DaggaStats, Part I: College Football Head Coach Year In Review

Hello and welcome to your annual-ish installment of coaching performance analysis, via colored numbers and bullet points and things.

I won't spend much time here describing the metrics I'm using; if you want to understand these numbers in more detail, read here and here and here.

But because those articles may inspire tl;dr - here's the description in a nutshell:

• The dataset covers the last ten years of FBS football. It includes the year, team name, name of the head coach at the beginning of the season, conference, record, and a series of four advanced metrics, described below.


• All four metrics have been re-scaled from 100 (high) to 40 (low) over the entire 10-year dataset. All four metrics are remarkably normally distributed and the mean average of each is very close to 70.


• Q stands for "Quality" and is a combination of the F/+ and Sagarin Predictor team metrics. These overall team performance metrics are both adjusted for strength of schedule so you should be able to compare teams across conferences with them.


• T stands for "Talent" and is a combination of Rivals and Scout star rankings for each team. It amounts to "average stars per player" for the last five recruiting classes; that metric is then rescaled from 100 to 40. Also, both Rivals and Scout ratings have inflated over time - in 2005 the average FBS recruit got 2.26 stars, today it's 2.65 - so I've corrected for that as well.


• C stands for "Coaching" but really it's just Q minus T. The underlying theory is that the coach is ultimately accountable when team falls short of its estimated talent, and should be given credit when the team exceeds its estimated talent.


• However, you cannot use C to compare coaches who work with different levels of talent. If you're coaching a team with a elite T score, it becomes exceedingly hard to get even an average C score. Conversely, if you're coaching a bottom-feeder team you automatically get an average C score by getting bottom-feeder results.


• RSI stands for "Replacement Suitability Index" and attempts to solve the problem of the hard-to-parse C score. You have to give the RSI an input - ideally, the talent level of a team whose coach you're trying to replace - and it evaluates other coach's C and T scores and gives you back a number that combines the two scores in a way that is relevant for the level of talent you've input.


• Unlike C, you can compare across conferences and talent levels with RSI.


• Unlike Q, you can rationally distinguish underperforming teams and overperforming teams with RSI. Generally speaking, coaches who manage to overachieve with above-average T are given a bigger boost in RSI, while coaches who underachieve with above-average T are devalued more, and the amount they are boosted/devalued depends on your input T score.


• Where not otherwise posted, please assume RSI is evaluated with an input of Texas-level talent. We'll be evaluating coaches based on how well they'd project to do at a school like Texas, given their current performance.

In this first part I'll present all FBS coaching performances in 2014, split up by conference, in reverse order of conference RSI average.

Ready? Steady? Go:

* * * * * *

#11 - Sun Belt Conference (mean RSI: 57.0)

HOT PERFORMANCE: In the Group of Five conferences, there might be no better story in 2014 than Willie Fritz and Georgia Southern. It's their first year in the FBS, his first year as coach either in the FBS or at Georgia Southern, and they won the Sun Belt in fairly dominant fashion, going 8-0 in conference and outscoring their conference foes 2 to 1.

Fritz was the coach at Sam Houston State a few years ago when they were makin' bacon in the FCS playoffs. Georgia Southern's been lurking in the shadows behind North Dakota State in the FCS the past couple years, wielding the Paul Johnson run-happy offense now seen at Georgia Tech to great effect. It's all translating well to the next level. Yet they won't play in a 2014 bowl because they're still technically in a transition year.

NOTEWORTHY: Scott Satterfield and Appalachian State also showed unusually well in their first year of FBS competition, going 7-5.

HOT SEAT: The second-year trio of Paul Petrino (Idaho), Doug Martin (NMSU) and Trent Miles (Georgia State) will need to start showing improvement soon or they'll run out of runway.

#10 - Mid American Conference (mean RSI: 57.7)

HOT PERFORMANCE: Northern Illinois continues its run of superb play with wholly unremarkable talent. But perhaps more impressive is P. J. Fleck's turnaround at Western Michigan. They were 1-11 with a Bottom Ten Q score in 2013, Fleck's first year - a classic example of why you shouldn't put too much credence into first-year coaching results.

NOTEWORTHY: Not much else. Mark Whipple has a decent-looking C score at U Mass, but it's their first year in FBS so their talent level is awful. A 53-ish Q score and 3-9 record in the MAC says they're still not very good.

HOT SEAT: Eastern Michigan might've avoided the worst record in the conference, but it's a goll-durn miracle because EMU edged out SMU for the worst Q-score in the nation. Granted, it's Chris Creighton's first year coaching a program that also had the worst Q-score in America in 2013 and that merits a degree of excuse. But when there's nowhere to go but up or sideways and your team chooses sideways, that's not a happy sign.

