Blog Entry

SEC math says back Bulldogs, doubt Tigers

Posted on: July 5, 2011 5:53 pm
Edited on: July 5, 2011 6:22 pm

Posted by Jerry Hinnen

Click over to the "expanded" version of the Major League Baseball standings here at CBSsports.com, and you'll see something interesting: Each team's record in one-run games.

Even 10 years ago, casual baseball fans would have shrugged at those records every bit as forcefully as they would have at "record in day games played west of the Mississippi River in which both starting pitchers wore mustaches." But thanks to baseball's stats revolution, even your average CBSSports.com-reading seamhead likely knows that over 162 games, every team's record in such close games will gravitate to .500.

This is an outgrowth of Bill James' pythagorean theorem for baseball, which, if you 've never heard of it, isn't nearly as complicated as its name might make it sound; the idea is simply that total runs scored and allowed (i.e., winning by many runs rather than just one) is a better indicator of future performance than straight win-loss record.

And though college football isn't nearly as stats-obsessed a sport as baseball has become, concepts like these are hardly new to dedicated followers of the pigskin, either. Numbers-driven magazine guru Phil Steele has been tracking "net close wins" for years, finding that teams that win or lose an unusually high number of one-possession games one season tend to lose or win a corresponding number the next season. (The current poster children for this phenomenon are the Iowa Hawkeyes, who lost four games in 2008 by a total of 12 points, went 11-2 in 2009 by winning four games by a total of eight points, then slipped back to 8-5 last year with all five losses coming by seven points or less.)

One Alabama blog, RollBamaRoll, has taken the next step where the SEC is concerned, actually performing the Pythagorean calculations for the 2010 SEC conference season. Though eight games is a tiny sample size for this kind of statistical work, the same calculations predicted (or would have) the downfall of such notable disappointments as 2005 Tennessee, 2000 Alabama, and 2009 Georgia.

So what do these approaches have to say about the SEC in 2011? Several things:

Georgia should be taken seriously in the East. Both Steele and the pythagorean wins agree: the Bulldogs were the unluckiest team in the SEC last season. Mark Richt's team suffered a league-high four "net close losses," and per their points scored/allowed should have won nearly two more games than they did in 2010.

Combine better fortune in competitive games with the Bulldogs' manageable schedule, and the numbers say Georgia should be poised to take a big step forward in 2011. (Steele pegs them as this year's East champions.) If they don't, the question has to be asked: if Richt can't engineer a turnaround this year, when can he?

Auburn is due for a sizable tumble. The next team to win a national championship without a healthy dose of luck will be the first, but Auburn might have enjoyed a little more than most last season; its seven net close wins were the highest in the nation, according to Steele. The pythagorean wins marked them as overachievers by nearly 2.5 games in SEC play alone. In other words, Gene Chizik and company shouldn't expect quite so many friendly bounces of the ball in 2011--and should in fact expect the opposite.

Of course, the numbers can't account for the expertise of Gus Malzahn or the fine recruiting classes assembled under Chizik's watch. But it's safe to say that between less good fortune, the Tigers' massive losses, and a brutal schedule, another top-25 season on the Plains will have been earned.

LSU remains the ultimate wild card. Steele tabulates the Bayou Bengals at five net close wins for 2010 -- usually an indicator of an impending backslide. But thanks to blowouts of Vanderbilt and Mississippi State, the pythagorean wins saw LSU as only slight overachievers in 2010, and (as we've noted before) Les Miles has an unusual knack for late-game decision-making that's given him a 22-9 record at LSU in close games. (Is it the grass?)

In other words, LSU could see Miles' dice-rolls come up snake eyes and the bottom drop out. They could continue to ride the Mad Hatter's hot streak back to a BCS bowl. Any and all possibilities seem to be in play.

Mississippi State may have to run to stay in the same place. With Dan Mullen still in Starkville and plenty of starters returning on both sides of the ball, State may seem poised to take the next step and challenge for a West championship. But there's also some indications the Bulldogs weren't quite as good as their 9-4 record last fall might indicate. Despite going 4-4 in league play they were outscored by 30 points over those eight games, making them the SEC's second-biggest overachiever according to pythagorean wins. And while Steele's net close wins indicator doesn't feel strongly about them, his magazine does note that State's average yardage margin of -36.5 yards per SEC game was third-worst in the conference.

Steele also recently introduced a new metric which shows that teams that take a big leap forward (or backward) over (or under) the baseline of their previous two seasons usually -- though not always -- regress back towards their previously-established mean. Aside from Auburn, no team in the SEC fits that profile better than the Bulldogs.

No reason here to not buy Alabama or South Carolina. Though the above "slipping and sliding" Steele metric is mildly doubtful of Carolina's ability to maintain last year's gains, neither presumptive divisional favorite has anything to worry about from this statistical perspective. In fact, thanks to several blowout wins and their losses to LSU and Auburn by a combined four points, Alabama was the second-most unfortunate team in the league last year behind Georgia.

If the Tide get a handful more breaks and have the defense we're all expecting? Look out.

Category: NCAAF
Comments

Since: Jul 6, 2011
Posted on: July 6, 2011 3:16 pm

SEC math says back Bulldogs, doubt Tigers

If you guys like this kind of math and coach analysis, you have to check out:

These guys specialize in this kind of stuff. Definitely take a look at just how good Mark Richt and Les Miles have been.

They are ranked the #8 and #9 coach respectively. Great stuff.

Since: Aug 9, 2008
Posted on: July 6, 2011 12:47 pm

SEC math says back Bulldogs, doubt Tigers

Lets say UGA loses the 'tough games' (Boise, South Car., UF, Miss St) this year.  they can still have an 8-4 season.  Honestly, they will win a few of those 'tough games'  I got them at the least spliting the first two and the last two so that brings them up to 10-2 which is realistic.  But, to be totally honest, I think the only loss will be to S.C. this year and the Dawgs will go 11-1.

Since: Sep 12, 2008
Posted on: July 6, 2011 12:46 pm

" HOT " seat

. . . gotta beat boise state in the GEORGIA dome . . .

Since: Oct 17, 2006
Posted on: July 6, 2011 11:23 am

SEC math says back Bulldogs, doubt Tigers

Go Dawgs! They are certainly a lot better than their end of season results the last three years. This team is deep and there are plenty of unknowns who will be making a name fo rthemselves this season. I expect an SEC East Title bid. Go Dawgs!

Since: Jan 3, 2010
Posted on: July 5, 2011 11:19 pm

SEC math says back Bulldogs, doubt Tigers

Ease up. Nothing wrong in trying to statistically predict the future.  Economists do this everyday.
Sounds like an interesting theory.

Since: Dec 7, 2010
Posted on: July 5, 2011 11:08 pm

SEC math says back Bulldogs, doubt Tigers

Blah, Blah, Blah. Let's play the season and then pontificate on what went right and what went wrong.

The views expressed in this blog are solely those of the author and do not reflect the views of CBS Sports or CBSSports.com