Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2018 Free Agency Cost-Benefit Analysis

Is Kirk Cousins the best free-agent quarterback in recent memory? Should Trumaine Johnson or Malcolm Butler have gotten the larger contract? And what makes a free-agent contract good or bad, anyway?

12 Sep 2014

Varsity Numbers: Follow the Bouncing Ball

by Bill Connelly

From the start of the 2013 season through two weeks of 2014, college football players have carried the ball 117,688 times on non-special teams plays -- 78,851 via rush or sack and 38,837 times via reception. On 2,403 occasions (2.0 percent of the time), the ballcarrier fumbled; 1,175 times, the offense recovered, and 1,228 times, the defense recovered.

That's a defensive recovery rate of 51 percent on average, and every time I've ever looked at fumbles, that rate has been between 49 and 51 percent. This results in something pretty convenient: You can basically say that fumbles are 50-50 occurrences. And if you do that, you can pretty easily count up the fumbles that happened during a team's season, look at the number the team recovered, and determine a rough "fumble luck" effect. I do it all the time. It's great.

There's one problem, of course, and you notice it if you watch a lot of football. A lot of fumbles take place because of a bobbled snap or a poor exchange between the quarterback and running back. Anecdotally speaking, it always seems like offenses recover a higher percentage of those. Meanwhile, if a runner is stripped at the end of a 30-yard gain, there might not be any of his teammates close to the ball, and it seems like the defense recovers a higher percentage of those. Is this borne out in the numbers?

The short answer: yes.

Fumbles and rates, non-sack rushes and receptions (2013-14)
Yardage Range Count Fumbles Fumbles Lost Fum% Lost%
-10 or fewer 229 156 51 68.1% 32.7%
-5 to -9 851 225 84 26.4% 37.3%
-2 to -4 4098 287 102 7.0% 35.5%
-1 to 1 16529 335 152 2.0% 45.4%
2 to 4 24067 213 126 0.9% 59.2%
5 to 7 13283 131 92 1.0% 70.2%
8 to 10 6395 97 58 1.5% 59.8%
11 to 14 4336 52 36 1.2% 69.2%
15 to 20 2525 46 26 1.8% 56.5%
21 or more 2501 33 22 1.3% 66.7%
TOTAL 74814 1575 749 2.1% 47.6%

If you exclude sacks and look only at intentional runs (and catches), you see pretty clearly that if a fumble happens in the backfield, where defenders are less likely to be hanging around en masse, offenses recover the ball nearly two-thirds of the time. Meanwhile, if the fumble happens more than 10 yards downfield, where blockers and stray teammates are less likely to be hovering, defenders recover the ball nearly two-thirds of the time. In all, the recovery rates correlate pretty strongly overall with where the fumble occurs.

As one would expect, sacks are a different story.

Fumbles and rates, sacks (2013-14)
Yardage Range Count Fumbles Fumbles Lost Fum% Lost%
-10 or fewer 718 123 75 17.1% 61.0%
-5 to -9 2102 186 91 8.8% 48.9%
-2 to -4 938 70 46 7.5% 65.7%
-1 to 0 279 30 18 10.8% 60.0%
TOTAL 4037 409 230 10.1% 56.2%

So basically, the more fumbles that are caused by sacks, or the more fumbles that occur downfield, the higher the recovery rate should be for the defense.

This is interesting, in part because it just is, and in part because I'm trying to figure out what to do with turnovers when it comes to my ongoing Five Factors series at Football Study Hall. (Well, "ongoing" might be pushing it since I've drowned myself in other projects and haven't written a Five Factors post since May. "Intended ongoing," maybe.) Football games are basically determined by efficiency, explosiveness, field position, finishing drives, and turnovers. But while I've been able to strip apart the different factors that go into those first four, determining what's (accurate) luck and what's controllable in the turnover realm is tougher. This at least allows us to be a little bit more accurate in talking about the lucky bounces part, even if it means the whole "fumbles are 50/50 propositions" thing is harder to get away with.

This week at SB Nation

Here are some of the pieces I wrote at SB Nation this week.

The Oregon Ducks vs. power teams narrative is dead (again)

The Numerical: How Notre Dame won by *that* much, plus more great stats

Let's pretend Georgia vs. South Carolina is a normal football game

Let's figure out Week 3's best non-Georgia vs. South Carolina games
College football projections: Week 3 F/+ picks

3 big questions for UCLA-Texas. Does either team have answers?
UCF at Missouri Preview: Tigers have the edge, but watch the big pass plays

Meanwhile, for the soccer fans among us, I went to England in August to write a Premier League Road Trip piece, meeting with Ian Darke, Arlo White, and other soccer personalities. It was a nice football palate cleanser, and the series ran this week. Enjoy.

Part 1: Nerds and Ian Darke
Part 2: Nottingham, dirty Leeds, and regulars at the hotel bar
Part 3: Etihad, odds and ends, and Hi Ho Sheffield Wednesday
Part 4: Liverpool
Part 5: Underdogs, sleeping giants, and a Golden Lion

Posted by: Bill Connelly on 12 Sep 2014