#9 - American Athletic Conference (mean RSI: 59.1)

HOT PERFORMANCE: Fuente esta en fuego! The former TCU OC has finally achieved a solid turnaround in year three at Memphis. He inherited a team with a 44 Q score and 62 T score and raised the Q nearly 30 points while the T score actually fell to 60. A top-tier G5 turnaround.

NOTEWORTHY: What the heck did Tony Levine do to get fired at Houston? He had an average year with average talent at an average FBS school, that's what. And if Houston decided to get rid of Tony Levine, why weren't they first in line to secure Chad Morris rather than letting him fall to SMU? Well, um.

I suspect expectations at Houston have been set too high by people who don't know football.

HOT SEAT: Willie Taggart is only in his second year but he should be doing better than 4-8. South Florida is located in too rich a talent hotspot to allow this kind of underperformance to continue for long.

#8 - Conference USA (mean RSI: 61.5)

HOT PERFORMANCE: Doc Holliday and Marshall pretty much lapped the field in 2014. Meanwhile the Skip Holtz Redemption Rollercoaster Ride continues. Louisiana Tech delivered an impressive year, two years after Holtz ran South Florida into a ditch, which was three years after he lead East Carolina to back-to-back C-USA championships.

NOTEWORTHY: In Old Dominion's first FBS year, Bobby Wilder delivered a 6-6 record. It's somehow the worst performance of the three new FBS teams, which I think says more about about the quality of the G5 conferences than it does about these schools.

Meanwhile Bill Clark's UAB squad made quite a lot of progress this year and as a reward, the whole program got sh*tcanned. Hello, Conference USA!

HOT SEAT: Todd Monken needs to show something in his third year because Southern Miss should absolutely be competitive in this league, and right now they're not.

#7 - Mountain West Conference (mean RSI: 62.8)

HOT PERFORMANCE: Jim McElwain put together a really solid three-year stint at CSU. That combined with his Alabama experience was going to land him a nice P5 gig this year. I'm a bit surprised it's Florida, but not shocked.

NOTEWORTHY: Bryan Harsin stepped in at Boise where Chris Petersen left off without skipping a beat. Not surprising. Meanwhile, Matt Wells continues to keep Gary Andersen's system in fine fettle at Utah State. Tim DeRuyter's Fresno State squad took a hard tumble in Q score this year after losing Derek Carr to the draft, but somehow managed to win their division and lose to Boise a second time in the CCG.

HOT SEAT: Bobby Hauck, you're in Las Frickin' Vegas. Why aren't you recruiting better than Bob Davie in New Mexico?

#6 - FBS Independents (mean RSI: 64.2)

HOT PERFORMANCE: Bronco Mendenhall went 8-4 after losing Taysom Hill in the fourth game of the season. They crapped their way through the next four games before suddenly turning it around and closing out the year with four wins, including the final one at Cal. Impressive.

NOTEWORTHY: I'm not sure that Brian Kelly knows how to manage highly-ranked recruits. This is the first stop of his career where his Q score fell as he got his own guys in the system. It's also the only stop of his career where he's improved the T-score of the team over his tenure. Coincidence? I'm starting to think, not.

HOT SEAT: I don't think any of these coaches are under a lot of pressure yet. Jeff Monken was the coach of Georgia Southern last year, and the players he recruited and developed went 9-3 in the Sun Belt this year. He needs more time.

#5 - Big Ten Conference (mean RSI: 68.7)

HOT PERFORMANCE: Meyer and Dantonio lead the pack, but look at Jerry Kill's year. A solid division contender with by far and away the worst T score in the league.

NOTEWORTHY: The numbers affirm that the Nebraska-Wisconsin-Oregon State coaching dance still makes no sense at all as a purely football matter. Putting aside one embarrassing loss each - or rather, considering it but only as one game in a 12/13 game season - Pelini and Andersen both had undeniably better years than Mike Riley at Oregon State, and looking at the last ten years both Nebraska and Wisconsin are an undeniably better programs than Oregon State. It's like somehow John Hodgman found himself in a threeway with Brangelina. Makes no sense.

HOT SEAT: Brady Hoke needed to be ejected from Ann Arbor with prejudice, and the M wasted no time. Still, the entire bottom half of the league is suspect going into next year. Beckman, Hazell, Flood and Wilson in particular need to show something. They're all a couple years in and the results are lacking.

#4 - Pac 12 Conference (mean RSI: 70.6)

HOT PERFORMANCE: Oregon is a bit otherworldly this year. It'll be interesting to see how Helfrich performs without Mariota around.

