Ah Yes, That OTHER Framegate Urban Myth - - Fumbles

[Galeb]

I think the situation and location of fumbles would have a significant impact on who recovers. A large number of fumbles occur because of bad snaps. In those cases, the QB, holder or punter is much more likely to recover the fumble. I suspect that fumbles that occur more than 5 yards past the LOS are much more likely to be recovered by the defense.

 

[Galeb]

Interesting, I must have reversed it in my strip sack comment earlier. I knew that it was one of the few that is slanted in one direction, but obviously I got the direction wrong.

A significant portion of QB fumbles are on bad snaps and fudged hand offs.

 

[Oswlek]

A significant portion of QB fumbles are on bad snaps and fudged hand offs.

So it could be that I'm right about "strip sacks" but wrong about "QB fumbles." Or maybe I'm just wrong overall. Either way, I edited the post above.

 

[ivanvamp]

Just using this year's numbers as data, culled from espn's team stats (2015 NFL Team Defense Stats - National Football League - ESPN):

Team Forced Fumbles
- Forced Fumbles: 226
- Recovered: 140
- Recovery % of forced fumbles: 61.9%

So the team forcing the fumbles has recovered them at a rate much higher than 50%.

Previous years:
2015: 207 FF, 140 rec, 61.9%
2014: 451 FF, 305 rec, 67.6%
2013: 435 FF, 308 rec, 70.8%
2012: 472 FF, 328 rec, 69.5%
2011: 447 FF, 303 rec, 67.8%
2010: 500 FF, 347 rec, 69.4%

So pretty clearly it appears (if I'm remotely reading these numbers right) that the team forcing the fumbles will tend to recover them more than 50% of the time. Much more. But maybe I'm not understanding the data correctly.

 

[Galeb]

Just using this year's numbers as data, culled from espn's team stats (2015 NFL Team Defense Stats - National Football League - ESPN):

Team Forced Fumbles
- Forced Fumbles: 226
- Recovered: 140
- Recovery % of forced fumbles: 61.9%

So the team forcing the fumbles has recovered them at a rate much higher than 50%.

Previous years:
2015: 207 FF, 140 rec, 61.9%
2014: 451 FF, 305 rec, 67.6%
2013: 435 FF, 308 rec, 70.8%
2012: 472 FF, 328 rec, 69.5%
2011: 447 FF, 303 rec, 67.8%
2010: 500 FF, 347 rec, 69.4%

So pretty clearly it appears (if I'm remotely reading these numbers right) that the team forcing the fumbles will tend to recover them more than 50% of the time. Much more. But maybe I'm not understanding the data correctly.

That's good info. These numbers are for forced fumbles only, so botched snaps and handoffs wouldn't be included.

 

[ivanvamp]

That's good info. These numbers are for forced fumbles only, so botched snaps and handoffs wouldn't be included.

Ok that's helpful. So maybe recovering fumbles is around 50% overall.

 

[VrabelMayeWin]

That's good info. These numbers are for forced fumbles only, so botched snaps and handoffs wouldn't be included.

If the numbers come out close to 50% that would make sense as those QB related ones would mostly be fallen on by the QB. It makes sense that other types of fumbles would be more often recovered by the defense.

 

[Concussion Protocol]

Just using this year's numbers as data, culled from espn's team stats (2015 NFL Team Defense Stats - National Football League - ESPN):

Team Forced Fumbles
- Forced Fumbles: 226
- Recovered: 140
- Recovery % of forced fumbles: 61.9%

So the team forcing the fumbles has recovered them at a rate much higher than 50%.

Previous years:
2015: 207 FF, 140 rec, 61.9%
2014: 451 FF, 305 rec, 67.6%
2013: 435 FF, 308 rec, 70.8%
2012: 472 FF, 328 rec, 69.5%
2011: 447 FF, 303 rec, 67.8%
2010: 500 FF, 347 rec, 69.4%

So pretty clearly it appears (if I'm remotely reading these numbers right) that the team forcing the fumbles will tend to recover them more than 50% of the time. Much more. But maybe I'm not understanding the data correctly.

