Week 4 DVOA Ratings

Week 4 DVOA Ratings
Week 4 DVOA Ratings
Photo: USA Today Sports Images

by Aaron Schatz and Danny Tuccitto

Despite their close loss to Cincinnati, the Buffalo Bills continue to lead the Football Outsiders DVOA ratings after four weeks. The Bills are in a close pack of teams at the top that also includes their division rivals in New England and the surprising Tennessee Titans (who will be the subject of an ESPN column later this week). Green Bay is right behind those three AFC teams in DVOA, but ranks first in our "estimated wins" metric because they've been particularly strong in red zone defense and in the second half of close games. The undefeated Packers also rank second in DAVE, which accounts for both DVOA and our preseason projections, and the Packers are currently the leading candidate to win this year's Super Bowl with a 15.7 chance according to the Football Outsiders playoff odds.

Week 4 is when we introduce opponent adjustments into our formula for the first time. The hardest schedules so far have belonged to Denver and Oakland; the easiest have belonged to Tennessee, Atlanta, and the New York Giants. The Giants rank 32nd in schedule so far and first in schedule remaining. (Note that the schedule ratings are based on DVOA, not DAVE.) 

The biggest gap between DVOA and non-adjusted VOA belongs to Detroit, which explains why the undefeated Lions are ranked just ninth in DVOA. Not only have the Lions had the league's 26th-ranked schedule, but they've also fumbled the ball five times on offense without losing one. The Lions actually have a below-average -12.6% DVOA for their narrow Week 3 win over Minnesota and a 1.6% DVOA for this week's win over Dallas.

For your enjoyment this week, we have a couple of graphics that Danny Tuccitto has been working on. With teams like Buffalo and Tennessee surprising so early, we were curious at what point early-season DVOA really established how good teams would be for the entire season. The answer appears to be "pretty early." This graphic shows you the correlation of DVOA after a certain week to final DVOA. The black line right above 0.7 represents when r^2 passes 0.5 -- in other words, when the current early-season DVOA accounts for at least 50 percent of the variability in the final DVOA. As you can see, offensive DVOA and total DVOA pass that line as of Week 4. So there's a pretty good chance that teams like Buffalo, Tennessee, and Detroit are going to be good all year. The correlation is even pretty high after Week 3 -- before opponent adjustments are applied. (We should note that, obviously, final DVOA does include the first four weeks of games, which is one of the reasons the correlation is so high.)

This next graphic looks instead at the correlation of weekly DVOA to final wins. This one will be lower, of course, because at the end total DVOA won't correlate exactly with wins the way it will with itself. Still, this shows that you can explain 50 percent of the variability in final win total with DVOA as of Week 7.

Anyway, on to the ratings. All stat pages are now updated through Week 4. The FO Premium database of DVOA splits is updated as well.

For the many people asking, we're trying to get the midseason KUBIAK update out by Thursday.

* * * * *

These are the Football Outsiders team efficiency ratings through four 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.

Opponent adjustments are currently at 40 percent strength and will steadily grow stronger until Week 10. As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.

DAVE is a formula which combines our preseason projection with current DVOA to get a more accurate forecast of how a team will play the rest of the season. Right now, the preseason projection makes up 40 percent of DAVE.

