Writers of Pro Football Prospectus 2008

01 Aug 2011

NCAA Resource-Centric Model: Part II

Guest Column by Kevin Haynes

Part I of this article is here.

One of the most overlooked aspects of the decline and fall of Jim Tressel is the fact that of all the teams in the Big Ten, Ohio State had the least reason to cheat. The school has such a strong competitive advantage that it’s almost as if it was created for the sole purpose of being good at football. Take the size of the war chest that the Buckeyes football team has at its disposal: not only is the program in Columbus blessed with the largest budget in college football, but the gap between their monetary resources and that of their closet competition is so sizable that it’s more than twice as large as eight of the 12 teams in the Big 10.

The Buckeyes have not just the good fortune to be located in the heart of a fantastic football state, but to also have little direct in-state competition for Ohio’s top recruits. A recruit choosing Cincinnati, Miami of Ohio, or Ohio University over Ohio State, straight up, is rarer than a Prairie View A&M win in the 90's. And although Ohio State certainly is a good school with plenty of alumni that contribute to society in all sorts of positive ways, there should be no disregarding the fact that its admission standards are not so high that mediocre students need not apply.

Ohio State (2006 to 2011)
Year Model's Win Projection Actual Wins +/-
2006 8.4 12 3.6
2007 10.3 11 0.7
2008 11.4 10 -1.4
2009 9.5 10 0.5
2010 10.1 11 0.9
2011 10.5* ? N/A
2011 9.1~ ? N/A
*-(w/ Pryor, et al.) ~-(w/o Pryor, et al.)

So does all that mean Ohio State shouldn’t get credit when it wins? It’s not like Tressel and his staff had to outsmart their opponents. Instead, they could simply rely on the faster, stronger, more talented Buckeyes to control the line of scrimmage, keep possession of the ball, and cruise to yet another win over a less capable team like Minnesota or Purdue. Yet a glance at the far right column in the above table shows that according to the resource-centric model’s predictions, Ohio State actually exceeded its projected regular season win totals since 2006 at a fairly consistent rate. So whatever Tressel was doing in Columbus, be it hiring the right conditioning coach, identifying which recruits would best fit his system and then rigging raffles to ensure those same recruits won prime merchandise at OSU’s summer camp, or staying up late to watch game film, it seems to have paid off.

It might be a stretch, of course, to accept the resource-centric model’s discrepancies with reality as signs of over or underachievement by coaches. They could just be computational errors that stem from the variables included (or not included) in the model, or problems inherent to the technique of multiple regression in and of itself. Furthermore, is it really fair to pin, say, Sam Bradford’s 2009 shoulder injury on Oklahoma’s head coach, Bob Stoops? Can we really give full credit to Gene Chizik for Auburn’s improbable 2010 National Championship? And can we truly make fair judgments about a coach’s ability with just five years (and, in many cases, less than that) of data? Probably not. But curiosity compels us to go on, so here are the best coaches of BCS teams since 2006 according to the plus-minus measure:

Best coaches of BCS teams since 2006
Coach School N Avg. +/-
Chip Kelly Oregon 2 2.53
Brian Kelly Cincinnati / Notre Dame 4 2.30
Lloyd Carr ~ Michigan 2 2.21
Jimbo Fisher Florida State 1 1.90
Gary Pinkel Missouri 5 1.68
Jeff Jagodzinski * Boston College 2 1.65
Bo Pelini Nebraska 3 1.57
Bob Stoops Oklahoma 5 1.29
Mike Gundy Oklahoma State 5 1.29
Bret Bielema Wisconsin 5 1.13
~ - Retired * -Fired (sort of)

Sifting through these results is a bit of a chore, and probably can’t be accomplished with much nuance in this short space, but a few high-level findings do stand out. There’s a lot of good coaching happening in states that begin with the letter "O," so North Carolina (see below) may want to change it's name to Orth Carolina. Also, Gary Pinkel might be the best coach that doesn't draw constant praise on studio shows.

