The New England Patriots’ prevention of fumbles is nearly impossible

If you watch Patriots games, you'll see that if Patriots running backs fumble, they get benched. If they do it again they get released. Belichick is maniacal about getting on his players about fumbling. That is a simple explanation. The team has been well-coached for a long time. If you think the Patriots' success is all from deflated balls, you need to explain how they out-scored the Colts 28-0 in the second half (and didn't fumble in horrible conditions).

I&#x27;m becoming more disappointed in this analysis.<p>&quot;But in 2007, something happened to propel [the Patriots] to a much better rate.&quot;<p>The author is using a 5-year average, and then concluding that something must have happened in a single year that is suddenly different. What that implies to me, is that they had a particularly terrible year in 2002. So, when the really bad year slides off of the 5-year moving average, then the 5-year average suddenly looks much better.

As someone who doesn&#x27;t follow the NFL I am surprised that they allow the teams to supply their own balls. Seems it would be less complicated and achieve an even playing field if everyone had to use the same balls supplied by the league.<p>Does this happen in baseball&#x2F; basketball as well? From an Australian perspective it seems unusual.

plays per fumble is not a very good metric - as fumbles go down it increases exponentially. If you look at the total fumbles, the patriots are #1 but only 7 ahead of #2. The two worst teams are 10 behind the next two teams. I didn&#x27;t run the numbers, but eyeballing it, it seems like the Patriots are pretty easily within a couple standard deviations there.

For those of you who aren&#x27;t keeping up with the up to the minute minutae of the New England Patriots (American football team) scandal (presumably because you have lives), head coach Bill Belichick gave a press conference a few hours ago where he provided a hypothesis of what might have happened: as many people know, quarterbacks like their game footballs prepared in specific (sometimes elaborate [0]) ways prior to use in games. Belichick says that the team went over the process they use for games a few days ago and they determine that the preparation process they use for their footballs (which he claims they do right up to the point they are given to officials for pre-game testing) increases the air pressure in the balls by about 1 psi [1]. Therefore, the balls appear legal when tested by officials, but after a few hours of sitting around, the pressure reduces to below the legal limit. He doesn&#x27;t specify the process used by the team, but presumably it&#x27;s some kind of rubbing that heats the ball up, and when it cools down the interior air pressure is reduced accordingly.<p>I&#x27;m sure someone will test this out and verify or discredit the claim.<p>[0] <a href="http://www.nytimes.com/2013/11/24/sports/football/eli-mannings-footballs-are-months-in-making.html?_r=0" rel="nofollow">http:&#x2F;&#x2F;www.nytimes.com&#x2F;2013&#x2F;11&#x2F;24&#x2F;sports&#x2F;football&#x2F;eli-mannin...</a><p>[1] <a href="http://www.boston.com/sports/football/patriots/extra_points/2015/01/bill_belichick_defends_patriots_on_deflategate_we.html?p1=feature_pri_hp" rel="nofollow">http:&#x2F;&#x2F;www.boston.com&#x2F;sports&#x2F;football&#x2F;patriots&#x2F;extra_points&#x2F;...</a>

Analyzing outlying statistics across all the available variables would be an interesting technique to try to predict other potential rules violations. This could apply to many different types of sporting events.<p>I wonder if anyone is working on this problem?<p>Despite a few nit picky issues I have with this particular analysis of the Patriots fumble data it&#x27;s probably the strongest evidence I&#x27;ve seen so far that there has likely been a persistent rules violation.

I&#x27;m not convinced that it&#x27;s okay to just remove all the dome-teams from the plot. Using the author&#x27;s data, here&#x27;s a plot of all the teams and total fumbles (not &#x27;fumbles lost&#x27; that the author focuses on):<p><a href="http://i.imgur.com/QP6LXWg.png" rel="nofollow">http:&#x2F;&#x2F;i.imgur.com&#x2F;QP6LXWg.png</a><p>The Patriots are not even the top team! It&#x27;s actually the Falcons that have the most plays per fumble. If it&#x27;s &quot;nearly impossible&quot; for the Patriots, what does that say about the Falcons? The author focuses on &quot;fumbles lost&quot; because that number is more dramatic if one is trying to prove that the Patriots did something wrong. That number in isolation only shows that the Patriots were extremely good at recovering their own fumbles.<p>Looking at the graph, there is a notable bend at the top 4 teams. Either the top 4 teams have fewer fumbles through training, or the top 4 teams are all cheating somehow. It&#x27;s hard to call the Patriots statistics impossible when they&#x27;re in 2nd place.

It would be interesting to get the stats on fumbles during kickoff&#x2F;punt returns (when a neutral kicking ball is in use, not a team-supplied offensive ball), and on plays involving interceptions or fumble recoveries (where they would be using the other team&#x27;s ball). If they rank highly in fumble prevention in those situations, it would suggest that this is a result of coaching, not using a particular ball.

Listen, engineers and scientists: 11 out of 12 balls, markedly different and underinflated compared to the other team&#x27;s balls? This is a priori evidence of tampering. Any jury would convict. We don&#x27;t need a Perry Mason-type confession; we don&#x27;t need to identify the culprit. The team should be strongly punished for breaking the NFL rules. I think the NFL should do an extensive investigation, after the initial punishment, which justly would be done BEFORE the Super Bowl. This investigation needs to determine whether the Pats have cheated with underinflated balls over a period of months and years. It can be done. Then is the time for scientific analysis, testimony by physicists, and so on.<p>My point is: We don&#x27;t need to identify the person responsible, and we don&#x27;t need a confession. Guilt is obvious, a priori. Punishment should be swift.

Is the author confusing &quot;fumbles&quot; and &quot;fumbles lost&quot; or am I?

Wouldn&#x27;t pass completion ratio be a better metric?

The game against the Colts the Patriots clearly cheated. It is beyond reasonable that 11 of 12 footballs are 2psi less than the rule allows. Especially since the 12th football of regulation weight is the football used for kicking as you want the football inflated properly. But using fumbles as proof of long term violation is tough because their are too many factors to consider. Patriots are a pass heavy team and have been for a long time. They use multiple RB&#x27;s and a fresh RB is less likely to fumble than one RB used heavily throughout the game due to mental and physical fatigue. WR&#x27;s fumble but are less likely than RB&#x27;s. An under-inflated ball is easier to grip so it would be interesting to see how WR&#x27;s who left have fared on their new teams and if any returned later, after playing poorly on new team, to play well for the Patriots. Since 2007 has Tom Brady&#x27;s completion % and interception ratio on the road improved drastically? All these things could be explained by an under-inflated but not definitively.