This ranking method is based on the 247 composite player ratings but differs in how the class scores are calculated. 247 currently calculated a class's score using the following formula where n is commitment's rank within the school's class, Rn is the nth recruit's composite rating, and σ (sigma - standard deviation) is 9:
In effect each recruit's base rating is adjusted by subtracting 70 from their composite rating (a linear adjustment). Their base rating is then weighted according to the player's rank within the school's class as an attempt to adjust for class size in the rankings. When reviewing it I was not in agreement with how rapidly the class rank weight fell with each successive recruit and was also disappointed to see the linear formula for the base rating. As you can see on this link as well as similar studies, talent has historically been shown not to vary linearly based on recruit ranking or rating. And with respect to class size I didn't agree with a formula that weights the 12th ranked recruit in a school's class less than half as heavily as the top recruit. Depending on class positional balance there is no reason why it should be generally held that the 12th most highly rated recruit in a class won't be able to contribute as much as the highest recruit from a playing time perspective other than their respective ratings. The rating formula, therefore, is where this adjustment should be made and not in the class size.
For class size purposes I chose 22 as the magic number for two reasons. The first is that it reflects a full starting unit on each side of the ball, the second is that it is roughly 1/4 of the 85 scholarship limit. So it made sense as a starting point. After the 22nd recruit the successive impact of each additional recruit's rating drops rapidly. Logically speaking if a recruit after #22 in a class contributes heavily then he will do so at the expense of one of the more highly rated recruits. However we can't simply stop counting at 22 because of the benefits of class size when it comes to attrition, injuries, depth, etc. Below are two graphs comparing 247's calculations with mine:
In the end the rankings are very similar in many cases, which I believe to be because the 247 nth recruit weighting formula approximates the talent distribution within most recruiting classes. Whereas the non-linear adjustment should preferably be performed on the recruit ratings themselves they are approximating that relationship in their class size adjustment.
Positionally Adjusted Rankings
The second score shown on the recruiting rankings is a positionally adjusted score. Whereas the first ranking score is based on the total number of recruits in the class, the positionally adjusted score instead considers the number of recruits at each position. The baseline numbers for each position are 1 QB, 4 RB/WR, 1 TE, 4 OL, 2 DT, 2 DE, 2 OLB, 1 ILB, 4 DB, 1 K, 1 P, and 1 LS. The positions with a baseline of 1 go down in value by half with each successive recruit. The positions with a baseline greater than 1 go down by 75% for each recruit after the baseline number.
Positionally Adjusted and Weighted Rankings
The final score shown on the recruiting rankings is a positionally adjusted and weighted score. This is a further modification to the second score where each position is given a weighted importance to a football team's success. As you can see from the basic ratings explanation above, the team's TE recruits (as well as K, P, and LS) will have the same effect on the overall positionally adjusted team score as the QB recruits. Few people would argue that proficiency at the tight end position is as important to team success as proficiency at the QB position in college football. This third and final score accounts for positional differences.
