I Don't Want My P-E-R

I think John Hollinger is a brilliant guy. He's arguably the flat-out smartest writer working for ESPN, and probably any other sports empire. He understands statistics and projections in a way that normal people don't, and if he starred in the show Numb3rs, episodes wouldn't last past the second commerical break. He can talk numbers for hours upon hours and remain giddy about it. I respect his intelligence, and I respect that he's doing something that he loves. Everyone who talks about him is in awe of his ability to crunch numbers and create formulas.

However, to be a Hollingerian, you need to love his Player Efficiency Rating (PER), and I don't. I don't like it quite a bit, in fact. If I was into hating things, I might hate it.

The PER (explained by the Honge himself here) is meant to measure the rate of an NBA player's per-minute statistical production, the idea being that such a system allows you to compare players who don't play the same amount of minutes, or play in systems that limit their statistical productivity (i.e. Shawn Marion might score more points than Rip Hamilton because his team gets more offensive possesions). Hollinger is quick to point out that the PER is not meant to be the final, definitive word on an NBA player's production, and he's always quick to point out the glaring flaw that the PER doesn't really measure defensive ability.

Those flaws aside, there are a few other problems I have with the system. Hollinger has often said that the benefit of the PER is that it gives a unifying number by which to compare players, since you can't just go to scoring or rebounding to tell the difference between Randy Foye and Chris Mihm. However, that only makes sense if the PER is meant for people who know nothing about basketball. Fans of the game understand the value (or lack thereof) of any type of statistic, and no matter how long Hollinger's system is defended (and there are a LOT of people who love it), the PER is just that: a statistic. Too often, the defense of the PER is that is unifies stats, which is, strictly speaking, untrue. The PER take a lot of stats into account, but then tosses in huge intangibles like "pace," filters the numbers through a formula, and ends up with a PER. This is still merely a number, and it doesn't clarify anything any more than comparing stats based simply on averages.

In addition, Hollinger has said that the PER is used to make "comparisons between players who play differing minutes, or in different systems or what-not -- comparisons which, using conventional stats, are almost impossible." Again, those comparasins are only impossible if one is completely on the outside of basketball looking in. It's actually quite possible to compare players who differ in style or position or number of minutes played, and that's by watching these players, taking into account their statistics, and using your brain and basketball knowledge to come to a conclusion. You've no doubt noticed that the system sounds remarkably similar to the PER, and indeed it is, the problem with the PER is that it takes the "use your brain" part out of it. Perhaps that's the unintended point, sadly, in that it tells people what to think, and that's also probably why I (almost) hate it. That's incredibly far from Hollinger's intention, to be sure, but it is indeed the unintended outcome quite often.

As I've said, if stats are your only concern, the PER is as decent a number as any. But if the PER is designed to determine a player's level of production overall -- aka how good of a basketball player he is, overall -- here's where it beings to fail. Here are some projections from the upcoming season's projected PER:

--Yao Ming will be the most productive player in the NBA this year. As with all PER projections, this is largely based on Yao's stats compared to "similar players at a similar age," and where their statistics went the following year. The problem is that Yao -- along with Lebron, Shaq, Steve Nash, and Kevin Garnett, among others -- are so vastly different than any player who has come before them that it renders comparasins all but moot.

--Also according to the projections, 12 players will be more productive than Tim Duncan. Now, I won't argue that Duncan's numbers will blow away the field this year, but again, this is why a stat like this doesn't work to measure how good a basketball player is.

--In addition, Steve Nash is ranked all the way at 19th, while Vince Carter, Shawn Marion, Allen Iverson, Luol Deng, Josh Howard, Jermaine O'Neal, Jason Kidd and Rip Hamilton are all below Atlanta's Josh Smith.

Now, I know that some will argue that I just don't get the PER. On the contrary, I do understand it indeed, I just don't see the need for it. Perhaps it's because I'm not a numbers guy, or because I don't like being told what to think, or because I don't see the point of a list that ranks basketball players with phantom stats and places the best player at #4 (Kobe, as if it needed to be said) and one of the most accomplished at #13 (Duncan), with the guy who should have won his first MVP award last year at #19 (Nash). And while I'm not closed to it, I don't see myself coming around on this one.

Again, Hollinger's a good and smart man. He's often the first to point out when his formulas lead to wacky conclusions or totally miss intangibles like Steve Nash's ability to not age after turning 30. He's also a very smart basketball analyst, as in evident every time he strays from mere numbers in his writings. So no, I don't want to ditch John Hollinger, I just don't want his stat.