College football turnovers luck in 2013

Sometimes the ball lies, Rasheed, at least in football.

We can go pretty deep into the numbers to gauge who's good and who's bad, who's efficient and who's not. But the ball is still pointy, and it bounces the way it chooses to bounce sometimes. More than in any other sport, you have to be both lucky and good to win a lot of football games.

In my offseason preview series each year, I've tried to get at this idea by discussing turnovers luck. The concepts are pretty simple: Over time, you're going to recover about 50 percent of all fumbles, but in a given year, you might recover 70 percent, or you might recover 30. The same goes with passes defensed; on average, you can expect to intercept about 22 percent of the passes you defense. (Passes defensed = interceptions + break-ups.) This is a bit mushier a concept, as we'll discuss below, but over time a particularly butter-fingered year will be balanced by a sticky one.

I'll be using the concept of turnovers luck throughout this year's offseason preview series (which, no lie, begins in under a month), but I thought I would break down all the numbers for you in one place first.

Fumbles

For whom did the ball bounce nicely in 2013? Oklahoma, for starters. The Sooners were quite good (lucky) at falling on the bouncy, pointy ball this past season, and it helped them to win an improbable 11 games. They were 3-0 in one-possession games and 7-0 in games decided by 14 or fewer points. They recovered all three fumbles against Oklahoma State (two of theirs, one of OSU's) in a game in which they trailed until the final 20 seconds, and they recovered all four fumbles against Alabama in the Sugar Bowl, taking one of Bama's back for a touchdown. OU played well in tight games, but the Sooners got help. (You might want to remember that when you're getting ready to put OU in your preseason top 5 this coming year. You know you're going to, and you know you're probably making a mistake.)

On the flipside, Akron managed to go 4-3 in one-possession games despite recovering fewer than one of every three fumbles. With OU bounces, the Zips might have gone 7-5.

Passes Defensed

As mentioned above, this concept is a bit mushier. Not every defensed pass is created equally. For one thing, not every defensed pass is RECORDED equally; one scorekeeper's definition of a break-up might not be another's. But we're not talking about QB hurries here (hurries are worthless because one scorekeeper calls almost anything a hurry, and another won't document any at all); the PBU is, from what I've been able to gather, relatively consistent from school to school and year to year. It's consistent enough, anyway, to use here.

There is also a skill aspect to defensed passes. Some defenders are simply more likely to catch a ball than others. The "They can't catch passes; that's why they're on defense" theorem isn't always true. But even a player like David Amerson regressed toward the mean over time. In 2011, Amerson rode a wave of good fortune, intercepting 13 passes while breaking up only five. On average, you pick off 22 percent of your defensed passes, meaning 18 passes defensed should have netted a normal player four picks. In 2012, he defensed 17 passes, almost the same as the year before. But he picked off only five (29 percent). Maybe his hands were better than those of the normal DB, but 2012 suggested that 2011 was pretty lucky.

With that in mind, I think we can still call certain outliers lucky or unlucky in 2013. South Carolina's offense had only four passes intercepted to 27 PBUs; that INT total probably should have been closer to about seven. Meanwhile, the Gamecocks also picked off 18 passes to 33 PBUs. On average, that would be closer to 11. I think we can safely say there was some good fortune there. And on the other side of the table, it's safe to say that Kansas State (13 offensive INTs to 17 PBUs) and Utah (three defensive INTs to 52 PBUs) were pretty damn unlucky.

"Diff. from Avg." is calculated by looking at the difference between a team's %INT and the national average of 22%. So for South Carolina, it was as follows: (22% - 12.9%) + (35.3% - 22%) = +22.4%.

Overall Turnovers Luck

While acknowledging that this isn't a 100% perfect concept -- that offenses recover 51% of fumbles, and that not every PBU is created equally -- here's a look at who most likely benefited from some serious luck-of-the-bounce in 2013.

Below you'll find both a team's turnover margin and an adjusted turnover margin based on what might have happened with normal fumbles luck and 22% PD rates for both offense and defense.

You'll also find a "Turnover Luck per Game" number. Here's how that's derived: On average, a turnover is worth about five points when you look at field position lost (by the offense) and gained (by the defense). So if you look at the difference between, say, Houston's turnover margin and adjusted turnover margin, you see the Cougars were +25 when they should have been about +3.8. They were +21.2 turnovers. 21.2 times 5 = about 106 points. 106 points divided by 13 games = about 8.2 points per game.

The five luckiest teams in the country when it comes to turnovers were Houston, Ball State, Mississippi State, Oregon State, and Florida State. Unluckiest? Indiana, Southern Miss, Troy, Rutgers, and Utah. Luck didn't necessarily play the role one might think with these teams. For a team like MSU, it helped a lot; the Bulldogs went 4-1 in one-possession games. Meanwhile, Southern Miss and Indiana were each 0-2. But Houston was just 1-4 in such games; turnovers helped the Cougars stay just close enough for their general inexperience to hurt them. And in the end, the five luckiest teams went 10-10 in one-possession games while the five unluckiest went 11-11.