Writers of Pro Football Prospectus 2008

05 Aug 2013

Maximizing Fumble Luck

Painting with a broad brush, fumble recoveries are a fairly random event. Offenses tend to recover approximately 50 percent of fumbles. The Mathlete at MGoBlog refines that analysis by breaking down college football fumble events into six distinct categories, and finds that the defensive recovery rate varies considerably depending on how and where the ball hits the ground.

Posted by: Brian Fremeau on 05 Aug 2013

21 comments, Last at 10 Feb 2016, 4:39pm by amyahlom


by Arkaein :: Mon, 08/05/2013 - 12:54pm

This all looks pretty similar to what FO wrote about fumble recoveries 8 years ago. From the FO Basics page:

"Football Outsiders stats treat all fumbles as equal, penalizing them based on the likelihood of each type of fumble (run, pass, sack, etc.) being recovered by the defense."

I think FO went into more detail, but it's apparently in PFP 2005 (which I have, but not with me at the moment) and not on the website. It's unfortunate, because I've wanted to reference the info in discussions about fumble recovery being random, since there is a bit more nuance to the subject.

by Nopal (not verified) :: Wed, 12/18/2013 - 10:38pm

"Football Outsiders stats treat all fumbles as equal, penalizing them based on the likelihood of each type of fumble (run, pass, sack, etc.) being recovered by the defense." Thanks on it.

by ebongreen :: Mon, 08/05/2013 - 1:13pm

It seems to me that at least one of the trends he's observing (specifically the quantity of defensive recoveries beyond the LOS) can be explained pretty readily by the proportion of players from each team likely to be in the area. Defensive players should generally outnumber offensive players beyond the LOS, sometimes dramatically so - e.g. max-protect pass vs. four-man pass rush.

A fumble downfield in those circumstances should present pretty bad odds for an offensive recovery.

by Sisyphus :: Mon, 08/05/2013 - 1:26pm

While you are correct about the numbers the offensive player is the first one who knows that a fumble has occurred giving him a critical lead in the response time. The question that I think might be interesting is what is the recovery rate by the person who fumbled as opposed to a different offensive player. That I suspect becomes relatively higher as you move further past the line of scrimmage as offensive players are looking at defensive players and defensive players are, for the most part, looking at the ball carrier.

by Guido Merkens :: Mon, 08/05/2013 - 4:50pm

That would be an interesting breakdown. On one hand, the offensive player knows he fumbled, and he's the most likely to land near the ball. On the other hand, he probably just fumbled because he got tackled, so he might be the only guy on the ground, and won't be able to get to the ball as quickly if it's not near him.

by BaronFoobarstein :: Mon, 08/05/2013 - 6:04pm

Facing also matters. On a strip sack there are probably more offensive players nearby, but five of them are looking the wrong way.

by Karl Cuba :: Mon, 08/05/2013 - 1:39pm

- Didn't Chase Stuart do a study on fumble recovery rates by type of location a few years ago that showed different rates of loss depending on where they occurred?

- Wouldn't it make sense to remove aborted snaps and botched hand offs from the data? They are pretty obviously not a skill you can credit to the defense.

- Someone raised Charles Tillman's propensity to force fumbles and he dismissed it as luck. By all means suggest he is an outlier and that most NFL players don't possess that skill but anyone who has ever seen him play knows that Tillman is more than just lucky.

by White Rose Duelist :: Mon, 08/05/2013 - 5:59pm

Re: Tillman - Forcing fumbles is a skill. Recovering them is not. So it is entirely reasonable that Tillman was better than an average defender at making a ball carrier lose his grip.

by LionInAZ :: Mon, 08/05/2013 - 8:09pm

Since Tillmam's forced fumbles come from punching the ball out of a player's arms, I've always wonder how many of them are clean, i.e., from hitting the ball, rather than punching the ball carrier's arms, which should be a penalty.

by DEW (not verified) :: Tue, 08/06/2013 - 3:27pm

Given that virtually all if not actually all of those punches are shown on replay after the play, I'd say very few of them come from punching the player's arms. I'd expect arm-punching to actually occur on the plays where he attempts to punch the ball loose but fails to do so--in part because his punch missed the ball and was blocked by the arm instead.

by snoopy369 :: Mon, 08/05/2013 - 1:46pm

Tenet. Tenet. Tenant is either someone who pays you rent, or the tenth Doctor short an n.

by Independent George :: Mon, 08/05/2013 - 1:58pm

But % of fumbles are recovered by the defense when the QB runs into his offensive lineman?

by akn :: Mon, 08/05/2013 - 4:39pm

I am desperately hoping there are enough butt fumbles in the coming years to turn them into meaningful statistic.

by Guido Merkens :: Mon, 08/05/2013 - 4:53pm

Last season, 0.00147%. But last season may have been an outlier.

by steveNC (not verified) :: Tue, 08/06/2013 - 10:24am

*Butt last season may have been an outlier.

by hassan :: Tue, 12/31/2013 - 5:53am

Thank you for this beautiful theme friend
Best flash games

by hassan :: Tue, 12/31/2013 - 6:39am

Thank you for this beautiful theme friend
Best flash games

by http://www.sia.fr/fr/airmax-fr.asp (not verified) :: Thu, 01/09/2014 - 12:13am

On ignorait mardi matin si le tribunal rendrait sa d

by burberry outlet (not verified) :: Fri, 01/10/2014 - 5:58pm

quelques voitures am

by vivian :: Sat, 12/06/2014 - 12:26pm

hehe nice post !

by amyahlom :: Wed, 02/10/2016 - 4:39pm