Writers of Pro Football Prospectus 2008

30 Jan 2008

Fumble Luck 2007

I've been looking at tons of stats to prepare the Friday Super Bowl preview. One of them was "fumble luck," the percentage of fumbles recovered by each team. As regular readers know, fumble recovery rate will generally revert to league average over time. It's a good way to pick out possible surprise teams for 2008, so feel free to think happy thoughts if you are a fan of... the Baltimore Ravens.

Worst fumble recovery (offense)
Baltimore: 5 of 25 (20%)
Tampa Bay: 4 of 14 (29%)
Jacksonville: 4 of 12 (33%)
Minnesota: 8 of 24 (33%)

Baltimore had the second-lowest fumble recovery rate of any offense since 1978. The only offense with a worse fumble recovery rate? The 2005 New Orleans Saints, who recovered just 4 out of 23 fumbles. Irony alert: We introduced the concept of fumble luck in the New Orleans chapter of PFP 2005, right before that season.

Worst fumble recovery (defense)
Tennessee: 6 of 19 (32%)
New Orleans: 8 of 24 (33%)
Oakland: 6 of 15 (40%)
New York Giants: 8 of 20 (40%)

Best fumble recovery (offense)
Dallas: 11 of 15 (73%)
New York Jets: 11 of 16 (69%)
Green Bay: 14 of 22 (64%)

(Whoops, I had Green Bay's numbers incorrect... 14 recovered, 8 lost, not the other way around.)

Best fumble recovery (defense)
Cincinnati: 15 of 18 (83%)
San Francisco: 9 of 12 (75%)
Tampa Bay: 18 of 28 (64%)
New York Jets: 6 of 10 (60%)

These numbers don't include special teams, or a handful of aborted snaps where the quarterback was able to make a play afterwards, rather than falling on the ball.

Posted by: Aaron Schatz on 30 Jan 2008

20 comments, Last at 31 Jan 2008, 2:46am by Mr Shush

Comments

1
by Stiller Fan in Austin (not verified) :: Wed, 01/30/2008 - 12:44am

Ooh, ooh, I get to be annoying statistic police...

8/22 is much less than 62%!!!
(GB Offensive Fumble recovery rate)

Also, that's almost 3 fumbles a game, congratulations, GB!

2
by Stiller Fan in Austin (not verified) :: Wed, 01/30/2008 - 12:46am

Or not, as I was dividing by 8 instead of 16 in my head...

3
by Gerry (not verified) :: Wed, 01/30/2008 - 12:52am

Aaron-- how about throwing out five more numbers?

The offensive fumble recovery rate for the New York Giants, the offensive and defensive recovery rates for the Pats, and the league average for offensive and defensive recovery rates?

Please?

4
by coldbikemessenger (not verified) :: Wed, 01/30/2008 - 1:25am

Gerry
nyg off
12 of 26
nyg def
10 of 24
ne off
8 of 14
ne def
12 of 25
nfl ave
12 of 25
these are all fumbles, special teams, everything.

5
by Dan (not verified) :: Wed, 01/30/2008 - 1:28am

Stiller Fan (#1) is correct. 8/22=36%; to get 64% of 22 they'd have to be at 14 of 22.

6
by seth (not verified) :: Wed, 01/30/2008 - 4:06am

interesting piece. i am curious to see what "fumble luck 2006" would have looked like.

i'm surprised that the ravens had such a high number of total fumbles (25)- which players are to blame for that?

7
by Bobman (not verified) :: Wed, 01/30/2008 - 4:39am

Glad to see this because when I looked at the overall turnover picture I noticed that Indy had a freakishly low number of fumbles lost. WARNING!!!

But the good news it isn't because they fumbled a lot and recovered a non-repeatably high %, but they just didn't fumble--Hell, their pro bowl WRs accounted for most of them. Since most of the discussion here centers on fumble recoveries being random, a team that fumbles seldomly is probably actually doing well and not just lucky. Right?

8
by Dom (not verified) :: Wed, 01/30/2008 - 6:35am

#7

Yeah, fumble recovery is a fluke but forcing fumbles (or not fumbling in the first place) is a skill

9
by mawbrew (not verified) :: Wed, 01/30/2008 - 9:26am

Yikes. This doesn't bode well for that fabulous Bengal defense next year.

10
by Tom (not verified) :: Wed, 01/30/2008 - 10:06am

say what you want about TBs fumble recovery rate, they recovered 4/4 fumbles against the redskins :-/

and also I dont see the point in splitting fumble luck into offense and defense since fumbles are equally as valuable on offense as they are on defense.

11
by Fred (not verified) :: Wed, 01/30/2008 - 10:18am

#1: perhaps it should be 14/22 (64%)

12
by Statistics (not verified) :: Wed, 01/30/2008 - 11:58am

#10: Perhaps the reason for the offense/defense split is that, while fumble recovery may regress heavily to the mean, that mean is different for offense and defense. If so, simply adding up those numbers would lose information. To combine them accurately, the stats would need to be represented differently, for example as z-scores.

13
by Wanker79 (not verified) :: Wed, 01/30/2008 - 12:10pm

Best fumble recovery (offense)
New York Jets: 11 of 16 (69%)

Best fumble recovery (defense)
New York Jets: 6 of 10 (60%)

Egads!

14
by Independent George (not verified) :: Wed, 01/30/2008 - 1:13pm

Do offensive recoveries include fumbled snaps? If memory serves me right, those are the only class of fumbles predominantly recovered by a single side (offense).

For reasons mentioned by others, I wonder whether differential against mean is a more pertinent number than percentages. For example, a team that recovered 2 of 6 fumbles is far less likely to improve than a team that recovered 8 of 24.

15
by Will Allen (not verified) :: Wed, 01/30/2008 - 1:27pm

Always nice to see reason for hope. If the Vikings can improve ball security, and get luckier, they may be able to overcome a tougher schedule. 8-8 again, I suppose. Sigh.

16
by fillylabinga (not verified) :: Wed, 01/30/2008 - 3:17pm

What is the average points per possession in the NFL?

Assuming the league average fumble recovery rate is 50% and that Baltimore cost themselves 7-8 possessions (we don't know their defense recovery rate) then I'd like to know roughly how many points that would translate to.

It seems like a rate like that could have cost them a win or two. And maybe a team like the Jets with an especially high rate got an extra win.

17
by Shawn (not verified) :: Wed, 01/30/2008 - 5:44pm

Why don't you include aborted snaps where the Quarterback was able to make a play afterwards? I would think how the ball bounces there is just as random as any other fumble. Just wondering the exact reasoning behind that choice.

18
by tgt (not verified) :: Wed, 01/30/2008 - 7:06pm

#6

Everybody.

19
by jimm (not verified) :: Wed, 01/30/2008 - 8:07pm

Hey Will,

No wonder you figure 8-8 if you go back to the 1999 season, the Vikes are 72-72 since then.

20
by Mr Shush (not verified) :: Thu, 01/31/2008 - 2:46am

Not all fumbles are created equal. If the quarterback takes a thirty-eight step drop then bobbles the ball backwards through his legs before standing around looking vacant, a defender is likely to recover. Hence the 2005 Saints.