John Hollinger, the statistical whiz who derived Player Efficiency Rating (PER), has released his projected 2007/2008 PER statistics for every player who played at least 500 minutes last season. PER is, as he explains it to Henry Abbott of TrueHoop:

Anyway, PER is a per-minute rating of a player's statistical productivity. It's great for measuring a player's "tangibles"; the one area it struggles is with defense because so much of that remains a black box for the analytical community.

I use it so much because it makes it very easy to make comparisons between players who play differing minutes, or in different systems or what-not -- comparisons which, using conventional stats, are almost impossible.

Here is how PER is derived:

I base the projections on a tool called similarity scores.

For each player, I use as a comparison the players from the past 20 years who are the most similar, based on age, height and stats over the past three seasons. Some players will have more comparables than others, depending on how unusual they are -- guys with freak heights (Yao Ming, Earl Boykins), freak ages (Dikembe Mutombo) or freak stats (Andrei Kirilenko) will have relatively few, while a more generic player like Al Harrington or Devin Brown could have over a hundred.

From that point, I see what their most similar players did a year later, and project those changes onto the stats of the player being studied. So, for example, the reason that Yao Ming's PER is projected to rise sharply this year is because the most similar players also saw their PERs increase sharply at the same age; similarly, Andre Miller is expected to tank because a number of similar players hit the wall at his age.

PER is normalized each year to 15; that is, if you have a PER of 15, you're an average player. After the jump, I've posted a table showing how Magic players fared last year, as well as Hollinger's projections for them this year, and their difference. Alternatively, you can click here to view the predictions directly on ESPN.com. While the raw numbers are free, the player profiles are not.

	Player	2006/2007 PER	2007/2008 PER (Proj.)	Difference
	Trevor Ariza	16.27	15.76	-0.51
	Carlos Arroyo	14.48	14.56	+0.08
	James Augustine	16.60	---	---
	Tony Battie	9.85	9.23	-0.35
	Keith Bogans	9.25	9.84	+0.59
	Keyon Dooling	11.16	11.54	+0.38
	Adonal Foyle	13.70	----	----
	Pat Garrity	6.40	----	----
	Marcin Gortat	----	11.19	----
	Dwight Howard	21.19	24.51	+3.32
	Rashard Lewis	20.78	20.10	-0.68
	Jameer Nelson	13.92	16.29	+2.37
	J.J. Redick	13.05	13.76	+0.71
	Hedo Turkoglu	14.26	14.38	+0.12

For one thing, it's nice to know that Jameer is expected to improve. For another, Marcin Gortat projects to be be better than Tony Battie was last year. And although Hollinger didn't project Adonal Foyle's PER this season, his showing from last season also exceeds Tony Battie's. Maybe we won't miss him so badly after all.

Regarding Rashard Lewis: he's rated highly -- 27th overall -- but he is expected to decline a bit from last season. I'm not John Hollinger, but I can assert with confidence that PERs tend not to increase after they decline. Keep in mind that 'Shard is stuck here until 2012, at which point he will be 33 years old and drawing a $23 million salary.

Uh-oh.

The biggest news here is just how highly Dwight Howard is rated. The 24.51 projection is the seventh-highest in the league and ahead of other All-Star big-men Kevin Garnett (9), Chris Bosh (10), Carlos Boozer (11), Amare Stoudemire (12), and Tim Duncan (13). Put it this way: Howard's projected PER this year is just 0.05 points below LeBron James' actual PER last year. Yes, they play completely different positions and completely different styles, but if Dwight Howard can be as effective for us as LeBron James was for Cleveland, why couldn't we reach the NBA Finals? Sure, it'd take some luck -- Cleveland swept Washington, which was playing without Gilbert Arenas and Caron Butler, in the first round -- but I wouldn't rule it out at this point.

Potential NBA champions? Heavens, no.

Potential conference contenders? You betcha.