The model may give too much credit to coaches when teams experience significant turnarounds from one year to the next (such as Michigan winning seven games under Lloyd Carr in 2005 and 11 games in 2006). However, because there does seem to be a solid case to be made for most of the coaches on the above list to be among the best in the college ranks -- Chip Kelly has turned Oregon into a true power, Jimbo Fisher has rejuvenated Florida State, and Brian Kelly’s run at Cincinnati almost single-handedly qualified him for the Notre Dame job -- the model does produce a good mix of expected and unexpected results. Now lets turn our attention to the other end of the spectrum: the worst coaches of BCS teams since 2006

Worst coaches of BCS teams since 2006
Coach School N Avg. +/-
John Bunting * North Carolina 1 -2.05
John L. Smith * Michigan State 1 -2.17
Mike Shula * Alabama 1 -2.29
Tyrone Willingham * Washington 3 -2.34
Ted Roof * Duke 2 -2.36
Chuck Amato * NC State 1 -2.54
Turner Gill Kansas 1 -3.23
Paul Wulff Washington State 3 -3.32
Walt Harris * Stanford 1 -4.99
Butch Jones Cincinnati 1 -5.17
* -Fired

It’s tough making conclusions on such limited data, but on the surface we can certainly make a few claims. First, a good number of these underperforming coaches seem to have gotten what they deserved: a pink slip. Second, asides from Mike Shula at Alabama and perhaps John L. Smith at Michigan State, each of the jobs on this list are not attached to schools that should be especially good at football due to their academic standards, program budgets, or geographical location in a talent-poor state. We don’t really have that much data for most of these coaches, so it’s not completely fair to consider the above results as 100 percent indicative of their abilities. Given the current employment status of so many of the names on this list, however, it might not be the worst idea for Butch Jones and his family to hold off on purchasing a home in Cincinnati for at least another season or two.

Kevin Haynes is a statistician in Cincinnati, Ohio. He earned a B.A. in English and Creative Writing at Clemson University and has a M.A. in Applied Economics from the University of Cincinnati. A lifelong Michigan fan, Kevin has spent the last four years making excuses for what’s been happening on the field in Ann Arbor, and swears that this season he won’t get too excited if the Wolverines beat Notre Dame.

Football Outsiders is always accepting guest columns that have a unique perspective on either the NFL or college football. Send your ideas or samples to mailbag-at-footballoutsiders.com.

Posted by: Guest on 01 Aug 2011

35 comments, Last at 04 Aug 2011, 2:52am by Solomon

Comments

1
by Aaron Brooks' Good Twin (not verified) :: Mon, 08/01/2011 - 12:59pm

I see a couple of coaches on the bad list who were afflicted with "Everyone graduated last year" syndrome (Gill, Jones).

2
by sundown (not verified) :: Mon, 08/01/2011 - 1:08pm

Turner Gill had 13 starters come back from Mark Mangino's last team, so that wasn't his problem. I thought he was a good hire--and think he will likely get the ship righted-- but they were an absolute disaster before showing a bit of life late in the year. The circumstances of Mangino's departure clearly shook everything up and the transition didn't seem to be all that smooth, but having everybody gone was not an excuse for Gill.

9
by Aaron Brooks' Good Twin (not verified) :: Mon, 08/01/2011 - 4:30pm

It's also a matter of which 13 starters, no?

Let's just say that the ingredients behind the 2008 team had left the cupboard by the time Gill arrived.

14
by sundown (not verified) :: Mon, 08/01/2011 - 6:30pm

In the Big 12 North, 13 starters coming back should have equaled a pretty good season. They weren't going to beat Nebraska, but every game (and 2nd place in the North) was there for the taking. Reesing was the major loss, but not having him wasn't why they lost to a Division 2 team. And the saving grace of their season was coming back from 45-17 down to beat Colorado. Awesome comeback for them, but how do you fall behind a Dan Hawkins CU team by that much?

19
by Marcus C. (not verified) :: Tue, 08/02/2011 - 12:33am

The year you're thinking of was 2007 when they went to the Orange Bowl. 2008 wasn't much to brag about, 8-5. In 2009 they collapsed at the end but still ended up 5-7 but with the Mangino distractions gone and all those guys coming back they should have bettered that record last season. Instead, they lost Gill's first game at home 6-3 to North Dakota State and lost their first 4 Big XII games by a combined score of 187-30. The cupboard wasn't that bare.

