APBRmetrics

kbomb · Joined: 09 Aug 2005 Posts: 9 Location: Gardena, CA

I noticed that as the league has expanded, the PER of the top players have increased as well. For instance, up until mid-90's, it was rare to see a player having a PER over 25. These days though, it is pretty common to have seasons where there are 5-6 players with PER over 25. Since the PER is standardized at 15, does this mean that the separation of PER between the top player and the average player become greater with bigger sample sizes? If that is the case, the comparisons of players of different era based on PER would be rather inaccurate.
Another question is does the standard deviation in PER calculation increases with a bigger sample size? If that's the case shouldn't we adjust PER accordingly when we compare players of different era?

Thanks

jkubatko · Joined: 05 Jan 2005 Posts: 702 Location: Columbus, OH

Neil Paine · Joined: 13 Oct 2005 Posts: 774 Location: Atlanta, GA

Clearly we are in a golden age of NBA superstars, which partially explains the record number of 25+ PER guys -- Lebron, Wade, Dirk, Kobe, Garnett, Brand, and Yao are all smack-dab in the middle of their primes (or not yet in their primes in the case of LBJ and D-Wade, which is a scary thought), and Iverson proved in '06 that he's still capable of similar greatness, history be damned (30-ish small guards who rely on quickness have a nasty tendency to break down quickly and irrevocably). Still, perhaps expansion/dilution has something to do with things as well... If so, maybe standard deviations away from the mean would be a more accurate way to account for cross-era differences in playing style and talent.

Nikos · Joined: 16 Jan 2005 Posts: 346

Could the referee's emphasis on hand checking be the reason guys like Pierce and Iverson had career years?