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 BUF 35.4% 1 19.9% 4 3-1 39.2% 3 5.6% 14 1.9% 12
2 TEN 33.4% 2 12.2% 10 3-1 26.1% 6 -10.9% 4 -3.6% 25
3 NE 33.1% 5 31.6% 1 3-1 44.8% 1 15.0% 27 3.3% 10
4 GB 29.7% 8 22.4% 2 4-0 39.2% 2 8.3% 18 -1.3% 19
5 BAL 24.2% 3 20.2% 3 3-1 -1.1% 22 -30.1% 1 -4.9% 29
6 NYG 23.4% 6 16.8% 5 3-1 21.4% 8 -2.7% 9 -0.7% 17
7 HOU 21.5% 9 15.7% 8 3-1 26.7% 5 11.1% 21 5.9% 6
8 NO 20.6% 13 16.3% 6 3-1 25.4% 7 6.8% 17 2.0% 11
9 DET 17.3% 4 9.4% 11 4-0 4.4% 15 -14.4% 3 -1.5% 20
10 NYJ 15.3% 7 16.2% 7 2-2 -15.1% 29 -16.5% 2 13.9% 1
11 SF 14.6% 12 6.1% 14 3-1 -3.4% 23 -5.0% 7 13.1% 2
12 OAK 11.7% 10 1.2% 19 2-2 32.8% 4 17.0% 28 -4.1% 28
13 TB 10.8% 19 3.5% 15 3-1 15.5% 10 13.3% 25 8.7% 4
14 CIN 10.4% 11 1.6% 18 2-2 0.7% 19 -9.9% 6 -0.2% 15
15 WAS 9.6% 15 1.6% 17 3-1 -0.4% 21 -10.4% 5 -0.4% 16
16 ATL 4.9% 16 6.5% 13 2-2 6.3% 14 1.2% 10 -0.1% 14
17 DAL 1.6% 21 -1.8% 21 2-2 1.7% 17 -3.7% 8 -3.8% 26
18 MIN 0.5% 14 -0.7% 20 0-4 3.9% 16 9.6% 20 6.2% 5
19 PIT -0.2% 17 12.8% 9 2-2 0.8% 18 4.8% 12 3.8% 9
20 SD -2.7% 23 7.3% 12 3-1 10.5% 12 9.3% 19 -3.9% 27
21 CLE -5.6% 18 -4.1% 22 2-2 -4.6% 24 6.8% 16 5.8% 7
22 ARI -11.9% 20 -13.6% 25 1-3 0.5% 20 13.7% 26 1.3% 13
23 CHI -12.9% 28 -5.0% 23 2-2 -12.3% 27 12.3% 23 11.7% 3
24 PHI -13.1% 22 1.8% 16 1-3 11.0% 11 21.6% 32 -2.6% 22
25 DEN -15.3% 25 -15.0% 26 1-3 -6.9% 25 13.3% 24 4.9% 8
26 CAR -15.8% 24 -15.2% 27 1-3 16.2% 9 19.2% 30 -12.8% 32
27 MIA -16.0% 26 -6.1% 24 0-4 8.2% 13 21.5% 31 -2.6% 23
28 SEA -24.0% 29 -23.2% 30 1-3 -11.0% 26 6.5% 15 -6.4% 30
29 IND -30.0% 27 -23.0% 29 0-4 -14.4% 28 4.8% 13 -10.7% 31
30 KC -36.1% 32 -26.5% 31 1-3 -23.5% 31 11.6% 22 -1.0% 18
31 JAC -38.3% 30 -22.2% 28 1-3 -31.1% 32 4.2% 11 -3.0% 24
32 STL -40.5% 31 -29.8% 32 0-4 -21.1% 30 17.3% 29 -2.1% 21
  • 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 BUF 35.4% 3-1 37.5% 3.0 4 4.8% 14 3.8% 11 8.2% 20
2 TEN 33.4% 3-1 33.0% 3.0 2 -8.7% 30 0.0% 18 7.2% 19
3 NE 33.1% 3-1 35.7% 3.0 3 7.1% 10 -2.1% 25 18.6% 27
4 GB 29.7% 4-0 34.1% 3.5 1 -5.8% 25 -0.5% 22 4.9% 15
5 BAL 24.2% 3-1 33.2% 2.3 13 2.0% 15 -4.1% 28 33.9% 32
6 NYG 23.4% 3-1 24.4% 2.6 7 -14.0% 32 10.4% 1 13.8% 25
7 HOU 21.5% 3-1 26.1% 2.7 6 -6.4% 27 -2.7% 26 4.8% 13
8 NO 20.6% 3-1 21.1% 2.8 5 0.0% 18 -0.1% 19 5.6% 17
9 DET 17.3% 4-0 30.3% 2.5 10 -5.8% 24 3.9% 10 10.4% 22
10 NYJ 15.3% 2-2 18.0% 2.1 18 -0.2% 19 5.9% 4 20.4% 28
11 SF 14.6% 3-1 20.0% 2.5 9 -6.3% 26 -4.9% 31 1.3% 6
12 OAK 11.7% 2-2 11.1% 2.6 8 17.2% 2 -4.6% 29 4.2% 12
13 TB 10.8% 3-1 14.5% 2.3 14 5.7% 13 5.3% 5 4.8% 14
14 CIN 10.4% 2-2 19.8% 2.4 11 7.3% 9 -1.4% 23 1.2% 5
15 WAS 9.6% 3-1 10.0% 2.4 12 -6.8% 28 2.3% 13 3.5% 10
16 ATL 4.9% 2-2 3.3% 2.1 19 -9.8% 31 4.5% 9 0.2% 2
17 DAL 1.6% 2-2 2.6% 2.2 16 9.9% 7 1.4% 15 0.9% 4
18 MIN 0.5% 0-4 4.6% 2.2 15 6.4% 12 4.6% 8 4.9% 16
19 PIT -0.2% 2-2 -5.7% 2.1 17 -2.1% 22 -3.8% 27 32.5% 31
20 SD -2.7% 3-1 4.4% 2.0 20 -4.6% 23 1.3% 16 10.7% 23
21 CLE -5.6% 2-2 0.2% 1.8 21 -0.5% 20 -1.9% 24 0.2% 1
22 ARI -11.9% 1-3 -5.8% 1.7 22 -1.7% 21 -4.8% 30 3.5% 11
23 CHI -12.9% 2-2 -6.9% 1.7 24 9.8% 8 -0.3% 20 10.0% 21
24 PHI -13.1% 1-3 -10.6% 1.6 28 0.6% 17 5.3% 6 7.2% 18
25 DEN -15.3% 1-3 -18.8% 1.6 25 21.3% 1 -0.4% 21 0.4% 3
26 CAR -15.8% 1-3 -15.2% 1.6 26 -8.4% 29 10.0% 2 3.1% 9
27 MIA -16.0% 0-4 -19.3% 1.7 23 11.6% 5 9.7% 3 2.1% 8
28 SEA -24.0% 1-3 -25.9% 1.6 27 1.8% 16 -5.3% 32 21.2% 29
29 IND -30.0% 0-4 -33.9% 1.1 29 6.6% 11 2.7% 12 13.7% 24
30 KC -36.1% 1-3 -47.2% 0.9 30 12.5% 4 0.8% 17 25.0% 30
31 JAC -38.3% 1-3 -36.1% 0.8 32 13.4% 3 4.9% 7 17.6% 26
32 STL -40.5% 0-4 -47.4% 0.9 31 11.0% 6 2.2% 14 1.8% 7


