Dropped Balls: The Patriots became nearly fumble-proof after a 2006 rule change backed by Tom Brady.

Deus Irae said:

This story has already been debunked. Thanks for bringing up, yet again, though.


First post on the board from this person is a hit piece that's been already been refuted.


I'm thinkin' there be trolls in our midst.

Click to expand...

A slow one at that
Joined: Jun 26, 2013

This theory is so stupid. They are going off fumbles lost and not fumbles. So because the Pats are good at recovering their own fumbles it's a negative.

Bobsyouruncle said:

I already saw this. The Patriots were not the best most years. Even over the period in the article they were only 3rd best in fumbles per play. Where's the article on NO and ATL who are better?

They are better than average most years but not all.

Better than average is not extraordinary. They are better than average at a lot of things, wins, points, etc. etc. So are other teams.

Click to expand...

They claim fumbles lost as total fumbles. We've lost 33 fumbles since 2010, but have 86 total fumbles. These stats include postseason too. Skewing information to further an agenda.

Also the rule was enacted for 2006, not 2007. So not sure why that article starts with 2007 and ignores 2006.

50-yard-line said:

Deus, do you have any links to the debunking? Because I'm really curious as to the explanations for this - because the premise seems ridiculous when I just look at team fumbles over the last 8 seasons.

Click to expand...

Let me just give you Unoriginal's post on the issue. There's plenty more but, even without getting into the fumbles v. fumbles lost problem, that one post alone is enough to kill the argument:

unoriginal said:

I respect the analysis but the suggested conclusion is infuriating and incomplete.

Here are the damning league rankings for the Patriots in terms of fumbles:

Code:

2003    14
2004    13
2005    12
2006    22
2007    1
2008    4
2009    4
2010    1
2011    4
2012    5
2013    24
2014    2

Clear evidence the Pats were using under-inflated footballs in every year since 2006. (Excepting 2013, when they inexplicably forgot to do so.)

Here are the unincriminating league rankings for the Colts in terms of fumbles:

Code:

2003    19
2004    4
2005    1
2006    2
2007    2
2008    2
2009    1
2010    4
2011    23
2012    14
2013    4
2014    31

In 7 of the last 8 years, the Pats put together an incredible run of rarely putting the ball on the turf. In the 7 years between 2004 and 2010, the Colts put together a similar, and more consistent run of never coughing the ball up. That both streaks coincide with both offenses running spread-out passing systems with elite QBs with famously quick releases and read progression is a meaningless correlation.

The implication that fumbles go up when you put in a bad QB (Indy 2011, Curtis Painter) or supply your QB with sub-standard WRs (NE 2013) is also a figment of your imagination.

What actually happened in 2013 is the Colts stole the Pats illegal footballs before the season and substituted their own rock-hard fumble-prone footballs in their place. The Pats detected the ruse this offseason and stole the rigged balls back.

Click to expand...

Danger Zone said:

I believe part of the debunking evidence that Deus is talking about is that the study doesn't take into consideration Fumbles Recovered.

For example, Edelman's fumble(s) (?) recovery during the AFCCG. I would not be surprised to learn that Edelman recovers more his own fumbles than any other player in the NFL.

Every Fumble Recovered is a lost data point because it doesn't get registered in the study as an actual Fumble.

Click to expand...

He updated it for all fumbles, but after his update showed the Patriots were still not the best he decided to cherry pick only non-dome teams (because they play like 4 games less outdoors) even though his data showed the same amount of dome teams above and below the median. And voila, it moves the Patriots to the top. What a clown.

Danger Zone said:

I believe part of the debunking evidence that Deus is talking about is that the study doesn't take into consideration Fumbles Recovered.

For example, Edelman's fumble(s) (?) recovery during the AFCCG. I would not be surprised to learn that Edelman recovers more his own fumbles than any other player in the NFL.

Every Fumble Recovered is a lost data point because it doesn't get registered in the study as an actual Fumble.

Click to expand...

You're telling me the guy compared fumbles we lost against the TOTAL fumbles of other teams?

Or that we fumbled almost as often as other teams, but just recovered more?

And yeah, I agree that QB is a big deal here - some QBs fumble a ton, the really good ones don't.

RayClay said:

The Patriots have won 14 of the last 18 coin flips.

Explain that.

Click to expand...

Its obviously cheating on the part of the Patriots. We cheat on the coin flip so we can get that much vaunted Defer Option.. which is exactly the scenario that 30/31 other teams would have given us anyway had they simply won the toss fairly.

Deus Irae said:

Let me just give you Unoriginal's post on the issue. There's plenty more but, even without getting into the fumbles v. fumbles lost problem, that one post alone is enough to kill the argument:

Click to expand...

Thank you - that was quite enlightening.

The "deflated" footballs add a 1 (ONE) motherf***ing extra MILLIMETER grip to the ball. Do you know how minuscule that is?! Regardless, that is under the extreme assumption they've been doing this since 2006.

Now get out of here. People like you are a cancer to the enjoyment of football.

One more to the **** list.

RayClay said:

The Patriots have won 14 of the last 18 coin flips.

Explain that.

Click to expand...

Obviously they deflated the coins too...

RayClay said:

The Patriots have won 14 of the last 18 coin flips.

Explain that.

Click to expand...

Magic. Now the NFL is going to contact Penn and Teller as part of their investigation to find out how the Pats did it.

