Hokepoints is Legit, Maybe Not Legit. Yo.

Submitted by Seth on December 17th, 2013 at 2:28 PM

george and jab

Moe (1) and Jabrill (2), via.

In last week's roundtable on the state of the conference I pulled out this table grading the new Big Ten's teams on their 2013 seasons (by Fremeau Efficiency Index) and their futures (by composite 247 score for the 2012-'14 classes):

West | East
School FEI Grade Rcrt | School FEI 2013 Rcrt
Wisconsin 13th A C+ | OSU 8th A A+
Iowa 30th B C | MSU 9th A B
Minnesota 49th C D+ | Michigan 29th B A+
Nebraska 51st C B | Indiana 62nd D+ C
NW'ern 60th C- C | PSU 65th D+ B
Illinois 75th D C- | Maryland 74th D C+
Purdue 114th F C- | Rutgers 98th E- B-
AVG 56th 2.0 2.0 | AVG 49th 2.1 3.0

That's about how I feel: A conference baseline of "C" (ie ranked around 50th) teams with one division recruiting at a "B" level and the other "getting the most out of" C level recruiting.

This I pulled from a spreadsheet of FEI and recruiting data that I'd like to mine further, because if you're looking at a chart it still counts as doing work.

Recruiting = legit, yo/maybe not so legit. So here's a new look at the old stand-by: recruiting on the Y-axis, performance on the X-axis, and a nice, heavy trend line with an R-squared of 0.46 to show an inconvenient-for-narratives correlation. Performance is FEI expressed as a percentile. The composite ranking is a bit more complex: the 2009 (5th year seniors) is weighted at 0.5 the 2010 and 2011 classes at full, the 2012 class at 0.40 and the 2013 at 0.10, which are arbitrary values I assigned based on expectations of how much a class contributes to a given team.


Blicking on it makes it cig.

It says they're correlated, but doesn't necessarily mean one is causing the other. FWIW the r-squared of the Rivals composite determined the same way was .4135; I haven't done Scout or ESPN yet. Look at how the correlation of recruiting %-ile of each class and 2013 performance %-ile changes by year:

Class 247 R-Squared Rivals R-Squared
2009 (5th yrs) 0.3681 0.3204
2010 (Jr/Sr) 0.3696 0.3776
2011 (So/Jr) 0.4596 0.3848
2012 (Fr/So) 0.4143 0.4134
2013 (RS/Fr) 0.4724 0.4076
2014 (recruits) 0.4417 0.4098

The highest correlation is to the freshman class, and the 3rd-highest is to the class that's not even on campus yet. There's a strong echo effect going on here, wherein the teams that are good today are getting the highest-ranked recruits. The diminishing returns from seniors, I would posit, are because they're the classes hit hardest by attrition, and most likely to have been recruited by a different coach or to a program in very different circumstances.

The other thing that immediately jumped out at me about that chart is look at all the color on top of the black trend line. Those gray dots are mid-major programs, who are largely outperforming expectations from recruiting, versus only one SEC team managing to do so. I bet that's a system bias in the recruiting rankings: there's little to parse between an under-the-radar guy who commits to Purdue versus one going to NIU except one of those is a Big Ten school.


Meeting expectations. Here's the BCS schools significantly (more than 0.1 difference between FEI %-ile and recruiting composite %-ile) outperforming their freshman talent acquisition rate:

Over 50% of KSU's starters went the JUCO route. [Jamie Squire, SBNation]
Team Conf FEI 247 Cp Diff
Arizona St Pac 12 90% 61% +30%
Kansas St Big XII 71% 43% +28%
Baylor Big XII 89% 62% +27%
Stanford Pac 12 100% 76% +24%
Missouri SEC 88% 65% +23%
Wisconsin Big Ten 84% 61% +23%
Okla St Big XII 89% 70% +19%
Michigan St Big Ten 88% 70% +18%
Duke ACC 68% 51% +17%
Arizona Pac12 73% 59% +14%
Washington Pac12 80% 69% +11%

I don't know what it means except ungh MSU and ungh Arizona of course would be on there. As for Michigan's bowl opponent, that's less of a "finds and develops" thing than that Bill Snyder is heavily reliant on JUCO transfers, who don't show up in the recruiting rankings so much. The only SEC team in there is Missouri, who wasn't in the SEC when they were rating their recruits and got a nice boost from much of their SEC-Eastmates being actually pretty crappy teams this year. It could mean they're overrated; it could mean the way Brian Fremeau does his statistics hurts teams that play a lot of tough conference opponents alongside a nonconference spate that Michigan's Team 7 would have found appalling.