127 comments, Last at 10 Oct 2011, 12:30pm

1 Re: Week 4 DVOA Ratings

Wow. Philadelphia with the worst defense in the league. Also, Buffalo, Tennessee, and Oakland all have better offenses than New Orleans.

32 Re: Week 4 DVOA Ratings

Wow. Philadelphia with the worst defense in the league.

And the Bears at #23. I initially looked at the three individual DVOA numbers for CHI and thought "they screwed up the total, because the offensive & defensive numbers cancel, and special teams is positive." But no, that defensive number is positive. Damn.

It's perfectly consistent with what I've seen on the field, mind you, it's just surprising.

34 Re: Week 4 DVOA Ratings

I think I may have really underrated Manning. All the safeties were putting out there are a step slow in their recognition.

37 Re: Week 4 DVOA Ratings

I am totally willing to believe that Steltz and Conte are significantly worse than Manning, but when they had the actual starters out there (week 1), things went the way they're supposed to in cover-2 (lots of underneath completions, no receiver ever running free behind everyone). I want to believe that Meriweather just needs a bit more experience in cover-2, and that plays like the long pass to Smith won't happen again.

40 Re: Week 4 DVOA Ratings

Steve Smith is killing everyone because he has a guy who gets the ball within 5 yards of Steve which apparently is all that's necessary.

61 Re: Week 4 DVOA Ratings


Patriots fans have a few years of game film we'd like you to watch on Brandon "The human missile" Meriweather. I'm sure he needs more experience in cover-2, but I'm also pretty sure that plays like the long pass to Smith will definitely, positively, absolutely happen regularly.

64 Re: Week 4 DVOA Ratings

In reply to by Malene, Copenhagen (not verified)

I'm pretty sure NE runs lots of Cover 2 principles in the secondary. Their front seven has traditionally been a 3-4 scheme but Belichick almost always has 2 deep safeties and a MLB bailing out. Sure its not always the same guy but Merriweather has seen plenty of action as a Cover 2 safety.

65 Re: Week 4 DVOA Ratings

In reply to by rtwhite78 (not verified)

Yup - and that's why they gave up a first round "Pro Bowl" safety for nothin. Absolutely nothin.

