DVOA Analysis
Football Outsiders' revolutionary metrics that break down every single play of the NFL season

Week 13 DVOA Ratings

by Aaron Schatz

After 13 weeks, the Football Outsiders DVOA ratings still have the Houston Texans number one, although their overall rating drops this week. Green Bay is still number two, and Pittsburgh is the big mover of the week, climbing up to number three. In fact, in Weighted DVOA, which lowers the importance of early games, the Steelers are now second behind Houston and ahead of Green Bay. Remember, their worst game of the year by far was in Week 1.

So, let's talk about Green Bay some more. Their narrow win over the Giants actually drops their overall season DVOA slightly. I'll admit, the whole Packers situation is a bit frustrating. The majority of the football commentariat describes Green Bay as unbeatable, as head and shoulders ahead of the rest of the league. Those of us who look at things a little closer know that the Packers have flaws. We know there is a difference between "unbeatable" and "unbeaten." And yet... let's face it, it just looks weird to not have an undefeated 12-0 team on top of the DVOA ratings. It's even worse now that they've dropped to third in Weighted DVOA. Doesn't Green Bay look like the best team in the league to all of us, even if we don't think they should be ranked among the best teams of all time? 

In order to talk about Green Bay properly, I think we need to separate the concepts of "greatness" and "dominance." The Packers are by no means a dominant team. They are fairly one-dimensional; that dimension just happens to be performing off the charts. (Although as of this week, the Packers have fallen behind the 2007 Patriots for the best passing DVOA ever.) The defense has lived off turnovers, but that element isn't as sustainable as preventing yardage. The DVOA ratings suggest that the Packers aren't much different from two other teams, New England and New Orleans. The difference is primarily one of degree -- the Packers are just a little better on offense and a little better on defense.

2 GB 26.5% 37.3% 1 12.4% 24 1.5% 12
5 NE 24.0% 34.5% 2 13.4% 26 2.9% 7
8 NO 19.1% 34.3% 3 14.8% 29 -0.4% 19

Some readers have suggested that perhaps total DVOA is not the proper measure to use when a team is so superlative in one area of the game. That's the "defense doesn't matter, because Aaron Rodgers can outscore anybody" theory. The problem with this theory is that it doesn't explain any team in the past. It's not an issue with the 2007 Patriots or 2010 Patriots, each of which ranked first in total DVOA as well as pass offense DVOA. It's not really an issue with the 2004 Colts either, who are fourth all-time in passing DVOA and were fourth in total DVOA in 2004. It's easy to argue that the Colts were better than the No. 3 Buffalo Bills that year, but hard to argue that they were truly better than either New England (14-2) or Pittsburgh (15-1). As far as total offensive DVOA goes (as opposed to just passing DVOA), the team that ranks fourth historically in offensive DVOA is the 2002 Kansas City Chiefs, who were the opposite of the Packers -- they finished fourth in the league in total DVOA but went just 8-8.

Is there something in the way DVOA is built that is missing an indicator of Green Bay's dominance? I don't think so. Instead of looking at Green Bay with DVOA, let's look at Green Bay with a fairly simple measure: points scored and allowed. Most readers know that we can estimate a team's wins and losses based on points scored and allowed, known as the Pythagorean projection. And here's the thing: By Pythagorean wins, Green Bay is not the best team in the league this year. Both San Francisco (9.6) and Houston (9.2) are ahead of Green Bay. Green Bay's totals of 420 points scored and 262 points allowed work out to a Pythagorean win percentage of .754, or 9.0 wins. The difference between Green Bay's actual win percentage and Pythagorean win percentage is the second-highest since the merger, behind only the 1992 Indianapolis Colts. And here's something perhaps more surprising: By Pythagorean projection, the 2011 Packers are no better than the 2010 Packers. The 2010 Packers had a Pythagorean win percentage of .757. It just so happens that last year's Packers fell far short of their Pythagorean projection, and this year's Packers are far ahead.

And now, a digression. We've been writing about the Pythagorean projection since we launched in 2003. We've always used 2.37 as the exponent in the equation; this is the exponent that now-Houston Rockets GM Daryl Morey found was most accurate for the NFL when he worked for Stats Inc. two decades ago. However, that exponent is based on the offensive environment of the league. We all know the offensive environment is a bit different now. Teams are scoring more points and allowing more points. So the exponent has changed, and 2.37 is not the most accurate way to estimate Pythagorean wins anymore. Actually, if we want to be as accurate as possible, each team plays in a different offensive environment. Saints games feature lots of points. Jaguars games feature fewer points. The exponent should be different for each team.