3
by Intropy :: Mon, 08/01/2011 - 1:29pm

Why is recruiting locality-based? I doubt players with eyes on the pros care all that much about proximity to home. Why isn't Alabama shopping in Iowa, BYU picking up players in Cincinnati, or Miami getting the top talent in Fairbanks? In the days of celluloid and crank-start cars proximity is a major factor. But in an era of internet video, cheap commercial air travel, and budgets in the tens of millions of dollars, that should matter little.

4
by trill :: Mon, 08/01/2011 - 2:18pm

When you equalize for prestige and resources, recruiting is mostly about coach-to-coach relationships. Recruiters need to know which players are putting in extra work, which are making their grades, which ones have problems at home, etc. For certain positions, especially DB, it helps a lot to be able to see the games live, as high school game film is invariably crappy. And while budgets are big, they're not bottomless; owning your home turf is efficient economically and in terms of talent. If you're not recruiting hard in your home state, the smaller schools in your state will be happy to pick up those 3-and-4 star recruits you don't have time to visit.

5
by Kevin Haynes (not verified) :: Mon, 08/01/2011 - 3:10pm

Maybe it shouldn't matter, and maybe it's mattering less these days, but for whatever reason, players overwhelmingly choose to attend colleges within driving distance of their home.

Check out OSU's roster and you'll see it filled with players from OH and the Midwest: http://espn.go.com/college-football/team/roster/_/id/194/ohio-state-buck...

Same with Florida: http://www.gatorzone.com/football/bios.php

Clearly it pays to be located in a talent-strong state. If Florida was in Rhode Island, it wouldn't be Florida.

6
by Intropy :: Mon, 08/01/2011 - 3:30pm

But we just heard that recruiters recruit close to home. Do the players prefer to stay near home or is that just who's recruiting?

15
by sundown (not verified) :: Mon, 08/01/2011 - 6:34pm

You'd like to play someplace where your friends and family couldn't easily see you? Even blue chips who could pick any school frequently like the familiarity of being close to home. They and their families have been fans of teams for years, they've known other guys who went there, the coaches are known to them simply by being from the same area, even before recruiting begins.

16
by Intropy :: Mon, 08/01/2011 - 7:22pm

I guess I just assumed it would be the same on the athletic side as the academic side, except even less biased because of the scholarships. I went to high school in Dallas, and several friends ended up at UT. But that didn't keep those with the grades and the money from going to MIT, Princeton, Harvard, Stanford, et al, and if tuition money wasn't a concern I'm sure a lot more would have gone far afield. Athletes with realistic NFL aspirations would likewise go to the top places for their chosen discipline, would they not?

18
by Marcus C. (not verified) :: Tue, 08/02/2011 - 12:02am

All things being equal, sure. But all things aren't always equal. First, the pool of guys with realistic aspirations of the pros is incredibly small, so focusing the discussion on just those few doesn't make much sense. There are also top football schools scattered all across the country, so they can pick one of those and still stay relatively close to home if they want. And there are factors like assuring playing time, looking for coaches and systems that best cater to your style of play, etc., that play a part.

20
by Intropy :: Tue, 08/02/2011 - 2:47am

All those factors you list at the end make perfect sense but have nothing to do with location. You make a good point about there being top football schools scattered around. Maybe it's inertia keeping those schools on top. They are good, so good players want to go there. They're all good, so the differences between them are minor or come down to preferences that will be different for each player. Add to that a small bias towards players staying near home and another bias for recruiters economizing by looking close and you get the present breakdown.

25
by Aaron Brooks' Good Twin (not verified) :: Tue, 08/02/2011 - 11:38am

Even amongst academically elite schools, most students are regional, and when multiple are clustered in a region (Ivy league) most students are local.

Penn primarily gets students from NJ/PA, Princeton and Columbia from NJ/NY, Yale from MA/CT/NY, Harvard from MA, Cornell from NY, etc. Especially for undergrad. It's more widely spread for grad school.

I would expect football schools to work similarly. For every good football player from Texas who goes to Ohio, 4 go to Texas/A&M/Texas Tech/TCU, etc.

