WOW! Nene, Kevin Martin, Horford, Tyrus, Millsap,West, Gordon, S. Curry......... Didn't expect to see some of those names. Defense must NOT factor into PERS rankings.
No, Hollinger and Basketball-Reference (BR) calculate PER differently. Here are today's Top 5 compared. Same number of games. So it's not because one site updated tonight's games and the other didn't. ESPN PER..BR PER..games..player 27.15 27.1 34 Paul 25.50 25.4 33 Wade 25.37 25.3 35 James 24.91 24.8 29 Nowitzki 24.25 24.1 31 Howard
I'm an Aldridge fan, and have been since he was at Texas, but I don't think his recent rise is entirely a result of the offense changing around him. McMillan has preached about wanting Aldridge playing on the low block since early last season, but Aldridge would routinely get muscled out, or just plain float out to the perimeter, seemingly on his own. Now, Aldridge is stronger, he's getting foul calls, he's more polished at the offensive end, and he knows he has that outside shot to fall back on - all of which make him a better post-up player. He's a better player this year, it's partly how the offense has taken advantage of his improvement that his numbers are looking so much better, but give Aldridge credit for getting his offensive game into better shape, too.
If he were just shooting 60 percent from deep over the last 10 games I would say that was just a hot streak. His entire game is different now, he is playing the way you expect a power forward to play. I love the new LaMonster and I hope he keeps it up.
It seems like a recent problem; just yesterday, Roy's PER went from 15.1 to 15.0, which is a neat trick considering he did not play. So BR may have recently changed something in their formulas (perhaps rounding down on every lost digit instead of only 4 or lower).
His stats didn't change, but everyone else's did, so he went from a tiny bit above average to average (15.0). There are definitely different formulas for PER. In fact each site used to keep theirs secret (do they still?), so how could they be the same.
It'd be really interesting to see a season-long chart of PER by player for our team. In the first few games it'd be all over the place for everyone, since there's a lot of variation from game to game. But it'd be interesting to see the overlaps in player performance over time. Then maybe overlay that with team win percentage. I know that Aldridge, Matthews and Batum's PER's have taken significant jumps recently, and Roy's took a drastic dump after the first dozen or so games. The big issue with our record is that there was a real bottoming out when Roy was sucking and nobody else was improving because we were still running the same Roy-based offense. If you take that section out our record would be drastically better. Here's a nice chart (from BlazersEdge--it also features Houston and is now out of date, but ignore that) that illustrates it: That nosedive is when our team went 4-9, or .307. If you take what I like to euphemistically call "that adjustment period" out of our record, we're 14-7, .666 for the season. That would put us around the 7th best record in the league. And maybe the 4th or 5th seed.
They can't be that different; they're within .15 points most of the time. I think Hollinger adjusts his PER formula a little or uses a different sample to normalize the numbers to 15 = average. BR.com says they use Hollinger's publicized formula, which produces pretty close results. It's not worth the money to get access to very slightly better data.
PER was created by John Hollinger and he explicitly defined it in at least one book. If sites don't use the equation, then it's not really PER. The complication is that the statistic is based on league stats, and that makes it complicated to figure... figuring out the Pace for the whole league, for example, takes a bit of number crunching. I'm not sure that most sites keep their numbers up-to-date on a daily basis. Ed O.
