With the current debate raging within Rockets fandom about Howard vs Asik, I figured it might be of some slight interest to recap a paper I had read from this year's MIT Sloan conference. While I know that most of the debate around these parts is about the difference between the offensive impacts of both, this paper presents a refreshing view of the defensive side of the story too. It's a take on measuring defensive effectiveness of interior defenders. Its quite a well written paper authored by Mr Kirk Goldberry from Harvard and Mr Eric Weiss from Sports Aptitude. [Sorry I havent properly cited the paper]
For those who want a quick explanatory summary, please allow me to try and present my understanding of the core concepts in it.
Firstly the paper rightly claims that offensive impact is easier to measure via stats but defensive impact needs us to dig deeper than what just defensive rebounds/steals/blocks and other plain stats tell us. I guess we all knew that.
The second point is that we can divide shots taken in the NBA into 3 zones(long range/mid range/close range). Again we Rocket fans have long known the relative strengths of the long and close range shots while decrying the inefficiencies of the mid range jumpers. It then suggests that offensive players should be analysed by looking at not just their accuracy across these 3 ranges but also how they distribute their shot attempts across these 3 ranges. Again, this is a concept we hang our hats upon many a time when we talk of say Harden vs Josh Smith (The paper tends to use Josh Smith at various points to illustrate inefficient shooting, so there's that for JSmoove proponents this offseason)
Average NBA offensive performance
Close / Mid / Long
35% / 41% / 24% -- Shot distribution of avg NBA player [35% from close, 41% from mid, 24%from long]
53% / 39% / 36% - Accuracy of avg NBA player from close, mid and long ranges
KD on the other hand is
Shot Distribtion 27/46/27 Accuracy 65/44/37
And JSmoove is
Shot Distribtion 45/43/12 Accuracy 61/37/30 [Surprisingly drives to the basket more than avg but as we have all seen, takes too many mid range jumpers and is bad at both mid and long range shots]
Anywhoo, this was just to set up the next part about how important are interior defenses, since as we can see the close range shots are pretty high percentage shots and also (apparently) the source of a higher number of offensive rebounds.
So now that we have broken down how offenses operate, a good inside defender can do one of two things. He can A)force the shooters to take less number of close range shots or B) make the shoter much mroe inefficient on the shots he does take in the close range. This forms the basis of our first comparison between messrs Howard and Asik. However, for this kind of comparison we should only take the shots which occur when such defender is close to the basket. This is of course a proxy for when the action is close to the rim rather than say on a fast break or when the offensive action is more perimeter oriented. Also not fair to judge interior defenders when they arent even there!
On average, in such situations players tend to shoot 58% of their shots in the close range, making these shots at a clip of 49.7%
Firstly looking at case A Howard comes in first in the league. This means that when Howard is present within 5 feet of the basket, opposing players simply choose not to go to the basket(only 48% of the time are close range shots taken). The percentage of shots they take distinctly shifts towards the mid range and long range. This is of course a testament to his reputation and intimidation as much as to his ability. The Big O ranks number 33 and doesnt really alter the shot distribution much in such situations. People stll take 58% of their shots in the close range against him.
Now looking at case B, Howard comes in 10th by reducing the fg% in close range to 45% whereas The Big O comes in slightly worse than average at 51%(number 33)
Now what follows is a little critique of my own. Of course the problem with the study is that it doesnt take into account how the defense of your teammates affects your own performance on defense. So it doesnt normalise across the fct that Omer is probably helping off his man a lot more and trying to make up for others defensive lapses as well. Also that O is Houstons only big man on the floor usually. Plus the fact remains that these numbers are nto normalised per 36 minutes or anything and the data is only gathered from Arenas with the tracking camera thingies. But overall it is still a fresh way to look at defense and I thought it would be cool to share it with you guys. Check out the paper its got other cool stuff too.
Hope you enjoyed my modest attempt at paraphrasing this paper. I just found it super interesting to know what kind of things they talk about at Sloan. They dont seem to be as geeky as we think but they do think quite deeply about the game indeed.