NOTEWORTHY: The rest of the league's top half is a motley mix of above-average. Shaw and Mora took a bit of a ratings dive this year but still acquitted themselves alright, while Whittingham, Graham and RichRod all upped their game but not to stratospheric levels. Chris Petersen and Steve Sarkisian are a step behind but also in their first year. It's not a murderer's row - Holgo would've placed second in RSI here, as would have Briles, Snyder, and Stoops - but it's very competitive.

HOT SEAT: Dykes, McIntyre, and Leach need to show some improvement soon. Dykes actually did a decent job of getting Cal off the mat in the first half of the season; but then he lost six of his last seven. Keep that up and he'll be revisiting the G5 in short order.

#3 - Atlantic Coast Conference (mean RSI: 70.8)

HOT PERFORMANCE: Probably not a coincidence that the two most innovative offensive coaches in the league are going one-two here. Paul Johnson and his option run attack, Petrino and his spread attack, both backed by quality defenses.

NOTEWORTHY: Kind of amazing that Jimbo is fourth on this list, huh? That's a direct reflection of the fact that FSU has one of the most impressive recruiting records of the last five years and still barely squeaks out wins over average teams.

Also: the reason Wisconsin is interested in Paul Chryst is because Barry Alvarez knows and trusts the guy. It's not because of his coaching tenure at Pitt, which has been so unremarkable they're thinking of changing the school colors to beige and eggshell.

HOT SEAT: Larry Fedora is lucky his team scraped their way to 6-6 because they didn't play like a bowl team this year, let alone the supposedly-fourth-most-talented team in the league. Repeat this performance again and he's in trouble.

#2 - Big XII Conference (mean RSI: 71.6)

HOT PERFORMANCE: People will talk Briles vs. Patterson, but by this Q metric there isn't much contest. TCU had the harder schedule, more efficient offensive and defensive production, comparable explosiveness, comparable wins, and the better loss, so Patterson gets the nod.

NOTEWORTHY: Here's how this chart reads: Patterson - gap - Briles - gap - Snyder - gap - Stoops - gap - Holgo - gigantic motherfrackin' cliff - five bodies at bottom of said cliff. The top half of the Big XII was a murderer's row of coaching on par with the SEC's top half. The bottom half, not so much.

Yes, Charlie Strong stands among the bodies at the cliff base, but that's primarily because he's the Cleaner. Remember: first-year results generally reflect the circumstances in which the coach was brought in, not the coach's performance per se. Sumblin fans should remember this, too.

Also: for anyone who thinks that five-star recruiting is a necessity for Texas to win in this league, I dare you to find any significant correlation between T and Q in this group. There isn't any. In this league we need to recruit for proper depth, useful development timetable, and match for scheme, so that we have no clear weaknesses for Briles et al. to gouge over and over again.

Once the depth chart is properly stocked and you're winning games, THAT'S when you worry about five-stars. Until then you're bound to be disappointed. So y'all can worry about Malik Jefferson if you like, but right now I'm more worried about QB (short-term) and DT/CB (long-term). Take care of those needs and the Maliks of the world will inevitably come.

HOT SEAT: I like the guy but Paul Rhoads needs to deliver better results at some point. Kliff Shirtlessberry also needs to get his offense into more consistent gear STAT because his defense puts up such little resistance, you'd think they do pregame warmups at Bill Cosby's wet bar.

#1 - South East Conference (mean RSI: 74.4)

HOT PERFORMANCE: Mullen and Saban top the list but you have to hand it to the entire top half of the league. It's frankly astounding that Gary Pinkel goes 10-3 and gets to the SEC CCG with a 74 T score, and still ranks no better than seventh in conference RSI.

NOTEWORTHY: Look at Muschamp. Now look at Sumlin.

Muschamp, then Sumlin.

Muslin, then Sumchamp.

Now look directly at the space between them. Keep staring.

Are you getting any stereographic 3-D pop-out yet? It's like they're the same, but somehow...inverted.

History may not repeat itself but it does rhyme, you dig?

Bizarro Muschamp.

(giggling, gurgling bong rip; in the distant aether, a practice DJ chops Bun-B)

HOT SEAT: Derek Mason wasted no time returning Vanderbilt to their proper station, didn't he?

Something tells me he's got two years to get to a bowl game, tops. Good luck with that.

* * * * * *

Here is the updated master spreadsheet, if anyone cares to take a look-see.

In the next edition I'll look a little closer at Charlie Strong's performance and introduce an analysis of changes in scores over time. In the meanwhile, let me know if there's any breakdowns of the data you want to see. I'll either add them in comments or in the next post.