Baseball Prospectus discovered this a few years ago and started replacing Pythagorean wins with something it called "Pythagenport" (after writer Clay Davenport). I've figured out a similar method to get better results for the NFL. Pythagenport finds a different exponent for each team based on their offensive environment. The equation that works best in the NFL is 1.5 * log ((PF+PA)/G). The improvement is slight. The correlation between Pythagorean wins and actual wins for 1990-2010 is .9120. The correlation between Pythagenport and actual wins for 1990-2010 is .9134. However, the improvement from Pythagenport is bigger in recent seasons because scoring has been higher in recent seasons. (In particular, it helps with the Colts, who have continuously outperformed the standard Pythagorean projection all decade.) We'll probably use the new method to update all the listings of Pythagorean wins on the site this offseason.

OK, digression ended. Let's get back to the Packers. No team this year has played in a stronger offensive environment than the Packers. Their Pythagenport exponent ends up as 2.63, and this gives us a more accurate projection of wins: 9.3 instead of 9.0. That moves the Packers ahead of Houston, but still they don't have the highest projection in the league. Here's a list of the top teams in the league this year by Pythagenport wins:

SF 10-2 288 161 .799 9.6 2.36 .798 9.6
GB 12-0 420 262 .754 9.0 2.63 .776 9.3
HOU 9-3 310 189 .764 9.2 2.43 .769 9.2
BAL 9-3 296 192 .736 8.8 2.41 .740 8.9
NO 9-3 393 269 .711 8.5 2.61 .729 8.8
NE 9-3 362 247 .712 8.5 2.56 .727 8.7
PIT 9-3 268 195 .680 8.2 2.38 .681 8.2
DET 7-5 333 277 .607 7.3 2.56 .616 7.4
CHI 7-5 291 242 .608 7.3 2.47 .612 7.3
DAL 7-5 283 244 .587 7.0 2.46 .590 7.1

Two notes. First, the Packers are better than last year if we use Pythagenport instead of Pythagorean (.776 to .759). Second, the lowest Pythagenport exponent, for those curious, belongs to Jacksonville at 2.27.

If the Packers are not a dominant team, how have they managed to go 12-0 this season? The main answer is consistency. The Packers have been absurdly consistent. After this week, the Packers lead the league with 3.3% variance. If that number holds until the end of the season, it would be the lowest variance of any team in the DVOA era, surpassing last year's Atlanta Falcons who were at 4.9%.

(For those of you who wonder how variance is computed, it's just the "VAR" function from Excel used on all individual single-game DVOA ratings for each team.)

Of course, there's a reasonable chance that the Packers' variance will change with four more games to play, any of which could be particularly good or bad. So I went back to compare the Packers' variance through 12 games with every team's variance through 12 games, rather than through the end of the season. Using this measure, the Packers finish second. Surprisingly, this year's Atlanta Falcons finish third. But the team that ranks first is even more interesting:

Lowest Variance after 12 Games, 1992-2011
Year Team Var. DVOA W-L
2005 IND 3.1% 39.2% 12-0
2011 GB 3.3% 26.5% 12-0
2011 ATL 3.4% 17.6% 7-5
2007 JAC 3.4% 15.1% 8-4
2006 SD 3.6% 28.1% 10-2
1994 PIT 4.0% 23.0% 9-3
2002 NYG 4.0% -9.0% 6-6
1999 OAK 4.3% 23.3% 6-6
2003 CAR 4.5% 2.0% 8-4
1999 MIN 4.5% 0.4% 7-5


Yes, that's the also-undefeated 2005 Colts who were both more consistent and more dominant than the 2011 Packers. As we all know, that team didn't end up going undefeated, and it didn't win the Super Bowl. The Colts lost to San Diego in Week 15, lost to Seattle (while resting starters) in Week 16, and then lost to Pittsburgh in their first playoff game. Like Aaron Rodgers, Peyton Manning was a pretty darn good quarterback. But he eventually did run into a couple of defenses that could slow him down a little bit.

The 1999 Raiders provide an interesting look at how even a team with both quality and consistency doesn't necessarily get all the breaks. The Raiders played very well but faced the second-hardest schedule in the league (through Week 13 of 1999) and played almost nothing but close games. Their six losses were all by a touchdown or less. Four of their six wins were by a touchdown or less, and another one was by just nine points. The team ended up third in overall DVOA for the season and didn't even manage a winning record, finishing at 8-8.

