On Sunday, we published a piece analyzing MAC recruiting rankings from 2007-09 and found multiple inconsistences between the rankings of classes and players and the results on the field. We also found several cases in which Rivals and Scout were dead on. This piece was not meant as a critique of those who evaluate prospects. They spend countless hours trying to get it right. Do they get it wrong? Yes, sometimes. But so do coaches, even ones who have reached the top of their profession like Nick Saban, Urban Meyer and Frank Beamer. Recruiting is an inexact science and it always will be. Here's how Scout.com arrives at its rankings, according to Allen Trieu, Midwest analyst.
The process begins at the state level where analysts view highlight packages, camp performances, and live games. Scout is represented in every state, or close to it, and two guys run the show in Ohio. From there, the state analysts report their findings to a regional manager, such as Trieu. If they believe a prospect is undervalued or overvalued they make the regional manager aware. Regional managers get together once a month — on conference call, I believe — to adjust Scout's national rankings. Thus, a prospect that might have looked just OK at a summer camp can raise his stock with a strong season.
What are the credentials of the analysts? All of them played high school football. Some played in college. Others are former coaches. All have 5-6 years experience in evaluating prospects. "All of us have experience in what a high caliber Division I prospect looks like, what a mid caliber Division I prospect looks like, and what a D-II caliber prospect looks like."
Other notes:
-Trieu said evaluating prospects has become easier in the past few years because of increased exposure through camps and combines.
-Scout does not take into consideration what scholarship offers a prospect has attracted when assessing him a grade. However, if a low-level recruit begins generating several offers, Scout will re-analyze his game tape to determine if they made an error.