31
by panthersnbraves :: Tue, 08/02/2011 - 4:15pm

Half of the cars at Duke have NJ license plates...

33
by Aaron Brooks' Good Twin (not verified) :: Wed, 08/03/2011 - 11:47am

I said "academically elite." =) Duke is for kids who didn't get into the Harvard of the North.

7
by MTR (not verified) :: Mon, 08/01/2011 - 3:35pm

I know citing a half remembered study won't be too compelling, but I recall one that found only two significant factors when trying to predict where recruits would go: how close the school was and how many wins they had the previous year.

8
by DisplacedPackerFan :: Mon, 08/01/2011 - 3:58pm

Because it does still matter. As you say it probably has less impact on the very top players, but the majority of the major teams are comprised of players from the local recruiting grounds.

Wisconsin (108) - 47 from WI, 2 MI, 8 MN, 7 IL, 1 IA, 12 OH, 14 FL, 5 TX. So 65 come from WI or states that border it, another 12 from a good state in their conference. But this is a team that doesn't have a good recruiting state.

Ohio State (112) - 73 are from Ohio (65%)

Alabama (112) - 64 from AL (57%)

USC (104) - 71 are from CA (68%)

Of those 4 schools, WI is clearly the weakest football power, and they have the largest percentage of players not from the home state.

So while the very best players may not care the majority of the roster from the powerhouse schools come from their own state. I would imagine if you look at starters on a team that the numbers likely are pretty close to the roster percentages. If 40-70% of your starters don't matter for wins and losses then locality based recruiting doesn't matter either. It's still a lot easier to get a player from your local recruiting grounds to play for you.

32
by Will Allen :: Tue, 08/02/2011 - 4:41pm

If Minnesota ever starts recruiting 80% of the in-state players who now go to Wisconsin, Iowa, Nebraska, Notre Dame, among other prominent schools, they'll become a consistent top-25 school. It is a testament of how poorly managed the program has been that other programs have had such a free hand at poaching talent from a top 15 (roughly) metropolitan area.

12
by Kibbles :: Mon, 08/01/2011 - 6:03pm

Financials. Recruiting is limited by time and money. It is more cost- and time-efficient to recruit a 4 star recruit 100 miles from campus than to recruit one 2,000 miles from campus. Those time and money savings can be used to then recruit more players.

Remember, recruiting is more than just scouting. You can't just watch film of a guy in Alaska and say "hmmm... he looks good" and have him magically attend your school. Recruiters have to build relationships with the recruit, with the recruit's parents, with the recruit's coach, etc. Recruits choose which schools they go to, and overwhelmingly they go to schools where the coach has built a strong relationship with them. Nobody is going to go across the country to play for a coach who he only knows through email and text messages when he can go 100 miles down the road to play for a coach who has personally met his family 3 different times. For the time and money it takes a Florida coach to visit one 4-star recruit in California, for instance, they could visit three 4-star recruits in central Florida. And since there's not much of a difference in value between one 4-star recruit and the next, a team's best bet is to stockpile as many as possible, which means maximizing contact.

Now, none of this applies to the truly elite prospects. Florida routinely goes across the country to recruit really high-end guys like Percy Harvin (who came from North Carolina, iirc). They send recruiters out to southern California to try to snipe 5 stars from USC. It's important to remember, though, that these elite prospects make up less than 20% of even the best teams in the nation (and less than 10% or even 5% of most BCS teams).

Add to that other human factors at play. It's true that most college recruits dream of making it to the NFL, but a team has to convince the recruit that they offer the best chance to make it to the NFL, which typically requires face time (unless you're USC, UF, Texas, or tOSU). Further, if a recruit feels multiple teams offer a great chance, they're going to fall back on other factors. Among those factors is how close a college is to home (and therefore how close he will be to his family). Most college-bound students prefer to remain relatively close to home. Also, if a football player grew up a major fan of a team, that'll also play into the decision-making process. Do you think it's any coincidence that Tim Tebow wound up attending the college that he grew up rooting for? When John Brantley was deciding between Texas and Florida, is it really surprising that he opted for the college where his father and uncle had played?

