I was wondering whether or not Texas Tech will be able to sustain its current number 14 position in the Rivals Recruiting Rankings.
When we look at the recruiting tables, we see that Tech already has 21 commitments while a number of other powers have so far added far fewer commitments such as Georgia (ranked 15, 16 commitments), Pittsburgh (ranked 16, 18 commitments), Tennessee (ranked 21, 18 commitments) and so forth
It would seem that once these teams sign their full classes, Texas Tech's ranking position might fall. So, is that actually true?
I wanted to see if that observation was correct, and to do so I tried to figure out how Rivals actually calculates their ranking. Intrigued yet?
More good stuff after the jump!
My understanding is that Rivals does not publish its formula (and if I'm wrong, I just wasted 2 hours and probably have in any case), but a quick search of the interwebs led me to an LSU Tigers message board, where the author purported that a statistics graduate from California-Berkeley had figured out the formula.
I went through the formula and tested it out on Texas Tech's incoming 2011 recruiting class. My results? I calculated 1287 points versus Rival's 1285 recruiting points (Feel free to check out now if you are unimpressed).
Now, I'm not sure how accurate the formula provided by the author is, and I certainly was not a mathematics major, so there are a few insights that I won't be able to give that someone much smarter than me probably could provide.
Nevertheless, the formula seems plausible and more importantly it seems to work! The formula also provides us with a much better understanding of what it would take for us to improve or score or sustain our rankings.
So without further ado, let's jump to the formula. I'll try to make this as painless as possible, and will be liberally borrowing from the poster on the LSU message board.
(Note that all data was obtained from the Rivals.com website. I use the terms recruit, prospect, commitment, player, etc interchangeably. In reality all names are all just commitments at this stage).
H * (n/(n+m)) + L * (m/(n+m))
Don't be intimidated!
Let's break this down, so that we can figure out what it's supposed to mean.
First, let's start with H and L.
H = Potential High Score whereby teams receive a score of 250 for every five-star, 140 for every four-star, 75 for every three-star, 20 for every two-star and 10 for every one-star.
L = Potential Low Score whereby teams receive a score of 18 for every five-star, 12 for every four-star, 8 for every three-star, 3 for every two-star and 1 for every one-star.
Only the Top 20 signees are counted in calculating the H and the L scores.
If we look at Tech's 2011 commitment list, the Red Raiders have received 21 commitments so far. Four of these players are four-star players (Amaro, Edwards, T. Morales and Williams), 16 of these players are three-star players, and 1 player is not rated.
Since we are only focusing on the Top 20 signees, we will ignore the unranked signee when calculating the H and L scores.
That leaves us with 4 four-stars and 3 16-star commitments.
The H Score would be calculated as follows:
(4 x 140) + (16 x 75) = 1,760
H = 1,760
(whereby a four-star player receives 140 points and a three-star player receives 75 points).
Similarly, the L Score would we calculated as follows:
(4 x 12) + (16 x 8) = 176
L = 176
(whereby a four-star player receives 12 points and a three-star player receives 8 points).
So our H=1,760 and our L = 176
That takes care of H and L.
What about calculating the variables m and n?
The variable "m" is easy. "m" is just a constant which in this case works out to be 50. It could be any number, but Rivals apparently chose to use 50. Why? I don't know. Anyone on DTN care to comment?
The Three Components of n
The variable n is actually made up of three components which I'll outline as follows (all recruits are considered when calculating n):
Any recruits which have a Rivals 100 designation are awarded points on a sliding scale: 10 points for being ranked 1-10, 9 points for being ranked 11-20, 8 points for being ranked 21-30 and so forth until the 100th recruit.
A similar sliding scale is used for players ranked Rivals 100 Juco and Rivals 100 Prep. In those cases, the sliding scale stops at 50.
In our current commitment class, TE Jace Amaro is the only Rivals 100 player (Texas has 8!) Amaro is ranked 48 overall in the country. On that basis Amaro (and of course Texas Tech) is awarded 6 points.
Among the Rivals Juco 100 players, CB Justin Williams is ranked 25 and so Texas Tech receives a further 8 points.
Among the Rivals Prep 100 players, LB Terrell Hartsfield is ranked number 12, and so Texas Tech receives another 9 points.