How does Dallas have 1 FF with 1 recovery? We know that they stripped Brady and it was recovered by Edelman. That doesn't make sense.

 

[ivanvamp]

To show you how lucky the 2011 New York Giants were, consider the FF stats I listed above. Then consider their fumble luck in their playoff run.

Wild Card Round vs Atl: No fumbles, no forced fumbles at all in the game.

Divisional Round vs GB:
- NY had no fumbles
- GB fumbled 3 times. NY recovered all 3.

Conference Championship vs SF:
- NY fumbled 1 time. NY recovered the fumble.
- SF fumbled 4 times. NY recovered 2 fumbles.

Super Bowl vs NE:
- NY fumbled 2 times. NY recovered both fumbles.
- NE did not fumble at all.

So over the course of the postseason:
- NY fumbled 3 times. NY recovered 3 times.
- NY's opponents fumbled 7 times. NY recovered 5 times.

So overall, there were 10 fumbles in NY's playoff games. NY recovered 8 of the 10. Small sample size, etc., but that's an awful lot of luck that went their way.

 

[ivanvamp]

How does Dallas have 1 FF with 1 recovery? We know that they stripped Brady and it was recovered by Edelman. That doesn't make sense.

No idea.

 

[Hoodie]

Deflategate!!!!!!!!!

The Patriots MUST be doing something to the balls that make them easier to recover!!!!!!!!

Not only are the Patriots doing something to make the balls easier for them to recover, they are making the balls harder for other teams to recover! I'm not sure what it is but Harbaugh and Tomlin will be happy to tell you.

 

[Oswlek]

Just using this year's numbers as data, culled from espn's team stats (2015 NFL Team Defense Stats - National Football League - ESPN):

Team Forced Fumbles
- Forced Fumbles: 226
- Recovered: 140
- Recovery % of forced fumbles: 61.9%

So the team forcing the fumbles has recovered them at a rate much higher than 50%.

Previous years:
2015: 207 FF, 140 rec, 61.9%
2014: 451 FF, 305 rec, 67.6%
2013: 435 FF, 308 rec, 70.8%
2012: 472 FF, 328 rec, 69.5%
2011: 447 FF, 303 rec, 67.8%
2010: 500 FF, 347 rec, 69.4%

So pretty clearly it appears (if I'm remotely reading these numbers right) that the team forcing the fumbles will tend to recover them more than 50% of the time. Much more. But maybe I'm not understanding the data correctly.

Just in this past thread we've seen two examples of ESPN being wrong in their data for 2015.

Here are a couple articles that elaborate on fumble recovery rates:

The Hidden Factors for Winning in the NFL

The definitive analysis of offensive fumbles

And then there is this article that has an Eagles bent, but states:

The long term NFL average recovery rate for “giveaway” fumbles is 52.31%. That means that an offense, on average, recovers 52.31% of the fumbles it commits.

Conversely, a defense recovers, on average, 47.48% of the fumbles the OPPOSING offense commits.

Note that those two numbers do not add to exactly 100%, which I assume is due to fumbles that go out of bounds and are therefore recovered by neither team.

 

[Galeb]

Just in this past thread we've seen two examples of ESPN being wrong in their data for 2015.

Here are a couple articles that elaborate on fumble recovery rates:

The Hidden Factors for Winning in the NFL

The definitive analysis of offensive fumbles

And then there is this article that has an Eagles bent, but states:

The numbers he posted are for forced fumbles. As stated before, that doesn't include botched snaps and hand offs.

 

[ReverendMaynard]

I think there's some confusion about how fumble recoveries are viewed by statisticians here.

