Ok, out on a limb here, but a quick eyeball perusal of the various metrics used in recruiting rankings seemed to reveal that the best predictor of success at the college level was position ranking. I was curious if anyone has done a more serious analysis to see if that correlation holds up.
The database I looked at was narrow (M recruits, 2002-2007, Rivals position rank data), but I was taken immediately that, overall, recruits ranked in the top 10-15 at their positions became strong contributors in their college careers at a high percentage, that there was an inflection point below 15, and another for everyone except O-line at about 30 -- ranked nationally below 30 for anything but O-line said that a player wasn't going to be a contributor, but more like a depth guy.
The O-line thing makes sense because there are more of them on the field, and they are probably more interchangeable than any other position.
The other thing that stood out to me was the number of top 10-15 position rank players committed so far this year to Michigan. With (likely) a couple more Top-Tenners to come, this class would be superior to any in the 2002-2008 period.
UFR, the correlation makes sense to me, since the forced rankings are the product of multiple observers of multiple players' relative impact on the games they play. More than the numerical ratings, which can be influenced more subjectively, the forced ranking at least differentiates the dominant from the good from the meh.
This also seems to support a correlation I've used since coaching kids in soccer, baseball and football starting in the mid-80's -- your success on the field is determined by your top 2-3 on the field and your bottom 2-3 on the field. If that set of 4-6 is better than your opponents', you rock. If not, you get rocked. The middle 3-5 in baseball, 5-7 in soccer and 16-18 in football are going to be a push. That may seem obvious, but it really shows how much happens at the margins, and how much match-ups matter. My monster on your meh: my advantage.
Michigan seems to be setting itself up these last couple of classes for some pretty good match-up situations. I'm especially intrigued by the slot-dots and hybrid safety-LBs.
A further musing is on how the spread is a great equalizer for quarterbacks -- it seems to take merely adequate to be good, and the good can become great due to the reduced demands on pre-snap read and ability to deliver spot-on throws over miles of field with correct trajectory, speed and timing. Data: HS position rankings for a selection of this year's QB performers -- McCoy (15), Bradford (12), Daniel (6), Herrell (7) -- oh yeah, and some guy named Tebow (1) to provide an outlier.
Anyway, Sunday morning coffee musings, FWIW. I'd really be interested if anyone has done any broader stats work on position rank as predictor.