Why is it "no way in hell?" They're similar in PER. If one believes Camby is a better defender, one could easily rate Camby the superior player. Camby's DRtg is certainly much better. Of course, win shares has Marc Gasol ahead. I think they're pretty similar caliber players and it doesn't seem startling for someone to rate one ahead of the other.
Well, I assumed that it was understood that Gasol was a borderline all-star, and Camby was...not. Gasol is clearly the superior offensive player, but maybe I'm underestimating Camby's defensive impact. So, just for kicks, I looked a bit deeper via 82games. The most comprehensive stats are PER, win%, on/off court (+/-), and simple rating. Simple rating Gasol +8.3 Camby +4.9 Win% Gasol 60.8% Camby 48.0% PER as C, as PF Gasol 22.1, 24.3 Camby 21.7, 20.8 Opponent PER as C, as PF Gasol 18.2, 11.9 Camby 19.3, 18.4 Clutch Win% Gasol 58.3% Camby 47.4% On/off court: off, def, total Gasol +10.4, -3.8, +14.2 Camby +3.4, -4.3, +7.7 Yes, Camby has greater impact defensively, but only slightly; in fact, Gasol's apparently a better man defender, but Camby's help defense is great. However, Gasol's impact on the offensive end is so much greater than Camby's that his presence in the lineup results in twice as large an improvement in point differential than Camby's does. I'm sorry, but the numbers are very clear--Camby's good, but Gasol is significantly better right now.
The numbers are clear if you put a lot of stock in +/- (and the related concept of win%), but I don't and few statistical analysts do in samples short of multiple seasons. If you use PER as a good proxy for production and DRtg as a good estimation of defense, as I noted before, Camby comes out ahead. Gasol PER: 19.9 Camby PER: 18.5 Gasol DRtg: 107 Camby Drtg: 100 So Gasol has a slight edge in PER but Camby has a large one on defense. This (your post and mine), to me, doesn't show that it's "clear" who's better. It suggests that they're pretty close and whom you regard as better depends on which statistical models you favour.
You're using one individual stat for overall production, and one team-performance stat for defense, but being disingenuous by ignoring the same team-performance stat for offense Gasol ORtg: 122 Camby ORtg: 111 Camby may have a large edge on defense, but by the same "statistical model" Gasol has a much larger edge on offense. DRtg is damn near the ONLY "statistical model" which favors Camby, which is not surprising considering who Gasol's teammates are. Also, you're ignoring the opponent PER stats as well, especially those which show that Gasol is vastly superior defensively at the center position (most of Camby's minutes this year have come at the PF spot).
Not being disingenuous, just using the stats I generally like (not for this particular comparison, just in general). I prefer individual stats for offense, but not for defense since defense has both an individual and a team component. If we want to be purist about "statistical models" (without mixing numbers from different systems) then ORtg/Drtg see them are pretty similar, as does PER. I don't, actually, consider Camby better. But I think they're pretty similar and I don't think your implication that Gasol is clearly and significantly superior is right (so much better, in fact, that it's ridiculous to consider Camby better).
Are you suggesting that defense has a team component, but offense does not? ORtg/DRtg has Gasol at +15, Camby at +11. That's a pretty big difference. If you see a 36% differential as "pretty similar", then I can't help you. I understand how, if you happen to focus on just the two measures you favor, that you might come to that conclusion. I prefer a more comprehensive approach. To each his own, I guess.
I'm suggesting that the vast majority of a player's offensive/rebounding contributions are captured by individual metrics. Defense has a much, much larger component that can't be captured by individual metrics. There is some offensive effect that is not captured by individual measures, but capturing that generally takes something like Adjusted +/-, which requires multiple seasons of data to be particularly reliable. Well, of course I'm going to use the ones I think are best. If by "comprehensive" you mean factoring in lesser measures, then yes, I prefer a less comprehensive approach.