75 Re: Week 4 DVOA Ratings

I feel like they run more Cover 3 than Cover 2 but that's just a hunch based on the infuriatingly large cushions the corners always seem to give everyone.

69 Re: Week 4 DVOA Ratings

It is a little soon to start calling out Conte but Steltz shouldn't be on an NFL roster. There isn't anything he does well and is slow and not particularly savvy. I really have no idea what the Bears coaches see in him. He isn't a key specials players either. I guess he shows up every day and knows everybody's names - this seems to be enough to keep him on the roster.

Meriweather isn't the answer either, if he was he would be playing for New England. Doesn't play to schemes (and never has) and doesn't wrap up when he tackles (and never has). It is beyond me why the Bears are paying this dork $3.5m instead of trying to use that money to keep Manning. I blame Ruskell (it fits his MO).

Getting Harris back will help, especially with discipline but Major Wright is going to keep getting neck and head injuries if he leads with the top of his helmet when he tackles.

The secondary is going to be an issue all year, too many question marks.

99 Re: Week 4 DVOA Ratings

I liked the style of offseason the Bears had. Bring in a bunch of one highly thought of players who underperformed. It worked pretty well with Okoye, didn't work with Gholston and is kind of working with Meriweather. I don't see why the Bears couldn't have kept Manning and made these moves though.

115 Re: Week 4 DVOA Ratings

They had an odd offseason, everyone expected them to go out and try to add a big name to push them over the top but instead they added added veteran depth and gave young players a chance to help them improve from within. I still think this team needs better production from DE2 and LCB, or the top passing teams can shred them.

Okoye being the best pickup probably tells us why he has worked out where others didn't. I would argue that whoever tried to make him into a two gap lineman needs his brain feeling for lumps. Grabbing a talented player and maximising their production is a nice bonus but you can only do this if other teams have screwed up their player development. I would argue that you are going to find more players where this applies in the front seven than anywhere else on the roster (maybe offensive line where some coaches are rather narrow in their run schemes).

123 Re: Week 4 DVOA Ratings

I think there is quite a bit of variation in secondary styles too. The Raiders run bump and run almost exclusively, the Bears run zones most of the time. Other teams like the Eagles run a lot of man, and other teams run a lot of combinations.

124 Re: Week 4 DVOA Ratings

Fair point. To some extent your CBs need to be able to do both man and zone or you end up very vanilla. I supppose you can disguise leverages and vary the number of deep men better if you stick to exclusively man coverages. The Bears problems aren't in their zone schemes it is that the CBs struggle like hell with playing man when they go cover-1 on third and short; they lose leverage or just don't keep up.

Front sevens tend to be either two gap or one gap. A mixture rarely ends well.

2 Re: Week 4 DVOA Ratings

Can you somehow factor Really Bad White Guy Hair as a DVOA positive? Between Curtis Painter and Pat McAfee, that would really help the Colts' ranking.

114 Re: Week 4 DVOA Ratings

One of my friends posed for a picture with him somewhere on a beach this summer and uploaded it to facebook. Hilarious comments ensued. He looks more like the guy who plays left field on my softball team than a professional athlete.

42 Re: Week 4 DVOA Ratings

They kind of won. The Chiefs stoppable force met the Vikings movable object. During this embarrassment of wooses, the Vikings lost. Since NFL rules mandate that if one team loses the other team must have won, the Chiefs were declared the winner. NFL rules may need to be rewritten to allow both teams to lose, which will partially balance all those games that moronic color commentators claim neither team deserves to lose.

4 Re: Week 4 DVOA Ratings

Did the Bengals not play that well this week? I'd figure holding the #1 offense to 13 points would have shown up well, but they dropped a place in defensive rankings.

8 Re: Week 4 DVOA Ratings

Remember that their position will also be affected by how their previous opponents performed and how others near them in the rankings (and their opponents) played.

9 Re: Week 4 DVOA Ratings

I don't think the opponent adjustments are very strong. For instance, the Ravens have -1.1% offensive DVOA but have faced PIT, TEN, and NYJ defenses, as the season goes along that rating should get better.

5 Re: Week 4 DVOA Ratings

Offense correlates with wins better than defense does? That's curious. Or at least, interesting.