So, let's circle back to the issue I brought up a few paragraphs ago, that we need to separate the concepts of "greatness" and "dominance." I think it's fair to say that the Packers are a great team, even if they aren't a dominant one. If the Packers manage to consistently play well like this over the next four games, that's a great accomplishment. And if the Packers actually manage to go 19-0 and win the Super Bowl with a bunch of seven-point wins, there is going to be an interesting argument over whether they truly qualify as "the greatest team in NFL history." Does the greatest team need to be the most dominant? However, the toughest tests the Packers will face will not be in their remaining four games. The toughest test will be the playoffs. That's why our current playoff odds report suggests the chances of the Packers going 16-0 are about 10 percentage points higher than the chances of the Packers winning the Super Bowl.

One other thing to consider when pondering the Packers' chances of going 16-0 or 19-0:

Weeks 1-7 40.5% 1 7.9% 19 -0.1% 16 32.5% 1
Weeks 8-13 32.9% 2 18.6% 29 3.8% 10 18.1% 9

Speaking of the playoff odds report, once again this week we've produced two versions: one which gives each team's Weighted DVOA as it currently stands, and one that attempts to adjust for the injured quarterbacks in Chicago, Houston, and Kansas City. In past weeks, I've reduced weighted DVOA by our rough estimate of the general difference between an average quarterback and a replacement-level quarterback, -13.3% DVOA. However, that's not how things have worked out in the two weeks since these teams changed quarterbacks. The Houston and Chicago offenses have declined significantly, while the Kansas City offense has barely declined at all. So it makes sense to adjust our ratings a bit more for Houston and Chicago, a bit less for Kansas City. What I ended up doing this week was figuring out the difference between each team's offensive DVOA before the quarterback change and after; then I reduced each team's Weighted DVOA by half that amount. It's a bit of a conservative estimate, but there's also no guarantee that just because Caleb Hanie had a horrible game this week, that means he's going to have a horrible game every week.

* * * * *

These are the Football Outsiders team efficiency ratings through 13 weeks of 2011, measured by our proprietary Defense-adjusted Value Over Average (DVOA) system that breaks down every single play and compares a team's performance to the league average based on situation in order to determine value over average. (Explained further here.)

OFFENSE and DEFENSE DVOA are adjusted to consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS DVOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season. WEIGHTED DVOA represents an attempt to figure out how a team is playing right now, as opposed to over the season as a whole, by making recent games more important than earlier games.

As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.

To save people some time, please use the following format for all complaints:

<team> is clearly ranked <too high/too low> because <reason unrelated to DVOA>. <subjective ranking system> is way better than this. <unrelated team-supporting or -denigrating comment, preferably with poor spelling and/or chat-acceptable spelling>

