Stat-wise, I think they're all close enough to where I can see a legitimate rating system putting them in any order. What separates Wilt and Kareem (and Shaq and Hakeem, for that matter) from Robinson, in my opinion, goes beyond the box score.
oh, come on. Stat wise, no one comes close to Wilt. He is the litmus test of stats for a center. If your system ranks him 3rd best center, and Jabbar is still behind him, there is something wrong with your pure stat ranking system. If you want to say there isn't something wrong with it, then tell me what is the point of it then...for fantasy league scoring? PER is hype. NBA Eff is clearly (statistically speaking--haha) just as good if not better.
If you're only looking at non-pace adjusted, per game numbers, it's Wilt by a landslide. That's the traditional way to do it. A coach or scout might want to track the effectiveness of a player on a possession by possession basis. PER would be more illustrative for something like that as a summary metric. It's a different way of looking at the problem, and I think it's a legitimate approach. I don't think PER is exact enough to where you can say with any confidence one player is "better" than another just because he has a PER of 27 instead of 25. It could be useful as a tool for more general assessments of how a player is performing. But I would never make a blanket statement that a player is as good as his PER. For whatever it's worth, I think PER was found to correspond better with adjusted +/- stats than NBA EFF, suggesting it is a better estimator for on-court contributions. I remember reading something like that on the APBRmetrics board a while back.
You know what your problem is, you think think stats are meant to be the end all - be all measurement of a players overall worth. When stats will never be able to tell the entire story, so PER doesnt have to do anything, it only tells a piece of the puzzle. And I dont have to claim anything, EFF is basically accepted as the dumbed down version of PER. The best stats are the ones that attempt to account for every aspect of the game. EFF does not, the simpler stat is misjudging players, because its treating all stats equal. The simpler stat is more flawed, and you have yet to answer my 20 rebound statement as the foundation to why its ridiculous to treat them all the same without accounting for pace.
When it is comes to centers, it's very hard to say. Because, Kareem put up some really good games against Wilt (one game was 39 point and 20 reb game), but this is an older past his prime Wilt (35). But the same Wilt was the Finals MVP that year. Kareem was past his prime against the younger Olajuwon, and for the most part, he was guarding Sampson. Olajuwon was way too quick and explosive for Kareem at that point. Olajuwon was averaging over 31 ppg in that series. Olajuwon got the better of the Shaq matchups (for few seasons to follow), but it wasn't total domination, like Ewing and Robinson encounters.
I think people who say Wilt wouldn't school today's centers haven't really watched Bogut, Perkins, Gasol, 2008 Shaq, etc, etc, play defense very carefully. We'll never know of course, and yes people play more intense defense now (especially during the season), but I think it's ignorant to just ignore the impressive magnitude of the dude's stats. I mean his basketball stats, by the way. This is not the Hangout.
PER is based on league/player averages (in this case, presumably Finals averages for both teams). In Wade's case (and Duncan's in 2003), no one else dominated during those Finals, which would mean that their own dominance would be compared to only non-dominant players playing in the series. In Jordan's case in '91, both Magic and Pippen clearly outplayed any non-Wade and non-Duncan player in the '06 and '03 Finals. In other words, there wasn't the deviation from other players in the series that Wade and Duncan enjoyed because, well, there were better players playing in Jordan's series. This also explains how a 41/9/6/51% series from Jordan in '93 only measured a 27.6 in PER -- because Barkley averaged something like 27/14/53% that series (and Pip put up like 21/9/8). This is exactly what the case is, since 41/8.5/6.3/51% (Jordan's 1993 Finals averages) is well beyond Jordan's season averages that year in every category, yet he recorded a 29.7 PER that season and supposedly only a 27.6 PER during the Finals. That's because he had other players playing in those Finals (Barkley, Magic, Pippen in both '91 and '93) who played at a very high level while Wade and Duncan didn't. Pretty easy to understand, actually. Basically, there was a greater deviation from Wade and the other players in the '06 Finals than there was between Jordan and other players in his best Finals because, well, there were better players having better series in those Finals.
Very good post. Also there seems to be a confusion about PER. PER was never meant to be calculated for players before the NBA-ABA merger because they didn't record enough stats to get an accurate PER. Hollinger himself never calculates PER before 1977 and if you've ever seen PERs before that year it's someone or some people making an estimation so take it with a grain of salt. To whoever said EFF is as good as PER, that is complete bogus. EFF assumes that all boxscore stats are worth the same amount (which it obviously is not). PER makes an honest attempt to weigh each stat by importance (eg turnover counts more negative than an assist count positive). EFF has a correlation to winning of something like .66 and PER has a correlation over .9