12 Re: Week 4 DVOA Ratings

It shouldn't be that surprising. Think about the teams that have been consistently good. The pats, colts, saints, sd, Pittsburgh, philly have all had top 10 offenses. Only the jets and ravens have had strong defense without good offensive play. ,

15 Re: Week 4 DVOA Ratings

Put another way, scoring points is better correlated with wins than stopping e opponent from scoring points. I think it makes more sense when put in that light. Stopping the opponent definitely helps, but you still need to score in order to win.

19 Re: Week 4 DVOA Ratings

You have to score more points than your opponent does to win. When this happens, it can be interpreted as
(your offense) - (their defense) >
(their offense) - (your defense)

In other words, the forces should be symmetric.

And yes, this topic has been discussed before.

The most natural expectation for how DVOA works is that it would be normalized so that the expected [mean] value of a game [summing the scores of both teams] would be zero. There are a few reasons this isn't so. One is that it is normalized using historical data, and since the league is becoming dominated more and more by offenses, the historical data no longer represents the mean behavior of a football game accurately.
Another reason is that there are small loopholes (such as penalties IIRC) that allow for the mean to vary from zero.

I'm getting off topic back to an earlier topic. We were talking about correlation.

"Stopping the opponent definitely helps, but you still need to score in order to win."

This doesn't bring us much further to an understanding.

I blame the Patriots, with their high-rated offense and bizarrely incompetent defense for throwing the whole system out of whack.

23 Re: Week 4 DVOA Ratings

Is this not probably about lower scoring games being more likely to be won by the inferior team, on the grounds that lower score -> lower absolute points margin -> greater impact from fluky plays. Good offense decreases variance (especially in mismatches); good defense increases it. Possibly.

104 Re: Week 4 DVOA Ratings

I disagree that the forces should be symmetric, because you're not looking at sample size for each. I think what's going on is something akin to Simpson's paradox.

If your offense is good, you will control the ball longer, so you will continue to have your good offense affecting the game.

If your defense is good, the other team will control the ball less, so you will have your good defense affecting the game less.

In other words, a team with a good offense will give itself more opportunities to have its better half affect the outcome of the game.

A team with a good defense will give itself fewer opportunities to have its better half affect the outcome of the game.

Hence having a good offense is more useful for winning.

113 Re: Week 4 DVOA Ratings

The assumption here is that good offenses have long times of possession.

A cursory review of the recent history of the Indianapolis Colts suggests otherwise.

22 Re: Week 4 DVOA Ratings

There's a greater spread in offensive DVOA between good and bad teams, as compared to defensive DVOA. I wonder if that has something to do with it.

50 Re: Week 4 DVOA Ratings

Yes, I think that's it. The great offenses are way better than the mediocre ones, but there is less difference among the defenses. Hence the overall difference between teams is dominated by the difference in offensive quality.

86 Re: Week 4 DVOA Ratings

If it's true (I haven't gone and checked) that Offensive DVOA has a higher variance than Defensive DVOA, that would absolutely create a larger correlation coefficient.

Assume the true data-generating process is:

[Wins] = a + b * [Offense] + c * [Defense] + d * [Special Teams] + e,

where a is a constant and e is noise (a standard linear regression model).

If the true effects of Offense and Defense are equal (b=c) but Offense has a higher variance, the correlation of Wins and Offense will be higher than that of Wins and Defense.

This is why in general correlational statistics are not that informative. It could be that Defense does more "work" in creating wins that Offense, but that this relationship is obscured by using correlation (which is also not comparable across samples) rather than, say, unstandardized regression coefficients, from which we can see actual marginal effects.

6 Re: Week 4 DVOA Ratings

At some point, the Vikings offensive ranking must start to approximate what I am seeing, does it not?

27 Re: Week 4 DVOA Ratings

They are 21st in the league in yds/play and have only turned it over 3 times. Sounds pretty mediocre to me as opposed to putrid.

I think they have looked pretty good at times and horrendous other times and the horrendous part sticks in your mind.

To me the defence has been more responsible for the loses than the offence - but that's kind of splitting hairs.

39 Re: Week 4 DVOA Ratings

Ya' gotta point about the turnovers, but in this day and age, if you can't go downfield when it is demanded, you just don't have much. Between inaccuracy and bad receivers, there just is nothing there.

41 Re: Week 4 DVOA Ratings

just have to outlaw the forward pass and we've got a squad.

