with Contested Percentages and total rebound %
- No correlation with defensive rebounds and contested rebounds.
- A significant (statistically and practically) correlation between number of contested rebounds per game and offensive rebounding percentage and correlation between the percent of team rebounds contested and team offensive rebound percentage. The R2 for all 30 teams being .29.
This basically confirms what we knew already. Many of the defensive rebounds any individual player grabs could have been taken by one of his teammates. In fact, Dave Berri eventually changed the Wins Score and Wins Produced formulas when convinced of the evidence to devalue defensive rebounds in player evaluation.
Then I looked at the SportVu rebounding data against the individual rebounding percentages for players. I found a largely similar pattern. Percent contested rebounding percentage found in the SportVu data have modest with a player's defensive rebounds, but they are more closely related to offensive rebounding percentages.
Below are the results of a simple regression for offensive rebounds for players with 4 or more rebounding opportunities per 48 minutes, showing a R2 of .499.
Using the same filter the R2 of .17 for Contested rebound percentage and defensive rebounds:
What the data released by the NBA does not show, is the percentage of contested opportunities a player is able to pull down. Some players may have fewer contested rebounds because the coaching scheme does not emphasize offensive rebounds. Alternatively knowing how often a player is successful in contested rebound situations may help a coach decide which players to send on offensive rebounds and the ones to drop back.
Comments