One of my favorite stats is Pythagorean expectation. It looks at a team's points for and points allowed, and then estimates what we should expect that team's winning percentage to be. If you then look at the difference between how a team did versus its expectation, you can get a sense as to whether the team won or lost more games than it "should" have. Just multiply the expectation by the number of games played, and subtract that from the actual games won.
The NCAA's stats archive goes back a full 10 years now, so it's possible to look over the last decade to see which teams stand out as winning more games than they should have (I call them overachievers for convenience) and which lost more games than they should have (underachievers). Since most teams' differences are less than a game either way, a team must have a difference greater than one or less than -1 to count as an overachiever or underachiever by my standards.
BIGGEST OVERACHIEVERS (SEASON)
2009 Wyoming: +3.37
2005 UCLA: +3.05
These were the only two teams to go above their expectation by more than 3 wins (though '06 Arkansas State came close with +2.99). Wyoming needed every bit of that to go bowling in Dave Christensen's first season there, proving that overachieving doesn't always refer to a team that won its conference unexpectedly.
As for UCLA, the Bruins somehow got to 10-2 in '05 despite being coached by Karl Dorrell. It should come as no surprise then that it was a completely unexpected occurrence since on a points for/allowed basis, UCLA played like the six- or seven-win team it was in every other season under Dorrell.
In the SEC, the largest overachievers were 2007 Mississippi State (+2.08) and 2001 Auburn (+2.04).
BIGGEST OVERACHIEVERS (DECADE)
The Wildcats from Chicago take this one and it's really not even close. Northwestern had a difference of more than one in a whopping six of the last ten seasons, and they just missed out on a seventh with a +0.99 in 2006. It's popular among some to label these differences as "luck," but if it happens this often, there's something more going on.
Randy Walker and Pat Fitzgerald consistently coaxed an extra win a year out of their teams than the number say they should have been able to, a sure sign of good coaching. When it comes to getting the most out of what you've got, Northwestern wins that contest easily.
Five teams came in tied for second with four overachieving years: Baylor, Cincinnati, Hawai'i, Oregon, and San Jose State. The Warriors and Ducks get extra points though for never having an underachieving year; the other three schools all had one apiece. The biggest overachiever in the SEC for the decade was Georgia, who had three such seasons (2002, +1.01; 2007, +1.15; 2008, 1.64) without an underachieving season.
After the jump, the biggest underachievers.
BIGGEST UNDERACHIEVERS, SEASON
2009 Oklahoma: -3.16
I find it interesting that the largest overachievers and underachievers happened in the same season. Anyway, this past season's Sooners take the dubious honor of being the only team to have a difference less than -3. With Sam Bradford going out at various times in the year plus three losses by a field goal or less (and a fourth by a touchdown), it's not hard to see how it happened. Of course the NCAA doesn't do nice things like filter out Oklahoma's 64-0 win over Idaho State, so the difference would move towards zero with I-AA contests filtered out. Alas, I don't have the time to filter out a decade's worth of I-AA games, so this stands.
Other big underachievers include 2006 Clemson (-2.94), 2004 Alabama (-2.90), and 2009 Louisiana Tech (-2.88). Yes, it just wouldn't be an underachievers list without the presence of Mike Shula or Tommy Bowden.
BIGGEST UNDERACHIEVERS, DECADE
To be clear, K-State is the "winner" of this category. These purple-wearing Wildcats underachieved in six of the last ten seasons, forming nearly a perfect mirror image of Northwestern. I say "nearly" because Kansas State underachieved by more than two games twice, including an impressive -2.87 in 2001. The other season was 2003, with a -2.09. Wait, wasn't 2003 the year the Wildcats beat down Oklahoma 35-7 in the Big 12 title game? Yes, and it was also the year they lost close games to Marshall and Oklahoma State, two teams that finished the year unranked.
