And we're not talking about NBA history are we? We're talking about Ben's 2003 season compared to Hakeem's 1993 season.

Anyway, it's almost like you didn't comprehend a single thing I was explaining by my earlier statement in the first place. Hakeem's 104 ORtg (on around 26% of his team's offense) is equivalent to someone putting up a 99 ORtg in 2003. He would still have a higher ORtg (and the difference isn't exactly by leaps and bounds, either) than Wallace given the usage/efficiency tradeoff, but once again the metric is measuring value. Two players can post identical ORtgs and % poss rates in the same statistical league environment, but if Player A played in more games where points are harder to come by, his production is going to be more important to his team winning their games, plain and simple.

You can be impressed with raw offense and per game numbers all you want, but without putting them into the right context, it's like buying a computer because you went by the clock speed of its CPU. Per game numbers/PER/ORtg + % poss rate gives you raw offense, WS gives you the CONTEXT to put those numbers into.

Regarding pace, my point was that high usage "star" players see diminishing returns from increased pace (or minutes for that matter), because they are already producing at peak level. Conversely, role-payers have more to gain from a faster pace or more minutes. I think it's telling that fast-paced teams are associated with a balanced attack, rather than a single dominant scorer; think Russell's Celts, Magic's Lakers, Webber's Kings, and Nash's Suns.

You're comparing players from different seasons where the league environment and defenses change from season to season. It makes comparing their per game numbers absurd.

"Just as Cooz unfairly benefitted with more DWS once Russell arrived, Hakeem's OWS was unfairly hurt by being on a bad offensive team (HOU was 21st out of 27 in ORtg in 1990)."

First of all, it wasn't "unfair" for the reasons I already mentioned. DRtg is also based on the assumption that the individual faces 1/5th of the opposing team's offense while on the floor, so DWS is going to tie alot of your value to how the rest of your team performs defensively (which reflects how defense works on the floor anyway).

Second, unlike defense, offense is much less team-oriented, and this is also reflected in ORtg (which is the basis for OWS). When OWS is calculated, it doesn't even look at what your team does - it compares your performance to the rest of the league on offense.

"Let's suppose a player sees his "real" statistical production RISE when his team starts playing at a faster pace, but his pace-adjusted stats show a DROP in individual production. Which measure is more accurate? If he's a high usage "star", perhaps his marginal utility can't really be much improved upon by the faster pace, simply because he was already near his peak of possible productivity. Meanwhile his lesser teammates disproportionately gain the rewards of a faster pace."

You need to explain the last sentence in your paragraph. But with what you're talking about, you still don't understand that the key here is measuring VALUE (which is the same point I have been making about the list that you put up with Hakeem). Those points scored in an offense where there are more possessions in a given time span don't have as much value as points in a slower paced offense.

In the example of Cousy, he went from scoring in a slower-paced offense to scoring in a faster-paced one, in a league where the overall pace also was faster. So those per game numbers are misleading despite the increase - the actual value of the point, assist, etc. (which contribute to wins) went down in the late 50s and 60s. Once again, you cry foul over this phenomenon, but I don't see how you can attribute the same amount of value to each case.

18.5 ppg/ 5.5 rpg/ 7.0 apg (36 mpg), .373 FG%, .814 FT%, 416 G, 21.1 PER
19.5 ppg/ 4.8 rpg/ 8.9 apg (36 mpg), .379 FG%, .795 FT%, 432 G, 18.2 PER

If we exclude his last season with boston (due to decline) and look at 1958-62:

19.8 ppg/ 5.0 rpg/ 8.9 apg (36 mpg), .376 FG%, .805 FT%, 356 G, 18.5 PER

On the second point, I meant OWS (not ORtg), which has the same issues that you acknowledged with regard to DWS. Just as Cooz unfairly benefitted with more DWS once Russell arrived, Hakeem's OWS was unfairly hurt by being on a bad offensive team (HOU was 21st out of 27 in ORtg in 1990). I feel like my point was pretty clear, even if I mistakenly said ORtg instead of OWS above.