13
by Intropy :: Mon, 08/01/2011 - 6:24pm

That all makes sense. Incoming students will always have their preferences whether it's go to the team I rooted for, go close, go far, go to the place with the best physics depart, or whatever else. But the rest of it dealing with relationship building and trip costs and whatnot strikes me as a problem to solve rather than an inevitability. One of these years someone is going to innovate somehow, whether it's recruiting bureaus spread around, training camps, or more likely something out of left field, and find themselves with a truly national recruiting pool.

30
by Kibbles :: Tue, 08/02/2011 - 3:04pm

I'm never going to bet against innovation. Never. 10 years ago, nobody would have guessed that text messaging would be as important in the recruiting war as it's become. With that said, there's literally no impetus for any of the "haves" to innovate. There's no real difference between 4-star guys in California and 4-star guys in Florida, so why should either USC or Florida devote resources to tapping into a pipeline that offers no advantage over the pipeline they already have.

Several of the smaller schools, or the "have nots", do have a vested interest in innovation, although they'll always be competing against a lack of funds and history. In the end, there are always going to be structural impediments to national recruiting.

17
by justanothersteve :: Mon, 08/01/2011 - 10:34pm

Very nice summary of some of the financial aspects of recruiting. A minor addition: Sometimes an assistant coach or booster has a pipeline from an athletic hotbed to a particular school. Many schools will have an assistant coach familiar with recruiting in Texas and/or Florida for that reason.

Minor nitpick: Percy Harvin was Virginia Beach. But the city south border is the NC state line so you were close. I only know because I was living there at the time.

10
by ClemsonMatt (not verified) :: Mon, 08/01/2011 - 4:55pm

Is the full list somewhere?

Given how announcers talk about Clemson, you'd think Dabo Swinney and Bowden would be in Ted Roof territory.

11
by Kevin Haynes (not verified) :: Mon, 08/01/2011 - 5:11pm

I can put it together if you give me a day or two. My hunch is that the model doesn't hold Swinney in as much trouble as the media does as CU (my alma mater!) doesn't match up that well resource-wise against the top ACC teams, let alone the SEC teams it has to recruit against.

22
by ClemsonMatt (not verified) :: Tue, 08/02/2011 - 10:00am

I saw you got your undergrad there. Me too.

I'm wondering if the model will hold him somewhat accountable since his GPR is so high though.

Was player turnover one of the variables investigated?

23
by ClemsonMatt (not verified) :: Tue, 08/02/2011 - 10:03am

I saw you got your undergrad there. Me too.

I'm wondering if the model will hold him somewhat accountable since his GPR is so high though.

Was player turnover one of the variables investigated?

24
by Kevin Haynes (not verified) :: Tue, 08/02/2011 - 10:24am

I used the number of returning starters on both sides of the ball, so in many ways player turnover is accounted for. What I can't quite account for is losing a superstar like Spiller and replacing him with someone who isn't Spiller...or Kyle Parker forgetting how to play quarterback from one season to the next.

Across the board, though, I like to think those types of "substitutions" even out because at good schools, good players are replaced with good players, and of course this occurs at "bad" schools, but with less talented players.

From my four years at Clemson, I think the biggest takeaway I had is being a Tigers fan is a fun but tough life. There is so much going right there - the tradition, the gameday atmosphere, the fans, the uniforms (which are totally underrated, in my opinion) - but like I said in a previous post, they also face some pretty tough obstacles when it comes to being a regular BCS bowl-qualifying team. SC is a good football state, but it's small, and CU has stiff competition in recruiting both from South Carolina and from the litany of programs that are within a few hours drive - UGA, GA Tech, all the NC schools, TN, and on and on. Throw in a budget that isn't by any means through the roof and above-average academic standards and you've got a steep hill to climb in order to be good every year. It can be done, I think, but it will probably take a coach who truly understands how to get the most out of what he's got to work with, and while I like Swinney, I'm not sure he's that guy. Really, I'm not sure how many of "those guys" are out there, though, so maybe when you find a likable personality who everyone can get behind, then you keep him for as long as you can and hope for the best.

27
by witless chum :: Tue, 08/02/2011 - 12:23pm

I think orange is an underrated uniform color. Remember the old Denver Broncos duds, when they called their defense the Orange Crush? My school, Michigan State, has a nice shade of green now, but they've worn some atrocities.

