the teams with scores less than 0 "sell" in the recruiting business and teams over 0 do not. There might be a financial incentive to hyping or over-hyping players that get recruited by the sellers.
Nevertheless, interesting numbers. Thanks for posting.
Hey, just a thought - since subtracting ranks like that might not be the best approach (ranks are not equal intervals), maybe create z-scores of each score (rather than rank) and then subtract z-scores. The ends of your distribution (teams at the top and bottom) will be less affected by that, but the middle portion might re-arrange some and it would be a technically better approach.
1. You have a good point that recruiting rankings are affected by cohorts. With absolutely no knowledge of the player, I assume when Alabama, a big Florida school, USC, or Texas offers someone a good distance away, that the kid is really good. Some of our recruiting ranking is probably due to this factor.
2. Perhaps these even out, but it seems that the recruiting rankings should be retabulated to consider the effects of transfers and serious injuries. In some cases, the coach is a factor in the player leaving, but not all the time.
3. Recruiting scores are also slightly biased against good recruiting schools since only 22 players start. Yes it is great having 4 star recruits on the bench for depth/injury purposes and to cover any mistakes you made in selecting players. A team with fewer high-ranked recruits can compete.
4. Achievements should factor out cupcake games against D2/FBS opponents.
The charts do confirm intuitive feeling on underachievers (Texas, Tennessee, Old Miss, Miami, UCLA, LSU, TAMU, and us), but USC seems a reach. The overachievers confirm (WASU,MSU,Stanford, Utah, K-State, and Wisconsin), but many others just say that the teams didn't have great talent, but did OK