It's not that all fumbles are considered a 50/50 proposition, as that would be demonstrably false. There are certain types of fumbles more likely to be recovered by the offense vs. others more likely to be recovered by the defense. What the statisticians do say about fumbles is that recovering them at a higher rate than the averages suggest is not a repeatable skill. A team that recovers a larger percentage of fumbles one season is no more likely to recover a larger percentage the next season than a team that recovered a smaller percentage. Since it's not a repeatable skill, and offers no predictive power for the next game/season, teams are not given credit (or penalized) for fumble recoveries in statistics that are meant to be predictive (such as DVOA).

 

[Galeb]

Are Buttfumbles considered forced?

 

[Oswlek]

The numbers he posted are for forced fumbles. As stated before, that doesn't include botched snaps and hand offs.

The numbers are wrong regardless. We know of two fumbles just for the Patriots in 2015 that aren't being counted. When a single team in a single half season has that much error, the entire data set is corrupted.

 

[Galeb]

The numbers are wrong regardless. We know of two fumbles just for the Patriots in 2015 that aren't being counted. When a single team in a single half season has that much error, the entire data set is corrupted.

I doubt many large data sets exist that don't contain errors. That said, I can't speak for the accuracy of said numbers other than the glaring mistake already pointed out.

 

[Oswlek]

I think there's some confusion about how fumble recoveries are viewed by statisticians here.

It's not that all fumbles are considered a 50/50 proposition, as that would be demonstrably false. There are certain types of fumbles more likely to be recovered by the offense vs. others more likely to be recovered by the defense. What the statisticians do say about fumbles is that recovering them at a higher rate than the averages suggest is not a repeatable skill. A team that recovers a larger percentage of fumbles one season is no more likely to recover a larger percentage the next season than a team that recovered a smaller percentage. Since it's not a repeatable skill, and offers no predictive power for the next game/season, teams are not given credit (or penalized) for fumble recoveries in statistics that are meant to be predictive (such as DVOA).

Both propositions are true. Different fumble types have different recovery rates and fumbles overall are considered a 50/50 proposition. A single team's data in a single season could slant in the direction of one fumble type, making it more likely that their expected recovery rate diverges from 50%, but you'd have to show that reasons exist to believe the same slant will continue into next year (or the second half of the year) for future expectations to stray from 50%.

 

[Koma]

To show you how lucky the 2011 New York Giants were, consider the FF stats I listed above. Then consider their fumble luck in their playoff run.

Wild Card Round vs Atl: No fumbles, no forced fumbles at all in the game.

Divisional Round vs GB:
- NY had no fumbles
- GB fumbled 3 times. NY recovered all 3.

Conference Championship vs SF:
- NY fumbled 1 time. NY recovered the fumble.
- SF fumbled 4 times. NY recovered 2 fumbles.

Super Bowl vs NE:
- NY fumbled 2 times. NY recovered both fumbles.
- NE did not fumble at all.

So over the course of the postseason:
- NY fumbled 3 times. NY recovered 3 times.
- NY's opponents fumbled 7 times. NY recovered 5 times.

So overall, there were 10 fumbles in NY's playoff games. NY recovered 8 of the 10. Small sample size, etc., but that's an awful lot of luck that went their way.

While not related to fumbles, something similar happened in 2007, also: Eli led the league in interceptions in the regular season. In 4 playoff games, he threw 1 INT and that was in the SB. In the 3 NFC playoff games, including 2 road games, he threw no INTs.

 

[Bobsyouruncle]

Actually, with the sole exception of 2013 the Patriots are consistently among the league leaders in fewest total fumbles:

Team Fumbles and Fumbles Lost: 2015 NFL Season

Sure, and the original claim passed on about deflategate used fumbles lost over many years to claim NE was a huge outlier.

When it was redone for all fumbles it showed what this says, the Patriots are pretty good but not necessarily the best.

And that sort of makes the whole point of that "analysis" moot. If they want to claim NE must be cheating by being 3rd best at fumbles, then the two teams ahead of them must be cheating too. But that's not what they wanted to claim, so it was nonsense.