Interesting SSOG. I know you've shown DPAR data before, but I'm still not familiar enough with it to trust it.
It's understandable, because some of it really flies in the face of conventional thinking. For instance, they said that, on a per-play basis, Ricky Williams was only the 13th best RB in the NFL in the season he led the league in rushing- he just got such a ridiculous amount of carries that he managed to compile some great stats. Sort of like Eddie George, but not quite as bad. Still, I've been following them for 3 years now, and both the stats and the writers are incredible. They have unflinchingly defended every single accusation of the system, they have constantly searched for flaws in the system and impose sanity-checks (for instance, last season SF and Hou were so brutal that Indy dropped from #1 to #6 in their rankings one week despite having a good game, just because their schedule adjustment was so laughable. They explained the flaw in detail, showed some historical comparisons to demonstrate how extreme the adjustment was, and then listed Indy at the top of the list in the #1 spot, but kept the #6 results that the formula put out). They are constantly improving the system, and after every improvement they run back and check it against previous years' data to ensure that they have improved the correlation between DVOA in one week and winning the following week (testing its accuracy as a predictive stat as well as as a descriptive stat). They have posted the correlation several times (I forget it offhand), but it's currently better than any other statistic out there (points scored, points allowed, scoring differential, record, YPA, YPA allowed, YPA - YPA allowed, etc). People also tested FO's formula against several internet power rankings, and again found a greater correlation between the FO rankings and winning the following week than any other power rankings on the 'net. At the end of every season, King Kaufman compares preseason playoff projections with actual playoff teams and lists who had the most accurate predictions, and for 2 straight years someone from FO has taken home the "prize". FO's formulas predicted Washington's dramatic turnaround at midseason last year, Carolina's dramatic turnaround at midseason the year before, and as far as I know, DVOA was the only thing on the face of the planet that called San Diego a playoff team before the 2004 season. They've also been referred to by numerous sources such as the New York Times (for whom they occassionally write articles) as "The Bill James of football".What I like best about them is that they don't just say "Here are these numbers, they're great"... they actually put their money where their mouth is and SHOW why the numbers are great. The explain where the numbers succeeded, explain where the numbers failed, and are constantly looking for ways to make them better. And they aren't just numbers people- they have volunteers who actually watch the games and chart data on such things as missed tackles, uncatchable balls, etc. They break down game film and discuss matchups and all of that. They place a huge emphasis on blocking and frequently will write articles about matchups between Offensive and Defensive linemen (like two straight pieces on the Johnathon Ogden vs. Dwight Freeney matchups). For someone like me, who is in love with all aspects of the game of football, they're really heaven on the internet.
I can't vouch for their Fantasy Football numbers, because I've never bought them or checked them out, but as far as DVOA and describing real-life occurances, their numbers are the best I've seen. Not perfect, of course- no numbers ever can perfectly describe such complex and variable interactions as occur on the football field- but they're very open about what the weaknesses are.
I know, I know, I sound like a walking Football Outsiders commercial. I guess I am- but off of the FBGs board, I'm very much the same way about Football Guys and VBD and such, which I think is leading the way in fantasy football. In fact, DPAR and VBD are both pretty similar- both attempt to measure how many points a player is worth above a replacement-level player at the same position- it's just that VBD measures it in fantasy points, and DPAR measures it in actual NFL points.