Re: 2012 NBA draft ratings
Posted: Wed Jul 04, 2012 5:02 am
Fruit salad...yummy yummy. A few more words on the Shaq vs. KG thing, trying to keep the apples, separate from the oranges, and grapes, etc.:
The core of the argument that it is by no means clear that Shaq is KG's better is that the belief in Shaq's superiority is (almost) entirely based upon offensive bias - one which informed consumers really should not fall victim to. A point prevented is as valuable as a point scored (to a first and very good approximation) and on this basis (via RAPM) 30 year old to 35 year old KG was clearly and strictly superior to 30 year old to 35 year old Shaq. And I don't think anyone really wants to disagree with this. Right?
But what about the comparative careers in their primes (call this from ages 26 to 29) as well as those of their youth?
First, a brief word on the latter. It is true that I gave Shaq's early years short shrift. This too was a natural experiment, and he entered the Association as a really Big Beast (and one who initially was able to play long minutes without getting injured). Comparatively, KG entered one year younger as a very talented, but very skinny kid. And even though in his 20 year old, sophomore campaign (when he was playing the same approximate mpg as 20 year old rookie Shaq) the Timberwolves improved by 14 games, this is less of an apparent impact than Shaq had. And despite these remaining the statistical dark ages, in terms of properly measuring defensive contributions, it has to be correct to believe that KG was not so defensively precocious as to be Shaq's equal when first entering the NBA.
So Shaq gets the nod between 20 and 25 whereas KG wins the rounds post-30, but what about their respective primes? Again, of course, there is incomplete information for comparison because of the lack of good information about defense. But what do we know?
1) Daniel can comment if he wishes, but I read his work to say that ASPM is decidedly not a good predictor of Kevin Garnett's RAPM between 2003-2011, for what includes his prime years, and (more than) all of the discrepancy shows up on the defensive ledger (http://godismyjudgeok.com/DStats/aspm-and-vorp/). Conversely ASPM is a good predictor of Shaq's RAPM, though perhaps it wouldn't have been in his prime.
2) Win Shares is a coherent, well thought out rating, though necessarily limited in its defensive measure. What does it say about the comparative ages of 26 through 29? By my calculation, if you take a straight sum of WS over these four years then divide this by total minutes played, you get a Shaq superiority of 0.000314 WS/minute. And what does this imply for a representative year, in terms of extra wins? Well, if you average the number of minutes for Shaq aged 27 to 29 (as Shaq's 26 year old year was the shortened 1998-99 year) you get a representative year of 2,836 minutes. Multiplying this by 0.000314 you get 0.9 extra Shaq wins per year, compared to KG, in their comparative primes.
But. Wait.
We know that WS has defensive shortcomings as a measure, and we know, via RAPM, that KG was a defensive standout in those same years. So, to believe that KG was inferior to Shaq in their prime is to believe that KG's undermeasured defensive contributions weren't worth 0.9 wins per year were he to have played Shaq minutes. Seems rather implausible.
But there is another important point: he didn't play Shaq minutes. He played more, and more games too. And that is another bit of bias undergirding the "Shaq is better" argument. The problem of Big Beasts is that their bodies face a lot more wear and tear, and they break down more. Over the common 14 years of their careers, between ages of 20 to 34 and excluding the shortened 1998-99 year, KG played 9.7 more games per year! And in the prime years in question, ages 27 to 29, KG played, on average, 267 more minutes per season - what implies (via WS numbers) 1.4 extra "KG wins" per year, from which you subtract the wins of an average replacement player, and then you basically get a wash.
But, again, there is the uncounted defense contributions of KG identified by RAPM, which if it is greater than nothing, must (?) tip things in KG's favor.
So, how might this story be better told? Shaq enters the NBA as the Big Offensive Beast, the incredible force that is firmly anchored in all our consciousnesses. KG, by contrast, enters more quietly and develops a more well-rounded game. Then by their prime years, it seems likely that KG's immeasurable defensive contributions have, if not tipped the balance in his favor, at least caught him up to Shaq's level. And then for the years that followed, things are much better understood: Big Beasts Break Down. Wiry, wily fitness fanatics...not so much. And then they keep on going, such that KG's legend is not yet ready to be fully told.