As for pace-adjusted stats, I shared some of my objections earlier in post #61. Let's suppose a player sees his "real" statistical production RISE when his team starts playing at a faster pace, but his pace-adjusted stats show a DROP in individual production. Which measure is more accurate? If he's a high usage "star", perhaps his marginal utility can't really be much improved upon by the faster pace, simply because he was already near his peak of possible productivity. Meanwhile his lesser teammates disproportionately gain the rewards of a faster pace. Cousy is a good example of this, since he played in both the pre-shot-clock era and in the 60's when scoring was off the charts. Let's look at his PER, since it's not tied to team performance beyond the pace adjustment. It turns out Cooz had a PER over 20.0 in his first 6 seasons (not counting his rookie year, when they didn't track minutes). But from 1958-63, his PER fell below 20, despite the fact that his ppg and apg both rose on a per minute basis, and his FG% improved. What changed in 58? Bos went from averaging 105 ppg to 110 ppg; and that was just a transition year; in 59 Bos avgd 116 ppg; in 60, they avgd 124 ppg.

For Hakeem's standards (and for the standards of star offensive players in 1990), it was just that, so-so. And for someone who keeps talking about the need for more complete stats, you're not exactly doing yourself a favor by using numbers that give you the least amount of info about player performance.

"ORtg assumes a player is as good (or bad) as the team he plays on. Thus a good offensive player on a bad offensive team is unfairly punished, while a bad offensive player on a good offensive team is unfairly rewarded."

I don't know if you ever read Basketball on Paper, but this is false.

"Adjustments for pace also unfairly reward players on slow teams, or playing in slow eras."

Unfair for what reason? You can't just make a statement without explaining your point.

"Since you have some special animosity towards Hakeem, let's look at Bob Cousy..."

I don't have any particular feeling about ANY basketball player, in terms of his on-court performance. I simply look at his production. Doesn't matter if I'm saying this about MJ or Sam Perkins.

But since you did happen to bring up DWS, you're right that it IS largely team-dependent (as well as something that is hard to get in the numbers) - it also happened to be the same season that Bill Russell entered the league and the Celtics became a better defensive team (not solely bc of Bill of course, but some of defense improvement certainly goes to his credit). But then again, it's nothing I don't even think that is anything necessarily "wrong" with the model. Defense is basketball is mostly team-dependent (in comparison to offense), and great defensive bigs and interior defenders are more valuable than perimeter defenders. Any player would become a alot better defender with a legendary defender behind him in the paint and other good defensive players around him as well.

Getting defense right statistically is still a work in progress, but this gets away a bit from what we we're discussing in the first place.

And since ORtg (and DRtg) is a major component in factoring WS, it's disingenuous to pretend it's an entirely independent measure. ORtg assumes a player is as good (or bad) as the team he plays on. Thus a good offensive player on a bad offensive team is unfairly punished, while a bad offensive player on a good offensive team is unfairly rewarded. Adjustments for pace also unfairly reward players on slow teams, or playing in slow eras.

Since you have some special animosity towards Hakeem, let's look at Bob Cousy; through the first 5 years of his career that we have DWS for, Cooz was one of the worst defensive guards in BBall; in 55 he recorded just 1.5 DWS in 71 games. Then in 1957 he suddenly became the best defensive guard in the game, recording at least 4.6 DWS a season in each of his last 7 years. Hmmm, I wonder what changed in 1957...? Maybe WS doesn't properly measure how much a specific player contributes to his team's defensive performance in this case....

BTW, as for "catch-all" stats, you continue to act as if I look at only WS for my information. There are multiple sources of information that I'm using to support my case. If several of them don't exactly have Hakeem blowing away his peers on offense, then perhaps you need to be asking yourself if your opinion is a bit biased - which is little apparent from this statement you made: "And while it might not be "scientific" enough for your tastes, it does better than OWS at evaluating individual offensive performance."

Really now, is this because it's not putting Hakeem where you feel he should be?