34
by ClemsonMatt (not verified) :: Wed, 08/03/2011 - 4:29pm

As a ChemEng student, what I mostly took away was it was really cruel hearing the game while I was studying in Earle Hall. I went to more games during the 4 years I was in Texas than I was able to make it to in the 4 years I was there.

As long as the coach considers academics important and doesn't horribly underperform, I'm ok with it.

I can take it when I interview and we've lost. I'm not sure I could do as well if we had a cheating scandal or something.

21
by Danish Denver-Fan :: Tue, 08/02/2011 - 9:09am

I've seen one college game my entire life, and it was a Notre Dame game last year. They were playing... ahm... Tulsa? ND had the game almost wrapped up (ball at the 20 yard-line, down two with time running out). Kelly then, inexplicably called a deep end zone pass to their star reciever which, of course, was intercepted; ball game.

Am I remembering that right? It seemed like a crazy call to me - can someone explain why Kelly still has a job?

26
by Travis :: Tue, 08/02/2011 - 11:49am

From Kelly's post-game press conference:

Q. I know you said you would call that same play again. 36 seconds left, a field goal kicker who hasn't missed in the last 18 tries, why not just set up a field goal?

COACH KELLY: Why not try to get Michael Floyd one on one against a 5'9" corner and call a timeout, here is what we're going to do. Second down, take a shot here. If we don't like it, let's throw that thing away.

Tommy [Rees] wanted to do all those things. Tommy is a gamer. You saw him competing out there. He knows the deal. He's a quarterback.

Again, to me this is how we play. We're going to play aggressive. We're going to play smart. If it's not there, we're going to throw it away. We're going to line up on third down and have another shot to get that thing even closer.

But I would make the call again and I would hope that the process of learning would have a different outcome.

28
by witless chum :: Tue, 08/02/2011 - 12:41pm

Wow, I'd missed that whole thing. What the hell is the value of going for a TD there? ND goes up by a TD, but they give Tulsa the ball with 36 seconds or so to try to pull off an, admittedly unlikely, TD to win. If they just run the clock down to three seconds, (they can, it's second down and Tulsa only has one remaining timeout) it all comes down to a pretty makeable field goal.

Edited to add:
Even if the QB throws it away when it isn't open, that would allow Tulsa to get the ball back, because they could call a timeout after stopping ND on third down, which would give them the ball with probably 28 seconds or so, need a FG to win.

After watching the highlight reel, maybe Kelly's real answer is 'Have you seen what my special teams is doing today? I wouldn't trust them to get laid in a whorehouse' but he doesn't want to say that, so he just blusters and doubles down.

29
by Danish Denver-Fan :: Tue, 08/02/2011 - 1:14pm

Yes, the above is the only thing he can say after a blunder like that. I just brought it up because Kelly appeared in the article, and it reminded me. I figured i must've missed something - I even looked up college FG percentage.

Considering how much scrutiny a guy like Barry Switzer goes under for an amazingly less controversial call you've gotta figure that the coach i gone if he does something like this in the NFL. It's 5 years before he can be anything but a recievers coach again. Right?

35
by Solomon :: Thu, 08/04/2011 - 2:52am

"Instead, they could simply rely on the faster, stronger, more talented Buckeyes to control the line of scrimmage, keep possession of the ball, and cruise to yet another win over a less capable team like Minnesota or Purdue..." or Michigan lately.

Mr. Haynes -- I imagine you are taking delight in the Buckeyes' problems of late, but remember that your school up north committed its own NCAA violations not so long ago. In addition, Michigan was guilty of the serious "failure to monitor" infraction, whereas Ohio State was not charged with this.

Ohio State is a good academic school -- it ranks in the mid-50s of U.S. colleges in the U.S. News & World Report Rankings, good for 6th in the Big Ten. It has tightened its admission standards significantly in the last 15 or so years.

We shall see if Brady Hoke can back up his arrogant talk so far. Stuff like (paraphrasing) "We're not rebuilding, we're Michigan for crying out loud." Mississippi State would beg to differ about the rebuilding part.

Rivalries aside, the wins modeling is interesting.