BuzzMD, what your first graph tells me is that even amateurs can spot the best talent and label it 5-star. It also could indicate, as you say, that those guys tend to go to the factories that produce A-A's.
For example, you cannot show that Reggie Bush is really considerably better in his first 3 NFL years (~3500 all-purpose yards) than Tiki Barber was (~2800) in a limited role. Some better, yes, but significantly? Was he a product of being surrounded by superior talent in college?
If I read the graph correctly, the difference drops considerably when you grant the 5-stars and focus on 4/3/2-star guys. There just isn't much difference, especially when you factor in the much greater number of 3/2-star players than 4-star.
On the second graph, there are about 10 programs that distinguish themselves. The rest are pretty much scattered in a circle. Though the overall trend is upward, the real distinct upward slope (and distinctly narrowing sigma) appears to occur at about 5500 rivals points, or about 1100/year. Between 2002-03 that would roughly include the top 50 rated recruiting classes, but from 2004 onward that is about the top 30-40 programs or so. There are a lot of 2/3-star players in that group.
So the second graph indicates a distinct correlation between the top 30-40 programs distributed between winning roughly 60-75% of games. Over a 12-game season (at 8% per win), that equates to winning between ~8 and ~9 wins.
I'd conclude that though there is a general trend, the result is not a significant advantage, except for the very top programs, and those have other advantages that quite possibly skew the correlation.
So yeah, I'll agree that there is some accuracy to the amateurs' assessments, especially wrt the 5-star level. However, I'd say the correlation becomes so broad, with so little inherrent benefit in that middle area between the factories and the have-nots, that a valid conclusion has to be tentative at best.