If I were AD and had a choice, KG's career please. Thank you very much.
The core of the argument that it is by no means clear that Shaq is KG's better is that the belief in Shaq's superiority is (almost) entirely based upon offensive bias - one which informed consumers really should not fall victim to. A point prevented is as valuable as a point scored (to a first and very good approximation) and on this basis (via RAPM) 30 year old to 35 year old KG was clearly and strictly superior to 30 year old to 35 year old Shaq. And I don't think anyone really wants to disagree with this. Right?
But what about the comparative careers in their primes (call this from ages 26 to 29) as well as those of their youth?
First, a brief word on the latter. It is true that I gave Shaq's early years short shrift. This too was a natural experiment, and he entered the Association as a really Big Beast (and one who initially was able to play long minutes without getting injured). Comparatively, KG entered one year younger as a very talented, but very skinny kid. And even though in his 20 year old, sophomore campaign (when he was playing the same approximate mpg as 20 year old rookie Shaq) the Timberwolves improved by 14 games, this is less of an apparent impact than Shaq had. And despite these remaining the statistical dark ages, in terms of properly measuring defensive contributions, it has to be correct to believe that KG was not so defensively precocious as to be Shaq's equal when first entering the NBA.
So Shaq gets the nod between 20 and 25 whereas KG wins the rounds post-30, but what about their respective primes? Again, of course, there is incomplete information for comparison because of the lack of good information about defense. But what do we know?
1) Daniel can comment if he wishes, but I read his work to say that ASPM is decidedly not a good predictor of Kevin Garnett's RAPM between 2003-2011, for what includes his prime years, and (more than) all of the discrepancy shows up on the defensive ledger (http://godismyjudgeok.com/DStats/aspm-and-vorp/). Conversely ASPM is a good predictor of Shaq's RAPM, though perhaps it wouldn't have been in his prime.
2) Win Shares is a coherent, well thought out rating, though necessarily limited in its defensive measure. What does it say about the comparative ages of 26 through 29? By my calculation, if you take a straight sum of WS over these four years then divide this by total minutes played, you get a Shaq superiority of 0.000314 WS/minute. And what does this imply for a representative year, in terms of extra wins? Well, if you average the number of minutes for Shaq aged 27 to 29 (as Shaq's 26 year old year was the shortened 1998-99 year) you get a representative year of 2,836 minutes. Multiplying this by 0.000314 you get 0.9 extra Shaq wins per year, compared to KG, in their comparative primes.
But. Wait.
We know that WS has defensive shortcomings as a measure, and we know, via RAPM, that KG was a defensive standout in those same years. So, to believe that KG was inferior to Shaq in their prime is to believe that KG's undermeasured defensive contributions weren't worth 0.9 wins per year were he to have played Shaq minutes. Seems rather implausible.
But there is another important point: he didn't play Shaq minutes. He played more, and more games too. And that is another bit of bias undergirding the "Shaq is better" argument. The problem of Big Beasts is that their bodies face a lot more wear and tear, and they break down more. Over the common 14 years of their careers, between ages of 20 to 34 and excluding the shortened 1998-99 year, KG played 9.7 more games per year! And in the prime years in question, ages 27 to 29, KG played, on average, 267 more minutes per season - what implies (via WS numbers) 1.4 extra "KG wins" per year, from which you subtract the wins of an average replacement player, and then you basically get a wash.
But, again, there is the uncounted defense contributions of KG identified by RAPM, which if it is greater than nothing, must (?) tip things in KG's favor.
So, how might this story be better told? Shaq enters the NBA as the Big Offensive Beast, the incredible force that is firmly anchored in all our consciousnesses. KG, by contrast, enters more quietly and develops a more well-rounded game. Then by their prime years, it seems likely that KG's immeasurable defensive contributions have, if not tipped the balance in his favor, at least caught him up to Shaq's level. And then for the years that followed, things are much better understood: Big Beasts Break Down. Wiry, wily fitness fanatics...not so much. And then they keep on going, such that KG's legend is not yet ready to be fully told.
If I were AD and had a choice, KG's career please. Thank you very much.