1 HOU 31.0% 1 33.6% 1 9-3 19.8% 5 -9.6% 5 1.7% 11
2 GB 26.5% 2 23.5% 3 12-0 37.3% 1 12.4% 24 1.5% 12
3 PIT 25.8% 6 32.3% 2 9-3 20.0% 4 -3.4% 9 2.4% 9
4 BAL 24.2% 5 20.5% 6 9-3 8.7% 12 -19.4% 1 -3.9% 28
5 NE 24.0% 3 23.2% 4 9-3 34.5% 2 13.4% 26 2.9% 7
6 NYJ 21.3% 4 21.1% 5 7-5 4.0% 18 -13.0% 2 4.3% 4
7 SF 20.3% 8 19.1% 9 10-2 0.8% 20 -11.3% 3 8.2% 2
8 NO 19.1% 7 19.8% 7 9-3 34.3% 3 14.8% 29 -0.4% 19
9 ATL 17.6% 10 19.8% 8 7-5 8.7% 13 -8.5% 6 0.4% 17
10 CHI 12.6% 9 13.1% 10 7-5 -8.3% 25 -10.6% 4 10.3% 1
11 TEN 11.5% 12 10.5% 12 7-5 7.9% 15 1.9% 12 5.5% 3
12 NYG 9.3% 14 11.7% 11 6-6 13.9% 7 5.8% 18 1.3% 13
13 DET 8.3% 11 8.5% 13 7-5 8.6% 14 -6.1% 8 -6.4% 32
14 DAL 4.5% 13 4.0% 15 7-5 8.8% 11 2.9% 14 -1.4% 23
15 PHI 1.2% 16 2.1% 16 4-8 10.1% 9 10.6% 22 1.8% 10
16 CIN 0.9% 15 1.1% 17 7-5 5.6% 17 5.5% 16 0.9% 14
17 MIA 0.6% 18 4.3% 14 4-8 2.9% 19 3.2% 15 0.8% 15
18 OAK -2.9% 17 -6.0% 21 7-5 5.9% 16 8.2% 20 -0.6% 21
19 DEN -4.3% 20 -1.8% 18 7-5 -2.9% 23 5.6% 17 4.2% 5
20 CAR -5.6% 24 -3.3% 19 4-8 18.6% 6 18.6% 31 -5.6% 30
21 SD -5.7% 22 -5.5% 20 5-7 13.1% 8 16.2% 30 -2.6% 27
22 BUF -5.8% 19 -14.2% 23 5-7 9.4% 10 13.7% 28 -1.6% 25
23 SEA -9.1% 27 -7.0% 22 5-7 -8.3% 26 1.2% 11 0.4% 18
24 TB -11.5% 21 -14.9% 24 4-8 -0.3% 21 13.6% 27 2.4% 8
25 WAS -11.8% 25 -15.5% 27 4-8 -10.3% 27 1.1% 10 -0.5% 20
26 CLE -16.0% 26 -15.1% 26 4-8 -6.2% 24 9.0% 21 -0.8% 22
27 MIN -16.4% 28 -20.7% 29 2-10 -1.8% 22 12.2% 23 -2.4% 26
28 JAC -17.7% 23 -17.8% 28 3-9 -21.4% 31 -7.6% 7 -3.9% 29
29 KC -20.3% 29 -15.0% 25 5-7 -18.1% 30 2.6% 13 0.5% 16
30 ARI -22.9% 30 -22.3% 30 5-7 -13.8% 29 12.6% 25 3.6% 6
31 STL -36.1% 31 -34.3% 31 2-10 -26.8% 32 7.9% 19 -1.5% 24
32 IND -38.9% 32 -40.0% 32 0-12 -12.7% 28 20.3% 32 -5.9% 31
  • NON-ADJUSTED TOTAL DVOA does not include the adjustments for opponent strength or the adjustments for weather and altitude in special teams, and only penalizes offenses for lost fumbles rather than all fumbles.
  • ESTIMATED WINS uses a statistic known as "Forest Index" that emphasizes consistency as well as DVOA in the most important specific situations: red zone defense, first quarter offense, and performance in the second half when the score is close. It then projects a number of wins adjusted to a league-average schedule and a league-average rate of recovering fumbles. Teams that have had their bye week are projected as if they had played one game per week.
  • PAST SCHEDULE lists average DVOA of opponents played this season, ranked from hardest schedule (#1, most positive) to easiest schedule (#32, most negative). It is not adjusted for which games are home or road.
  • FUTURE SCHEDULE lists average DVOA of opponents still left to play this season, ranked from hardest schedule (#1, most positive) to easiest schedule (#32, most negative). It is not adjusted for which games are home or road.
  • VARIANCE measures the statistical variance of the team's weekly DVOA performance. Teams are ranked from most consistent (#1, lowest variance) to least consistent (#32, highest variance).

