Damn, how long did that take? Cool idea, definitely. Looking at the 2007 team, there are a couple of points where I disagree a little:
Mankins as a 6 and Neal as a 5- We saw what happened in the SB when Mankins had a bad game and Neal got hurt. Personally, I'd rank them both at least a point higher.
Welker as a 7- had the most catches in the NFL, probably the best slot receiver in the league.You're probably going to argue that he was pretty much feeding off of Moss, which is fine--I disagree with it, but that's one that we're probably not going to settle on anytime soon. Also, putting Welker as a 7 goes against your 6/7 distinction
Asante Samuel, IMO, should have been an 8 in 2007. And, as Deus mentioned, Ellis is much better than a 5 as a KR.
Mostly, though, for the way that it's being applied, there seem to be 2 assumptions that I don't agree with:
1) That the averages work out: i.e., that having a 5 and a 9 is equivalent to having two 7s.
2) That positions are equally important: having a 10 at QB and a 5 at C is the same as having a 10 at C and a 5 at QB.
All in all, though, I think the fundamental premise of it is really cool, and definitely a really interesting way to attack this issue.