Alabama gets a mention for also having six underachieving seasons, though last year's +1.18 kind of makes up for one of them. Bama also gets the distinction of being the only team to underachieve by more than two games on three occasions in the past 10 years: -2.01 in 2000, -2.45 in 2003, and -2.90 in 2004. In all, Mike DuBose and Mike Shula can be blamed for five of the six bad seasons, with Shula turning in a strong -1.83 in his final season to boot. That does mean Nick Saban had an underachieving year, which was 2007 with a -1.07. It shouldn't be too hard to figure out which game got away from him.
There a lot of other consistent underachievers, far more than overachievers. That makes sense, considering it's easier to underachieve than overachieve. Some notables:
- Florida: Yes, one of the decade's kingpins was a serial underachiever with no overachieving seasons (though 2006's +0.87 comes close). Stever Spurrier's final team in 2001 was his best since '96 yet it not only underachieved numerically (-1.30), but also by not even winning the SEC East. Two of Ron Zook's teams unsurprisingly underachieved: -1.14 in 2003 and -1.72 in 2004. The final was in 2007, when Tim Tebow's Heisman winning heroics weren't enough to overcome a bad defense and prevent a -1.02.
- Penn State: The nadir of Joe Paterno's tenure was bad, with 2003 and 2004 going below -2 in both seasons. They go along with not quite as bad underachievements in 2002 and 2007.
- Clemson: This team didn't just feel like an underachiever; the numbers back it up. The Tigers have underachieved each of the past four seasons: -2.94 in 2006, -1.33 in 2007, -2.20 in 2008, and -1.23 in 2009 despite the breakthrough ACC title game appearance.
- Illinois: The Illini put a up great +2.04 in 2001's Big Ten winning year, but it's been all downhill ever since. The two years after under Ron Turner went -1.84 and -1.02. Zook isn't absolved here either as he's gone -1.88 in 2006, -1.54 in 2008 and -1.46 in 2009. If you're counting along, that means Zook has underachieved in five of his eight seasons without overachieving once.
- Michigan State: Sparty has a general air of underachievement like Clemson does, and it's earned too. John L. "Slappy" Smith turned in three in a row from 2004-06 (-1.56, -1.55, -1.14). Mark Dantonio has a pair of his own, with -1.14 in 2007 and -1.43 in 2009 sandwiching a nice +1.53 in 2008.
- Vanderbilt: Bobby Johnson's done a great job with the turnaround there for the most part, but trying to change a culture of losing into a winning one is no small task. Vandy had four underachieving seasons: 2003, 2004, 2006, and 2009. The Commodores had no overachieving seasons.
UPDATE, 2/11, 4:00 PM
As Poseur pointed out in the comments, Bill James (who first made the observation of Pythagorean expectation) considers differences luck and not over- or underachieving. I don't disagree with that to a degree. That's why I focus on larger differences than smaller ones; large differences are less likely to be due to pure luck alone.
That's also why I spent many more words on the trends that encompass the decade with relatively little on the single season anomalies: if the same luck keeps happening over time, it's probably not luck. When serially bad luck follows Ron Zook from Florida to Illinois or keeps happening at Alabama throughout the tenures of mediocre coaches who don't develop talent well, luck is probably not the whole of it.
While Poseur is right that close games can come down to the bounce of the ball or the way refs call something (see the SEC ref scandals of 2009 if you need proof), the fact of the matter is that 60 minutes of football go into a game's outcome. Unless you're getting screwed by the refs all game or there's a comical number of turnovers, the vast majority of plays in a game are determined by decisions, skill, and execution. Sure it's unlucky for a team with a +22 turnover margin for a season to be -2 in turnovers, give up its longest passing TD of the season, miss its only extra point of a season, and get stopped on fourth down on a play that works more than 90% of the time all in the same contest and lose by one point. However, when that many things (and more) go against a team at once, perhaps it didn't show up ready to play and couldn't changed the outcome through better focus and preparation. And the vast majority of close games don't come down to the bounce of a fumble or the outcome of a penalty.