1 HOU 31.0% 9-3 33.9% 8.5 5 -1.4% 25 -8.0% 29 16.2% 23
2 GB 26.5% 12-0 28.3% 9.7 1 -1.1% 24 -0.6% 19 3.3% 1
3 PIT 25.8% 9-3 25.9% 8.5 4 -0.3% 22 -12.0% 31 15.8% 22
4 BAL 24.2% 9-3 22.0% 8.8 2 4.8% 3 -14.9% 32 22.8% 29
5 NE 24.0% 9-3 23.5% 8.8 3 -0.9% 23 -5.4% 27 10.1% 10
6 NYJ 21.3% 7-5 20.0% 7.3 9 1.9% 11 -2.3% 24 13.4% 18
7 SF 20.3% 10-2 25.9% 7.8 8 -3.9% 32 -10.6% 30 4.9% 3
8 NO 19.1% 9-3 18.6% 8.1 6 -2.8% 28 1.8% 16 13.1% 15
9 ATL 17.6% 7-5 14.0% 7.9 7 2.4% 10 -3.9% 25 3.4% 2
10 CHI 12.6% 7-5 11.9% 7.0 11 1.5% 14 -0.9% 20 13.3% 16
11 TEN 11.5% 7-5 15.9% 7.0 10 -0.1% 21 -1.6% 23 17.5% 25
12 NYG 9.3% 6-6 6.9% 6.9 12 0.5% 19 4.6% 9 18.1% 26
13 DET 8.3% 7-5 8.8% 6.9 13 3.1% 9 0.4% 17 8.2% 8
14 DAL 4.5% 7-5 9.3% 6.4 15 -2.5% 27 2.1% 15 18.3% 27
15 PHI 1.2% 4-8 1.4% 5.9 19 0.6% 18 3.6% 12 13.5% 19
16 CIN 0.9% 7-5 6.9% 6.6 14 3.2% 7 -1.0% 21 6.5% 5
17 MIA 0.6% 4-8 1.4% 6.4 16 1.9% 12 10.2% 4 13.3% 17
18 OAK -2.9% 7-5 1.7% 6.1 17 1.9% 13 2.2% 14 23.9% 31
19 DEN -4.3% 7-5 -2.5% 6.0 18 1.5% 15 2.6% 13 7.2% 6
20 CAR -5.6% 4-8 -2.5% 5.4 21 -3.4% 31 14.0% 1 16.8% 24
21 SD -5.7% 5-7 -0.2% 5.1 23 0.1% 20 5.9% 8 10.7% 11
22 BUF -5.8% 5-7 -6.1% 5.7 20 4.2% 4 3.6% 11 27.3% 32
23 SEA -9.1% 5-7 -6.9% 5.4 22 0.6% 17 -6.5% 28 12.1% 13
24 TB -11.5% 4-8 -18.7% 5.1 24 12.0% 1 -0.3% 18 20.9% 28
25 WAS -11.8% 4-8 -15.3% 5.0 25 -3.2% 30 4.5% 10 12.6% 14
26 CLE -16.0% 4-8 -10.3% 4.8 26 -3.0% 29 13.2% 2 5.1% 4
27 MIN -16.4% 2-10 -14.5% 4.4 28 3.2% 8 7.0% 7 13.8% 20
28 JAC -17.7% 3-9 -19.4% 4.3 29 8.2% 2 -5.3% 26 15.1% 21
29 KC -20.3% 5-7 -19.5% 4.4 27 0.7% 16 10.1% 5 23.3% 30
30 ARI -22.9% 5-7 -16.7% 4.1 30 -2.5% 26 -1.0% 22 8.5% 9
31 STL -36.1% 2-10 -39.9% 2.8 31 3.8% 5 9.5% 6 7.6% 7
32 IND -38.9% 0-12 -43.9% 2.1 32 3.3% 6 12.2% 3 10.8% 12


359 comments, Last at 10 Dec 2011, 11:50pm

1 Re: Week 13 DVOA Ratings

I don't think this was meant to be funny, but I got a chuckle out of it... "Green Bay is still number two, and Pittsburgh is the big mover of the week, climbing up to number three... So, let's talk about Green Bay some more."

2 Re: Week 13 DVOA Ratings

Wait, Green Bay has given up more points than the Patriots? Really? That amazes me. Green Bay really does have a poor defense, doesn't it?

8 Re: Week 13 DVOA Ratings

Green Bay allows 6.3 yards per play, highest in the league, but New England is 6.2 themselves.

(Baltimore and Pittsburgh tie for first at 4.6)

105 Re: Week 13 DVOA Ratings

I for one have more trust in the Patriots' defense's ability to stop/slow down Aaron Rodgers than I have in the Packers' defense's ability to stop/slow down Tom Brady.

115 Re: Week 13 DVOA Ratings

The difference is that Green Bay has guys like Matthews, Woodson and Williams, who can make 1 or two plays that can swing the game. I don't think the Patriot's defense has that dimension. So all else being equal, you've got to favour GB's D slightly.

119 Re: Week 13 DVOA Ratings

Arrington has really good ball skills. Thought he was getting fluke interceptions at first but the last few have been kind of impressive. GB does scare me though.

248 Re: Week 13 DVOA Ratings

So have the Jets, Saints, (the bears have given up 5 less points), (Dallas has given up 3 less points), the 7-5 Raiders, Broncos, Benglas, Lions have all given up more. The 7-5 falcons have given up 3 less.

The patriots defense gives up a ton of yards, but its pretty good at keeping people out of the endzone, and creates a significant amount of turnovers.

