Sunday, January 15, 2012

2011 NFL regular season wins and stat correlations

With the conference championships teed up for next weekend, here are correlations between several stats and wins during the 2011 NFL regular season. This isn't a sports blog, and I'm not making any audacious claims about being able to provide special insights. Just the raw correlations for entire teams here, next to the same for last year for comparative purposes:

Team passer rating
Team points scored
Yards gained per pass play
Total yards gained
1st downs
Total yards gained per play
Pass yards gained
Turnover ratio
Time of possession
3rd down conversion %
QB hits allowed
Sacks allowed
4th down conversion %
Rush attempts
Pass attempts
Yards gained per rush play
Rush yards per game
Total team penalty yards
Offensive penalty yards
Pass:Run ratio
Special teams

Average net punt
Average yards gained per kick return
Average kickoff distance
Average kickoff return yards allowed
Field goal %

Rushing yards allowed per game
Points allowed per game
Opponent's passer rating
Yards allowed per pass play
4th down conversion % given up
Passing yards allowed per game
3rd down conversion % given up
Sacks made
Defensive penalty yards
1st downs given up per game
Total yards allowed
Yards allowed per play
Yards allowed per rush

* not statistically significant at 90% confidence

Quarterbacks are king. Teams that move the ball through the air, win. Teams that can't, don't. The difference between having a play maker with a rifle for an arm like Jay Cutler and a dud like Caleb Hanie is starting off 7-3 and finishing 1-5. Need we even mention Peyton Manning?

The 2011 season felt to this fan more like 2009, which Steve Sailer deemed the year only passing mattered, than it did 2010. Indeed, the correlation with wins and the statistics presented above between the 2011 and 2009 seasons is .84, even more rigorous than the .76 correlation between 2011 and 2010.

OneSTDV, on why the NFL is the most popular sports league in the US:
The NFL is now the ultimate sports behemoth, perhaps as a result of helmets and pads covering up black superstars and whites still dominating the marquee positions of QB and head coach.
I'm unclear as to how strong the relationship between professional sports players' demographics and the popularity among the public of the sports they play is. The NHL and MLS are both a lot whiter than the NFL while remaining far less popular than that football behemoth is. Over the last few decades, the MLB has become increasingly non-black and yet has experienced a long-term decline in popularity not just among blacks but also among whites to such an extent that football, not baseball, now clearly deserves to be referred to as America's favorite pastime. But it's hard to argue that in football's most important position, whites are in fact overrepresented relative to racial composition of the country as a whole.

While assertions that "defenses win championships" inevitably abound during postseason play, the offensive correlations are considerably and consistently stronger*. That total rushing yards allowed--not yards per attempt, but total run yards given up--correlates more strongly with wins than any other defensive stat does says a lot about offensive predominance. When defenses give up a lot of rushing yards in a game, it's often because the opposing team spent a good chunk of the second half on the ground to burn off the clock and avoid costly turnovers while protecting an early lead that the losing team's offense couldn't match.

To argue that offense matters more than defense does might seem like arguing that it only matters how hard your side pulls on the rope in tug-of-war, not how hard the other side does. Yet defenses in the NFL are mostly fungible. Tampa 2 is a bit of an exception, where the outsized importance of the middle linebacker is akin to the quarterback's, albeit to a lesser degree, but the phrase "cookie cutter" gets (over)used so often because it is accurately descriptive. If an offense can move the ball in the air against a standard NFL defense, that team is likely to be found still playing in January.

It's sometimes asserted that special teams are one-third of the game. It's not true in terms of duration and rarely feels like they are as important as downs from scrimmage. The table above confirms that feeling.

Turnovers and penalties? Turnovers, yes. Penalties, not so much. The Colts and Cowboys were the least penalized team in the league this year. The Raiders and Lions were the most penalized (even if penalties don't appear to be significant in the outcome of games, it is telling that the notoriously dirty Raiders and Detroit, which was embarrassingly chippy this season, accumulated the most fouls).

* Of course, especially stout defenses are still helpful. The Ravens and 49ers both have shots at the Super Bowl primarily because of the strength of their defenses (or, in the case of San Francisco, the astounding and league-leading +28 turnover ratio that continued in the upset against New Orleans). To oversimplify, next weekend features the Ravens and 49ers representing defensive primacy and the Giants and Patriots the power of offensive dominance. Should make for some enjoyable spectating!


KPres said...

I think "defense wins championships" is supposed to mean that defense matters more in the playoffs and even the Superbowl itself, than offense. The idea being that those games are played under more pressure and intensity, making it more difficult to execute a complex offensive scheme. It would be interesting to see the correlations there, particularly in the SB itself.

You can get more in-depth statistical analysis of football here:

They've probably gone over all of this stuff.

Jokah Macpherson said...

Another way in which defense could be said to win championships is that it is more likely to be the differentiating factor among teams who both have a strong passing offense. During the regular season and early rounds of the playoffs you can coast on your offensive superiority but in games where the opponent matches up with you in this category, the relative importance of defense goes up. I don't know how to statistically test this theory, though.

I think a lot of football's appeal over other team sports is due to its short season, which heightens the sense of drama.

Ed Tom Kowalsky said...

Statistical reports I'd like to see:

the respective winning percentage of white and black starting quarterbacks over the last 30 years

the respective winning percentage of white and black head coaches over the last 20 years

the correlation between the whiteness (or blackness) of NFL rosters and winning percentage over the last three years

PS--I've always believed the ability to run the football (as measured by running back yards per carry) and stop the run (as measured by opposing running backs yards per carry) is likely to be the strongest index of winning because it denotes the ability to control the line of scrimmage. And as we all know, the game is won or lost in the trenches.

Audacious Epigone said...


Looking at the last ten SB winners, and their respective offensive and defensive regular season rankings:

2010: Packers, 9th and 5th
2009: Saints, 1st and 25th
2008: Steelers, 22nd and 1st
2007: Giants, 16th and 7th
2006: Colts, 3rd and 21st
2005: Steelers, 15th and 4th
2004: Patriots, 7th and 9th
2003: Patriots, 17th and 7th
2002: Buccaneers, 24th and 1st
2001: Patriots, 19th and 24th

Four of the last ten have been offensively dominant, six of the last ten, defensively. The argument makes sense.


Very well could be true. It does appear that defense becomes relatively more important in the post-season.


I'm pretty sure someone has looked at this before, although I'm not sure off-hand who it was. Steve Sailer has excerpted him (or the site) a couple of times.

Steve Sailer said...

It's interesting how the passer rating is irritatingly opaque, but works pretty well.

Anonymous said...

There are intangibles too, like just getting "hot" at the right moment and clicking on all cylinders, like the Giants last year. Not so hot during the regular season, but win the SB? As far as these statistics are concerned, it obviously needs to be a bit of both. The Bears D of 06 was pretty well rounded, ranking 5th in the league, the Colts were 2nd in total O and not too bad on D. The Colts had more going for them offensively and the Bears had little to no O and couldnt build a decent Time of Possesion to keep Manning off the field.