Everybody talk about fumbles but what about dropped passes? If the balls were underinflated and easier to catch, you would think that the dropped passes rate for Pats would be among the best in the league. Nope, in the last 5 years, Pats were in the bottom ((worst) five teams three of those years.....

http://www.sportingcharts.com/nfl/stats/team-receiver-drops-percentage/2013/

2001 3.8 30/31 second best

2002 4.9% 24/32

2003 4.9% 21/32

2004 3.3% 31/32 second best

2005 4.7% 17/32

2006 5.2% 15/32

2007 5.5% 11/32

2008 5.3% 16/32

2009 6.5% 5/32

2010 7.8% 1/32 worst

2011 3.2% 32/32 best


2012 6.5% 5/32 (fifth worst)

2013 6.6% 2/32 (second worst)

2014 3.7% 27/32 (sixth best)
Facts- sucks when they get in the way of a good theory.....

Bobsyouruncle said:

I just looked at the new update, the article debunks itself. After concluding "it could not be random!" No ****, fumbles aren't random. And it's "extraordinary!"

They run fumbles lost for the last 25 years to find the best 5 year period. And..... The Patriots "extraordinary" run puts them......drum roll.....4th best! In 25 years.

No team has ever had a run like this! Except for the four others that had a run better in just the last 25 years.

Click to expand...

Well, that's not quite what statistical randomness means, though I agree that the word random here is often used to confuse people. It simply means the difference in means is statistically significant (likely at p<.05), which means that the variation in the number of fumbles (really, lost fumbles) per play is highly unlikely to be due to random variation.

The problem is that there's a ton of problems with both the dependent variable (lost fumbles per play). Notably, the result of a fumble (i.e. whether it is lost or recovered by the same team) is random, for all intents and purposes. That means the difference between the Patriots average and the league average could be that the Patriots happen to recover their fumbles at a 60% clip versus the 50% clip of the league average. Assuming the Patriots fumbled infinite times, this would fall to 50%. But because teams only fumble a handful of times each season, a random difference in the rate of fumble recovery can make all the difference here.

Once you change fumbles lost into fumbles, the Patriots are still very good but they're not impossibly or historically good or even outside an expected distribution. In fact, a lot of the data being presented is due to the bizarre methodological choice of "5 year periods," where a single outlier season can cause a team to appear in his stats up to 5 times.

The more interesting question begged from the analysis (because the causal inference drawn is total bunk and would never pass peer review, but it does get you a national outlet in a Slate thinkpiece I guess) is how the Patriots are so good at recovering their own fumbles. Obviously a large part of this is random, but there could be an underlying statistical cause. For example (this is just a musing), a strip-sack could be more likely to be recovered by the fumbling team, and the Patriots' fumbles may disproportionately come from strip-sacks.

TheBostonStraggler said:

First post? I'd bet the mortgage check against that (regardless of the 1 next to its screenname).
Wouldn't surprise me at all if he was the other poster what started a 'hey guys, I'm a fan but the fumble numbers make me believe the Patriots are very bad people' thread from the other day.

Click to expand...

The technical name is "Concern Troll"

primetime said:

Well, that's not quite what statistical randomness means, though I agree that the word random here is often used to confuse people. It simply means the difference in means is statistically significant (likely at p<.05), which means that the variation in the number of fumbles (really, lost fumbles) per play is highly unlikely to be due to random variation.

The problem is that there's a ton of problems with both the dependent variable (lost fumbles per play). Notably, the result of a fumble (i.e. whether it is lost or recovered by the same team) is random, for all intents and purposes. That means the difference between the Patriots average and the league average could be that the Patriots happen to recover their fumbles at a 60% clip versus the 50% clip of the league average. Assuming the Patriots fumbled infinite times, this would fall to 50%. But because teams only fumble a handful of times each season, a random difference in the rate of fumble recovery can make all the difference here.

Once you change fumbles lost into fumbles, the Patriots are still very good but they're not impossibly or historically good or even outside an expected distribution. In fact, a lot of the data being presented is due to the bizarre methodological choice of "5 year periods," where a single outlier season can cause a team to appear in his stats up to 5 times.

The more interesting question begged from the analysis (because the causal inference drawn is total bunk and would never pass peer review, but it does get you a national outlet in a Slate thinkpiece I guess) is how the Patriots are so good at recovering their own fumbles. Obviously a large part of this is random, but there could be an underlying statistical cause. For example (this is just a musing), a strip-sack could be more likely to be recovered by the fumbling team, and the Patriots' fumbles may disproportionately come from strip-sacks.

Click to expand...

Mine was a typo, the 5 year period was all fumbles. I agree with everything you said.

Skree said:

The technical name is "Concern Troll"

Click to expand...

d%#*!@t, I overslept the day they taught Troll 101. So now I have no idea about Troll methodology such as 'Concern Trolling' (love that term).
I really have to go back to school

http://www.sharpfootballanalysis.co...umble-more-often-when-playing-for-other-teams

Our statistical genius managed to count Special Teams fumbles in his analysis, but not Special Teams touches. Why does that matter? Because Wes Welker and Brandon Tate account for 24 Special Teams fumbles with the Dolphins and Bengals respectively.

Of course, Special Teams use "K-Balls" which the teams do not have access to.

Danger Zone said:

Its obviously cheating on the part of the Patriots. We cheat on the coin flip so we can get that much vaunted Defer Option.. which is exactly the scenario that 30/31 other teams would have given us anyway had they simply won the toss fairly.

Click to expand...

Hey--we have won the toss this year in 14/18 games, so....