"The Mathlete's rank-by-offers method"
In reference to Christian Turner who may be underrated, the front page has mentioned "The Mathlete's rank-by-offers method" a few times in the last 24 hours.
I suspect this is an algorithm he developed and houses on his desktop. Firstmost, I care to know more about this. In particular, how do you weigh various offers, Mathlete (by quantity *and* quality? how?). More than that, I anticipate a ranking tool like this could succeed online. After all, 247 is presently gobbling up the market in no small part because it "invented" the Composite, which is a superior tool, and avid fans naturally flock to superior tools.
My long-time theory (unscientific; so far, dependent on moralizing eye tests and hot takey gut feels) has been that the top reason the recruiting services produce rankings that are broadly predictive of future success is because the services *try* to carbon copy the opinions of the best coaches, as expressed in scholarship offers.
I suspect a rank-by-offers algorithm could quantify that correlation (which may just be interesting to me). Furthermore, a rank-by-offers method may stand as even more predictive of future success (All-American teams and high draft picks) than the current system. And that question must interest many people.
I'd wager there wouldn't be 14 had they been under the previous staff these last two years....
we haven't had more than 3 players drafted in one year in a long time.
I've heard the same about Urban Meyer
It works a lot better when you are landing 7 of the top 50 recruits in a single class.
Verified offers could make for interesting metrics as the OP postulates
April 10th, 2017 at 10:25 PM ^
Everything's unreliable to some extent; this is another signal and I think it makes sense to include it. Ensemble methods are really good.
It is also worth noting that whenever U-M snags a low-ranked prospect, that player will look better under this metric, on average, just by virtue of having a Michigan offer. Treacherous.
April 10th, 2017 at 10:06 PM ^
There are also offers that are not comitable. Coach offers a kid and says there are four people ahead of them. (Not sure how that actually works in communication to the players.)
Offers between schools can carry different weights, but how do you weigh a commitable vice a non-commitable offer? Even Harbaugh seems to use the non-commitable offer as a way of saying hello.
Be careful, Mathlete! They'll steal it! Before there was the 247 Composite, there was the MGoBlog turd ferguson Composite.
As guy above says - an offer nowadays is like shaking someone's hand. We give out tons of offers a year. What do they really mean? Just like when guys ranked #600, 700 have a Bama offer and Iowa board goes crazy they landed said guy because "Saban wanted him".
Also I'd be interested in how this adjusts for early committs who get less offers over time, and the "committability" of offers
I'd also be curious if he only factors in power 5 schools that actually recruit nationally, or do prospects in, say, the midwest/southeast that get more offers primarily because there is a higher density of schools get ranked higher than those in California where your average MAC/CUSA/AAC/Sun Belt school is unlikely to get a player and thus less likely to offer?
Holy run-on sentences!!! Punctuation, man! Punctuation!!
I read his comment and it flowed in my head fine. But when I read it with your comment in mind it was pretty amusing. LOL :)
factor in the commitable/not commitable/early verbal/solid verbal offers?
The only way to really rank these kids is to...well... .... pay them what they are worth.
Just saying...