I'm a little late to the party, but I had to respond to jayparr...
Regarding star ratings, you are right that signing classes full of low-rated players is not a recipe for success. If you do that, then on average you will have less talent than other teams. Even though every 4 and 5 star player doesn't pan out, more of them become superstars than do 2 and 3 star players. That's a fact. So star rankings are fairly accurate when rating entire classes, which is evidenced by the obvious correlation of recruiting rankings and team performance.
That said, star rankings are not remotely reliable at predicting the success of an individual player. As much as work as they try to put into the rankings, a ton of it is guesswork, and there are way too many factors to consider future success, including work ethic, body maturation, and coachability. I would MUCH, MUCH rather have a 2 star (who I'm sure will end up 3*) that our coaches prefer than a 4 star that the scouting services ranked higher. They liked him, and they clearly liked him a lot to offer him with so much time and so few 'ships left.
I don't have the time to go find the stats, but I've seen them before, and they very much back up what I just said. The teams that are always at the top of the recruiting rankings are generally at the top of the polls as well (assuming they have a decent coach). Even so, if you look at the eventual success of individual players (measured by All-Americans or NFL draftees) compared to their star rankings, the rankings aren't even close to accurate.