ulosturedge
07-19-2010, 11:47 PM
NBA HD: Are standing reach/wingspan overvalued?
Written by Tom Haberstroh on Thursday, July 1, 2010
Last week, I looked at the relationship between max vertical height and rebounding. In the comments section, the brilliant readers offered up suggestions to expand on this topic. Let’s see what we can do.
One of the quibbles with the piece was my choice to look at max vertical height instead of say, weight or standing reach. Well, thanks to Draft Express, I can run those as well. Keep in mind, the measurements were taken at the draft and likely have changed slightly (weight).
Let’s start from the ground up. If we run a multivariate regression that predicts total rebound percentage in the NBA with height and weight as our predictor variables, what do we find? Those two pieces of information do a pretty good job of predicting rebounding performance– the adjusted R2 = .6584 and both factors were statistically significant at a p <.05 level). DeJuan Blair may stand six-feet seven-inches tall but he packs about 270 pounds, so he’s a refrigerator down low. DeMar Derozan, by comparison, is as tall as Blair but about 60 pounds lighter and grabs about a third of Blair’s share of rebounds. On the flipside, height gives players an obvious advantage as well (R = .756). Interestingly enough, weight had the strongest correlation of any tested variable (R = .772).
This makes sense and shouldn’t surprise anyone. It’s tough to find a short stick grabbing boards left and right. But what if we add another variable to the mix? Let’s replace height as a predictor variable with standing reach instead, as some have called for. The logic being, that necks don’t grab rebounds– hands do. So, does standing reach and weight predict rebounding rates better than raw height and weight? Turns out it doesn’t, at least not by this method. The adjusted R2 slides down to .647 which is lower than the predictive strength of raw height and weight. The correlation between rebounding and height is .756 whereas with standing reach it is .739.
So, the next thing I wanted to do was see if wingspan is a significant predictor of rebounding once we control for height and weight. Surprisingly, it isn’t at the p < .05 level (p=0.52) while height and weight remained significant. The adjusted R2 was .6576 which tells us that the model’s goodness of fit didn’t improve after adding wingspan as a predictor variable. Hmm.
I’m having some trouble explaining that finding. You’d think that wingspan would definitely help to explain a player’s rebounding in the NBA but that doesn’t seem to be the case. Sure, it certainly helps to have longer arms but on the whole it doesn’t seem to be of great tangible benefit in rebounding. Need examples? Javale McGee has the tallest reach of any of the 241 players in the study (he can actually come within 7 inches of touching the rim standing up) and yet he’s merely an average rebounding center. He and Chris Kaman share the same height but Kaman’s wingspan is 6 inches shorter than McGee’s. Who’s the better rebounder? The heavier Kaman. You’d think McGee would put that 7’6″ wingspan to better use on the boards.
Oh, and Patrick O’Bryant.
http://www.hardwoodparoxysm.com/2010/07/01/nba-hd-are-standing-reachwingspan-overvalued/
I stumbled on this article when I was trying to look up Alonzo Gee's wingspan and weather it might help him in guarding small forwards. Anyways the boards have been slow, and I thought this was kind of interesting. What do you guys think of the value between Height, Standing Reach/Wingspan, and weight as it pertains to rebounding?
Written by Tom Haberstroh on Thursday, July 1, 2010
Last week, I looked at the relationship between max vertical height and rebounding. In the comments section, the brilliant readers offered up suggestions to expand on this topic. Let’s see what we can do.
One of the quibbles with the piece was my choice to look at max vertical height instead of say, weight or standing reach. Well, thanks to Draft Express, I can run those as well. Keep in mind, the measurements were taken at the draft and likely have changed slightly (weight).
Let’s start from the ground up. If we run a multivariate regression that predicts total rebound percentage in the NBA with height and weight as our predictor variables, what do we find? Those two pieces of information do a pretty good job of predicting rebounding performance– the adjusted R2 = .6584 and both factors were statistically significant at a p <.05 level). DeJuan Blair may stand six-feet seven-inches tall but he packs about 270 pounds, so he’s a refrigerator down low. DeMar Derozan, by comparison, is as tall as Blair but about 60 pounds lighter and grabs about a third of Blair’s share of rebounds. On the flipside, height gives players an obvious advantage as well (R = .756). Interestingly enough, weight had the strongest correlation of any tested variable (R = .772).
This makes sense and shouldn’t surprise anyone. It’s tough to find a short stick grabbing boards left and right. But what if we add another variable to the mix? Let’s replace height as a predictor variable with standing reach instead, as some have called for. The logic being, that necks don’t grab rebounds– hands do. So, does standing reach and weight predict rebounding rates better than raw height and weight? Turns out it doesn’t, at least not by this method. The adjusted R2 slides down to .647 which is lower than the predictive strength of raw height and weight. The correlation between rebounding and height is .756 whereas with standing reach it is .739.
So, the next thing I wanted to do was see if wingspan is a significant predictor of rebounding once we control for height and weight. Surprisingly, it isn’t at the p < .05 level (p=0.52) while height and weight remained significant. The adjusted R2 was .6576 which tells us that the model’s goodness of fit didn’t improve after adding wingspan as a predictor variable. Hmm.
I’m having some trouble explaining that finding. You’d think that wingspan would definitely help to explain a player’s rebounding in the NBA but that doesn’t seem to be the case. Sure, it certainly helps to have longer arms but on the whole it doesn’t seem to be of great tangible benefit in rebounding. Need examples? Javale McGee has the tallest reach of any of the 241 players in the study (he can actually come within 7 inches of touching the rim standing up) and yet he’s merely an average rebounding center. He and Chris Kaman share the same height but Kaman’s wingspan is 6 inches shorter than McGee’s. Who’s the better rebounder? The heavier Kaman. You’d think McGee would put that 7’6″ wingspan to better use on the boards.
Oh, and Patrick O’Bryant.
http://www.hardwoodparoxysm.com/2010/07/01/nba-hd-are-standing-reachwingspan-overvalued/
I stumbled on this article when I was trying to look up Alonzo Gee's wingspan and weather it might help him in guarding small forwards. Anyways the boards have been slow, and I thought this was kind of interesting. What do you guys think of the value between Height, Standing Reach/Wingspan, and weight as it pertains to rebounding?