Amaro: 6 points
Williams: 8 points
Hartsfield: 9 points
In total, Texas Tech receives 23 points for landing three Rivals 100 from high school, JUCO and Prep.
(Aside: Anyone else see how recruiting a top ranked JUCO or Prep player can unduly skew this score? Anyone believe that a Rivals 100 recruit which is ranked number 10 is equivalent to a Rivals 100 Prep player which is also ranked number 10?)
Rivals Position Rankings
Rivals awards the following points based on position rankings:
- 24 points for players which are ranked number 1
- 18 points for players which are ranked 2 to 5
- 8 points for players which are ranked from 6 to X (range varies by position)
Each position apparently has a different cut off point. I'm not clear what those cut offs are, so I based my calculations conservatively only counting Tech commitments which were in the Top 5 by position. Those players are Tony Morales and Jace Amaro.
Tech has 12 of its 21 players which have received a position ranking. I'll list the players in order of position ranking.
- T Morales (OL): 1
- Amaro (TE): 4
- Brewer (QB): 12
- Williams (RB): 12
- A Morales (OL): 18
- Evans (DE) : 27
- Marquez (RB): 29
- Edwards (WR): 29
- Washington (ATH): 33
- Castleman (DL): 34
- Clark (OL) : 35
- Daniels (ATH): 66
(T. Morales: 24) + (Amaro: 18) = 42. Tech receives 42 points.
Bonus points above 3.0 Average Star Rating
The final component of the "n" variable awards additional points based on the number of basis points a team receives over 3.0.
Texas Tech's current average star-rating score is shown as 3.05. The actual score is 3.0476. Bonus points are awarded as follows:
3.0476 - 3.00 = 0.0476 x 100 = 48
On this basis Tech's n score benefits by receiving a further 48 points!
If a team has a 2.99 average star rating, too bad. They're out of the money!
One of the keys to getting a high recruiting score is keeping the average star ranking above 3.00.
The University of Texas currently boasts the country's top rated recruiting class with an average star rating of 3.78, which would mean that Texas receives 780 additional points on their n score - just another reason to hate on Texas.
Adding Up the N Score
We calculate the N Score as follows:
Rivals 100: 23 points
Rivals Position Rankings: 42 points
3-Star Average Bonus: 48 points
Total n = Score: 113 points
Putting it all Together
Ok, let's see how that all works together. The original formula is:
H x (n/(n+m)) + L x (m/(n+m))
Texas Tech's values are as follows based on the earlier discussion:
H = 1,760
L = 192
m = 50
n = 113
1760 x (113/(113+50) + 192 x (50/(113+50))
Dammit! Off by 1.5%.
I don't know if Rivals actually uses this specific formula (and there could be some rounding issues in there), but at least in the case of Texas Tech's score, it seem pretty accurate.
What is even more interesting is understanding which variables are the most sensitive, and the impact that those sensitivities have on Texas Tech's recruiting score.
According to the formula, the most likely variables which can affect Tech's current scores by order of impact are:
- Addition or subtraction of a 4 star or 3-star prospect (impacts H and L values)
- Upgrade or downgrade of a current 3-star or 4-star prospect (impacts H and L values).
Average-star score falling below 3.0 implying a loss of the 3-star bonus points (impacts n)
Changes to the average star score
- If one of our current commitments leapfrogs into the top 1-5 position rankings, the n score could improve marginally, but at this late stage in the recruiting process, that is unlikely to happen.
- If one of the current recruits is able to crack the Rivals 100 list then the n score could also improve slightly, but that also is probably unlikely at this stage in the recruiting process.
Implications on Tech's Recruiting Rankings
Based on the above discussion, all Tech has to do is sign one more three-star prospect and its current score will improve by a 100 to 200 points (or even higher if we can steal a couple more Rivals 50 JUCO and Prep guys!).
When we look at the other schools currently ranked below Texas Tech, those schools which a) have less than 20 recruits, b) have received 3 to 4 four-star commitments, and c) have average star rankings above 3.0 seem most poised to bypass Tech at some stage in the recruiting process.
There are about 15-20 schools which fit this description. Without analyzing each specific team, I would predict on the basis of how I believe the ranking formulas work, that Texas Tech would most likely end up with a recruiting class ranked between 20-30 in the country - a great outcome for the coaching staff's first year.
If anyone else cares to crunch the numbers of those other teams or road test this formula, feel free!