An Argument against Fumble Recoveries and Luck

by Jerome's Friend

Throughout time immemorial humanity has attributed unexplainable phenomena to magic, God, superstition, or luck. Applications to the insanities that occur in football are no different.  Footballs veer ever so slightly left of the upright. There are obstructed views. Blown calls. All of these things and more we commonly attribute to the quirks of fate, or, luck. That is, until we fully understand what’s actually going on.

I think we do this when trying to understand fumble recoveries.  Check out this series of articles by Brent over at

What’s the Deal with Fumbles? (Should we worry about Bryce Brown?)

Fumble Luck… Again

Forced Fumbles Skill and Give/Take Recovery Rates

He has expertly provided a strong case (which supports the work of several others) that fumble recoveries are a function of chance.  Despite his evidence, and despite my own research, I still find it hard to believe.  I’ve performed dozens of regressions (at least 60, both linear and logistic) comparing this statistic to that, from this year to that, hoping to find some relationship that explains the seemingly random nature of fumble recoveries.  Do they correlate to yards?  No.  Do they correlate to wins?  No.  The overall strength of a defense?  Sacks?  All no. I am usually the first to say, “Well if the numbers say it, then it must be true.”  And these numbers indeed suggest fumble recoveries are a function of luck.  So what’s my beef?

Maybe it’s philosophical.  Primarily, I’m a believer that there is no such thing as luck or chance.  I believe we make our own luck, be it good or bad.  And fumbles, be they accidental or forced, and recoveries, by the offense or defense, must be a product of concentration and skill rather than the whims of chance.  But, for the life of me, I just can’t accept the latter.

I think fumble recoveries are the product of the type of fumble.  Fumbled snaps (lack of concentration on the part of the center or quarterback) are more likely to be recovered by the quarterback, right? At that point there are fewer defenders around the ball.  Fumbles by running backs (lack of concentration and technique on the part of the quarterback or running back) are likely to be recovered by the offense if it occurred on the offense’s side of the line of scrimmage, and likely to be recovered by the defense if the fumble (forced or not) occurred on the defense’s side of the line of scrimmage.  I would think that’s what the odds would say. Right?

As it turns out, I think they do.  Luckily for me, Chase Stuart over at categorized the types of fumbles and the types of recoveries for the 2000-2011 seasons.  Here is his chart:

Fumble Recovery Rates - FootballPerspective

His chart tells us how each type of fumble was either recovered by the defense (DEF), offense (OFF), the fumbler (RBF), or whether the ball went out of bounds (OOB).  At first glance, note the prevalence of the red (DEF) color.  For the first four bars, all quarterback related, fumbles are most likely recovered by the offense.  This makes sense, since the fumble most likely occurred on the offense’s side of the line of scrimmage and should most likely be recovered by an offensive player.  The exception is sacks, which also seems to make sense (to me anyway).

The last two bars illustrate events that most likely occur on the defense’s side of the line of scrimmage. During these fumbles, the number of defensive players surrounding the ball in the defensive zone (swarming?) probably outnumbers the number of offensive players.  Again, this seems to make sense.  (Note: this is my interpretation of the data, not necessarily Stuart’s.)

So what can this tell us about luck?  My hard stance: when a football veers just left of the upright, it’s because of a poor snap, hold, or kick; when there is an obstructed view, someone is out of position; a blown call is a result of poor judgment.  Lastly, a fumble recovery is a product of the type of fumble, the location of which is a function of some player’s concentration and skill (or in some cases, lack thereof). Is all of this definitive proof that fumble recoveries are not a product of chance?  Probably not, but it’s still a valid argument.  Measured in the aggregate, fumbles and fumble recoveries appear to be random products of luck.  But when broken down and dissected, when we calibrate the microscope a bit deeper, we can see things a little differently, and perhaps clearly.