3 Re: Week 13 DVOA Ratings

2 weeks ago, the Bears offensive DVOA was close to 0%. If you replace their current value with that they grade out at about 20% overall DVOA right in the thick of things with the other good teams in the league. Arguably, that offensive rating might have undervalued them because of the huge issues the Bears had in weeks 2 and 3.

Lovie Smith might have been right when he said this was the best group of players he's had in Chicago, but we'll probably never learn how far they could have gone now.

320 Re: Week 13 DVOA Ratings

Agreed, and I can't tell you how sad that makes me. Nobody's fault; it just sucks.

(Well, maybe it's Jerry's fault for not getting a better backup, but honestly the Bears weren't getting out of the first round without a healthy Cutler.)

4 Re: Week 13 DVOA Ratings

"The defense has lived off turnovers, but that element isn't as sustainable as preventing yardage."

This might be true from a league wide perspective but am not sure if it is right as far as individual teams are concerned. This is the 3rd consecutive year that GB has been near the top of the league in Turnovers/INTs. Recovering fumbles might be luck but forcing them and interceptions are an ability

6 Re: Week 13 DVOA Ratings

Recovering fumbles might be luck but forcing them and interceptions are an ability....that statistical analysis reveals is unsustainable.

There you go finished it for you.

16 Re: Week 13 DVOA Ratings

More precisely: less sustainable than the ability to hold an opposing offense below 400 yards every once in a while.

The Packers gave up three pass plays of more than 40 yards to New York. When was the last time that happened? The Randy Moss coming-out party? As Aaron's table shows, the Packer D is bad and getting worse. Dom Capers admitted in his conference call that there are some fairly basic routes the defensive backs just can't cover. Opponents will continue to call these plays until such time as the Packers work out how to defend them.

139 Re: Week 13 DVOA Ratings

I'm not terribly surprised that the GB defense is getting worse, especially since the last game was the worst of all. They're missing Nick Collins, and against the Giants their two middle linebackers, and for some time another safety (Woodson). Injuries matter. Jarrett Bush is a fine special teams player, an excellent gunner. As a safety, not so much.

29 Re: Week 13 DVOA Ratings

Here are GB's INT statistics since Capers took over:

2009-30 INTs
2010-24 INTs
2011-23 INTs-On Pace for 30.6666

I guess Capers is just a lucky Defensive Coordinator. Hopefully his lucky rabbits foot continues

37 Re: Week 13 DVOA Ratings

Not necessarily, you are assuming a few things:

1. That the ability to intercept is equally distributed throughout teams. Only 1 Charles Woodson, Nick Collins, Ed Reed, Charles Tillman, etc exist

2. That coordinators have similar philosophies. Just because 1 Coordinator believes that an aggressive approach is the best one does not mean others do

*People readily acknowledge that more skilled Players are more adept at forcing Turnovers than other players. You would be in the Minority to say that Woodson or Polamolu force more turnovers than Peprah simply because of luck or playing time.

61 Re: Week 13 DVOA Ratings

No, I am not making those assumptions. Football has been played for a while, with many different combinations of players and coaches. Yes, it is possible that this set of players and this coach are a unique combination, and are thus showing a much greater degree of sustainability that other combinations have shown. Absent a much larger sample size, however, that isn't a wager I'd make, at least not one of much size.

91 Re: Week 13 DVOA Ratings

Dom Capers was a defensive consultant for the New England Patriots in 2008.

The Patriots play a similar style of defense ... in terms of emphasizing more on generating turnovers and stopping in the red zone and on 3rd down than giving up yards. The Pats led the league in turnover ratio last season.

I don't think it is chance that the Packers or Patriots D are generally "creating" more turnovers than other teams.

I also think that current statistical methods fail to interpret this completely. I am sure Belichick cares less that his D is 30+ place in yards or in the bottom quarter in DVOA. He would care more about whether he thinks the execution is better or worse than he thinks it should be.

NOTE: This is not meant to be a comparison between the Pats and Packers D, nor a hint on which coach might have applied such a style first, nor any sort of opinion whether the Patriots or Packers defense are any good or not. (I am sure I missed something in the disclaimer)

110 Re: Week 13 DVOA Ratings

And here we see the necessity of looking at INT per pass attempt. Since NE and GB score a lot and have vulnerable pass defenses, they see a lot of pass attempts--close to 20 percent more than the median. They still produce a lot of interceptions per attempt, but it no longer looks quite so exceptional.

126 Re: Week 13 DVOA Ratings

Wait - why do we care about INT per pass attempt?