As for the underperformers:

Imagine if Kevin Wilson turns Indiana into a conference power for 10 years, then was fired ostensibly for having the lowest graduation rates in the conference, and then Indiana went 0-10 versus FBS competition the following year. Ladies and gentlemen: Cal.
Team Conf FEI 247 Cp Diff
California Pac12 26% 77% -51%
Tennessee SEC 43% 87% -44%
Florida SEC 58% 98% -39%
Arkansas SEC 37% 72% -36%
Texas Big XII 66% 100% -34%
Kentucky SEC 33% 62% -29%
Kansas Big XII 31% 57% -26%
Colorado Pac12 28% 54% -26%
Penn State Big Ten 51% 75% -24%
N.C. State ACC 33% 57% -24%
Purdue Big Ten 20% 43% -23%
Virginia ACC 43% 63% -20%

Horray Michigan isn't on the list. The next 10% is USC, Maryland, Nebraska, West Virginia, Texas Tech, Illinois, North Carolina, Oklahoma and LSU. We're at –9%, in a pile with Miami (YTM) and Georgia.

After Cal, however, the teams on this list suggest Michigan has to do a lot more than wait for the 2012 and 2013 kids to mature. Tennessee, Florida, Arkansas and Texas had comparable classes in 2010 and 2011, though there's some coaching changes in there that change things a bit.

Ohio State broke even, but that's about as well as they can do since they're 98th percentile in recruiting. That is the goal.


Trending the Good Way. Michigan's top-o'-the-country recruiting has been of recent vintage. Other teams with significant recent leaps from their norms: Texas A&M, Vanderbilt, Miami, Baylor, Purdue, KSU, Zona, Ohio State, and Syracuse. Of these just Miami is around the same base level. You'll recognize the teams on the opposite side of the ledger: USC, Cal, Texas, Oklahoma, Kansas, Penn State: fading stars.

But knowing where teams are today doesn't help. I want a comparison from recent history so I grabbed three 2010 teams with a similar profile to 2013 Michigan's and tracked them over the next three seasons. And that was pretty useless but here it is anyway:

Team FEI Rec So Fr Recr 1 yr 2 yrs 3 yrs
2013 Michigan 0.122 7-5 6th 4th 13th ? ? ?
2010 Team 1 0.124 8-5 3rd 3rd 3rd 10-2 (7-2)
6th in AP
10th in FEI
7-6 (5-4)
30th FEI
9-4 (6-3)
18th FEI
2010 Team 2 0.120 8-5 24th 18th 33rd 7-6 (4-5)
16th in FEI
11-2 (6-2)
5th AP
3rd FEI
8-4 (4-4)
20th AP
25th FEI
2010 Team 3 0.123 8-4 19th 28th 32nd 7-6 (2-6)
47th in FEI
8-5 (4-4)
62nd FEI
6-6 (3-5)
42nd FEI
2010 Team 4 0.130 5-7 29th 28th 10th 10-4 (6-2)
22nd AP
32nd FEI
11-2 (7-1)
11th AP
25th FEI
8-2 (7-1)
12th AP
26th FEI

Not really helpful. Team 1 is USC, which Lane Kiffin. Team 2 is Texas A&M, which Johnny Football. Team 3 is Mississippi State, which very much isn't Michigan fergodsakes. Team 4 is Clemson. If Devin turns into 2005 Vince Young next year, upside seems pretty possible; cliffs are everywhere. Then again look what Lane Kiffin could do with just slightly more incoming talent, and the worst Borges hater would take a midseason tackle-over intermission any year to avoid goat boy.



December 17th, 2013 at 3:43 PM ^

Re: system bias. This is a huge problem with recruiting. Its a weird system. Coaches start the recruiting process, send out offers, then the recruiting gurus use those offer lists to help their evaluations, and this leads other coaches to evaluate the players again or for the first time.


December 17th, 2013 at 6:14 PM ^

is worth at least 1-2 stars.  That is, the average star ratings of the recruits have to be adjusted for the ability of the staff to develop players.  I would add these:

UM = 0

MSU = 1.5

OSU =0

Wisc = 1 (the old regime)

Iowas = 1

Rest =0 



December 17th, 2013 at 7:51 PM ^

I'm really surprised to see FEI rank us so high in the Big Ten, especially ahead of Iowa and Nebraska. It lends credence to the idea that everybody (except Minnesota) gets up for Michigan, while it's tough for us to play our "A" game week in and week out.

Zone Left

December 17th, 2013 at 8:48 PM ^

There's another way to look at seniors--having a lot of them probably correlates highly with FEI success, just like having highly ranked freshmen correlates with past success Think Wisconsin...retaining seniors and developing them is a path to success.

I'd guess the three strongest measurable components of FEI success would be recruiting, mean player seniority, and last year's record. The remainder is probably a good estimate of coaching ability.