We're not interested in which team gets the most INTs per pass attempt. We're interested in which team gets the most INTs.

Maybe if we're picking players for the Pro Bowl, we could get into debates about which players are better based on INT/attempt. But if we want to know which defense is doing better at intercepting passes, a simple total will suffice.

128 Re: Week 13 DVOA Ratings

No, we're interested in rate to determine which team is better at INTs.

If Green Bay played Denver and Tebow didn't throw a pass, would zero INTs mean that Green Bat suddenly got bad at intercepting the ball?

149 Re: Week 13 DVOA Ratings

No, we wouldn't care about whether GB was good or bad at intercepting the ball in that situation.

What do you want DVOA to do? Do you want it to measure what has happened or do you want it to measure what might happen? I'm content to let it measure what has happened.

162 Re: Week 13 DVOA Ratings

What if a team plays in such a way that causes teams to throw more often, baiting them into more interceptions? This might raise the interception total but not the rate (or even lower the rate). The opportunities to get an INT are caused how both sides call their plays and interact.

The Interception Rate and Total are both useless all by themselves.
The rate tells you something, but the picture is incomplete because the situation that those passes are in matters. Is the rate high because there are more passes in desperate situations? Or high because of the pressure? Or low because the opponents threw mostly 'safe' passes?

The total tells you something, but the picture is incomplete because it tells you nothing about the number or context of the opportunities.

30 interceptions on Hail Mary's at the end of half/game is not the same as 30 tipped INTs or 30 caused by disguised coverage or pressure.

Hmm, isn't there a stat that tells you, based on situational context, the value of plays? Hmmmm somewhere maybe.

253 Re: Week 13 DVOA Ratings

No, we don't need INT per pass attempt, thats useless.

We need INT per DRIVE. Thats whats really important.

255 Re: Week 13 DVOA Ratings


Well, GB and NE are 1st and 2nd in this, just as people suspected.

259 Re: Week 13 DVOA Ratings

I can see where interceptions per drive could be more useful, but I don't see why interceptions per pass attempt is useless. A team with more pass attempts against it should make more interceptions.

EDIT: That is, the Packers and Patriots could have the same true ability to intercept passes, but if the Packers have faced 10% more pass attempts, over the name number of drives, they should have 10% more interceptions. Any variance would be indicative of differences in ability.

272 Re: Week 13 DVOA Ratings

It's more complicated than that. You have to add down and distance, and score, and time remaining. Takeaways are quite heavily correlated to game situation: teams that are in the lead intercept more passes than teams that are behind. Not all pass attempts (or drives) are equally likely to produce an interception.

If only someone had invented a play-by-play statistic that uses a baseline to calculate something like a value over average.

44 Re: Week 13 DVOA Ratings

Depends if you mean a particular team, or any team this year, or any team in history. Within a given year, out of 32 teams, let's go with a Bonferroni-corrected 10 in a row.

118 Re: Week 13 DVOA Ratings

How many times in a row would a team need to win the coin toss for you to conclude that they were demonstrating a sustainable skill?

There actually is a mathematical answer to your question, that depends on what your threshold is for determining it was a sustainable skill. For example, winning three coin tosses in a row is probably meaningless (there's a 12.5% chance of that happening). Winning 20 coin tosses in a row has less than a 1 in a million chance of happening by random chance, so I think in that case it's pretty reasonable to say the team is somehow cheating on the coin toss.

352 Re: Week 13 DVOA Ratings

The browns lost 12 coin tosses in a row this year. Think they were cheating to lose? No matter how many times you lose a coin toss in a row, the odds are still 50/50 on the next one. If it comes up heads 1000 times in a row, its still
50/50 on the next one.

Just ask Rosencrantz. Or is it Guildenstern?

353 Re: Week 13 DVOA Ratings

Yeah, if the prior 11 tosses have already come up heads, then the probability of getting 12 in a row is 50%. If you're starting from scratch, the probability is .024%. Those are two entirely different things.

121 Re: Week 13 DVOA Ratings

Look, I'm not a GB fan, but what the man says is true. Certain teams do create more fumbles and certain teams do create more interceptions. The former is an approach, the latter a scheme. Do you really believe that all DC's prioritize tackling above ball hawking? cause some actually don't. And don't you believe that some schemes are set specifically to fool a QB into throwing grab passes? Some are.

If GB fits into all that is debatable. That it exists is not.

199 Re: Week 13 DVOA Ratings

Sure. Interception rate is a predictor for itself. It's just not a very good one. Yes, the Packers almost certainly are better than average at creating interceptions, whether because of their ball-hawking defense, their high-scoring offense, or both. However, they are also almost certainly lucky, and the luck related element can be expected to regress. They don't have to be only one of lucky and good.

202 Re: Week 13 DVOA Ratings

I think that this season has provided a very good comparison for explaining the Packers' ball hawking D. Last year the Colts defense ranked 24th, exactly the same spot as the pack this year and have slipped to dead last without the pressure exerted by Manning's offensive prowess. While anyone who knows anything about football would happily concede that Woodson is an excellent player and Williams is becoming one, it should also be clear that a very good offense can also have an impact on the level of risk that an opponent takes with the ball. Whether taking that risk makes any sense is debatable, that some teams do so when faced with trying to keep up with a Rogers or a Manning is less so.

271 Re: Week 13 DVOA Ratings

I don't think that's a fair comparison. Some of it is on Manning being out, but their secondary (and pass coverage in general) is worse than last year and they've been awful since injuries hit the DL. The defense was actually doing pretty well through the first 3-4 games and have been horrible since Nevis and Foster got hurt.

28 Re: Week 13 DVOA Ratings

I think you're misunderstanding "sustainable." You can be very good at takeaways for part of a season, have a bad stretch where you record few, and still finish near the top of the league.

Being good at something doesn't mean you can do it constantly.

33 Re: Week 13 DVOA Ratings

I guess I disagree about how you use "sustainable" then. Sustainable is more about the season long ability/trends for a team that any small amount of games. A team that regularly records 20+ INTs/Season can be said to have "Sustained" a particular level of skill in Intercepting Passes

45 Re: Week 13 DVOA Ratings

Sure, if that definition comforts you, use it. But they don't play football by season, they play it game by game. And game by game, turnovers are highly variable. There's are times when sportscaster cliches are useful: "Live by the big play, die by the big play." Yardage and per play statistics are much, much more consistent than turnover statistics. It's not very likely that a defense, which gives up a large amount of yards and points per play and per game will improve substantially, while there's a fairly good chance that a team that records a high number of takeaways will see a regression, or at the very least high variance.

73 Re: Week 13 DVOA Ratings

I just checked, and here is GB's INT breakdowns per game:

0 3 2 3 2 1 2 3 1 2 3 1

Only one game (New Orleans) without an INT. Never less than 3 INTs across any two consecutive games, and never less than 5 over any three consecutive games. That's pretty darn consistent.

The biggest problem is that, like almost any team, GB will force more INTs against worse QBs (0 against Brees, 1 against Manning, though they did get 3 each against Rivers and Stafford).

Considering that GB is one of the most consistent teams in DVOA history as Aaron has just pointed out, I think it's worth considering that GB is simply better than average at making consistent INTs on defense, just like they are better than average at completing passes and scoring TDs on offense.

It's mostly been the same core players getting the bulk of the INTs for the past few years (Woodson, Williams, and Collins), spread around enough so that one player having an off year can't drag the whole team down. Woodson was off last year, and Collins is out this year, but the other guys are good enough to pick up the slack. The teams that aren't as consistent generally don't have as many proven interceptors in their secondaries.

52 Re: Week 13 DVOA Ratings

I think most people inclined towards statistical analysis would operationalize "sustainable" as "showing strong self-correlation over time". This is the methodology used for the analyses that show, for example, that "clutch hitting" isn't a "sustainable skill".

One team recording 24+ interceptions in three consecutive seasons could be evidence of a sustainable skill, or it could be merely small sample size. If getting interceptions is truly a sustainable skill, then you would expect to see it being sustained by more than one team in the history of the NFL. So you could analyze the self-correlation across seasons of all NFL teams' interception rates, say since the 1978 rule change. That would give you a large enough sample to indicate whether the ability to get a large number of interceptions is a "sustainable skill", or whether the number of interceptions that GB's been able to get in the past three seasons is more likely to be statistical noise.

If there is no evidence of interceptions being a sustainable skill in the last 40+ NFL seasons, that doesn't mean it's impossible that GB has now developed a new and unique skill that no one in football has ever developed before. Perhaps there's some confluence of Dom Capers' coaching methodology (even though he's been coaching in the NFL for the past 25 years) and the talent in the GB defense (even though I don't think anyone is claiming that GB's defense has some heretofore-unseen talent and ability).

But I wouldn't bet money on it.