One thing I have liked is the use of Harvin running the ball. Harvin is a great runner.... It's not just that he's fast, he always seems to make more yards than you would think and breaks so many tackles.

10 Re: Week 4 DVOA Ratings

Maybe New Orleans has turned it over a lot more than Tennessee on Offense and Tennessee's games have had very few numbers of Possessions? I do agree though that I would much rather play vs Tenn's Offense than NO's, especially considering that 2x already Tenn has been held to 17 or less

18 Re: Week 4 DVOA Ratings

Tennessee also put up 26 against Baltimore, which is by far the strongest defense according to DVOA. Glancing at last week's ratings, it seems like the opponent adjustments bumped TN up and NO down. (makes sense when you think about it... denver is the only bottom-half defense TN's faced, while jacksonville is the only top-half defense NO's faced.)

20 Re: Week 4 DVOA Ratings

The comparison of Baltimore's effort against Pittsburgh to their effort against Tennessee shows why jump models are preferred in some settings.
In a jump model, you wouldn't have one value for an expected output from Baltimore in any game. Rather, you would postulate the existence of different levels of effort from the Ravens (high effort vs. low effort) which would have different expected results. A jump model would undoubtedly be more accurate in terms of modeling the fact that teams play differently from week to week. But of course, this kind of model is far more likely to fall into the trap of over-fitting given the small amount of data involved (only 16 games/season).

My point here? Don't be surprised if the Titans' rating slips in the future, since it's likely they are overrated from taking advantage of an uncharacteristically weak effort by the Ravens.

I could say the same about the Bills, but my friends accuse me of having a negative bias about any AFC East team not coached by Belichick. (But really...don't count on having that kind of luck with tipped balls again, Bills fans.)

14 Re: Week 4 DVOA Ratings

I'm surprised at the difference in the future schedule for NE and NJ. They were supposedly going to have it a lot harder from the preseason projected DVOA Future Schedule.

21 Re: Week 4 DVOA Ratings

For the Pats, the games against Indy and the NFC East (except for the Skins) probably look easier than they did a month ago. The Jets don't have Indy but they do have KC, who have also plummeted.
OTOH, Buffalo is doing much better than expected. Well, the Jets have them twice and the Pats only have them once more, so that helps explain why the Jets' schedule is rated as tougher than the Pats'.

29 Re: Week 4 DVOA Ratings

Actually, every team in the NFCE, except the Eagles, has moved up in the rankings since preseason. The big movers still on the Pats schedule are MIA (10 to 27), PHI (3 to 24), PIT (1 to 19) and IND (15 to 29). And they've already played SD (preseason #4). Wow - I guess I see why it's changed so much.

16 Re: Week 4 DVOA Ratings

The Eagles is clearly ranked because the Eagles have played the toughest schedule in the NFL. Any rating that takes into account the fact that the Eagles have played the toughest schedule in the NFL is way better than this. I mean they are the Dream Team, so anyone who beats them is great, and they've lost 3 times, so they must have the toughest schedule 'cause I don't see any 0-4 Dream Teams out there.

17 Re: Week 4 DVOA Ratings

Why not show the correlation between week 4 DVOA and week 5-16 DVOA so that it's out of sample and it gives us a better indication of what it means? This would tell us how much information we have on the future of the team.

24 Re: Week 4 DVOA Ratings

The Titans ranking I find kind of surprising, but what's really shocking is the Giants ranked 6th. I haven't really paid much attention to them, but from what I have seen, they seem average. Have they been recovering from all those early injuries?

78 Re: Week 4 DVOA Ratings

Few turnovers (including recovered fumbles) and at least "pretty good" in every offensive/defensive area (lowest category ranking is 12th in run D) gets you a pretty high DVOA ranking. The passing offense has looked more explosive than in the past, and they're the best red-zone offense in the League (partly a sample-size issues; many of their TDs have been over 20 yards, so they haven't had many opportunities to settle for a short-range FG).

As for the injuries, they've mostly adapted rather than recovered. Aaron Ross is a weaker starting CB than Terrell Thomas was, but he hasn't been a pure liability. They've filled Jon Goff's MLB role with 2 6th-round rookies, Greg Jones in the base offense and Jacquian Williams in the nickel. (Williams, in particular, looks like a find.) Undrafted rookie Jake Ballard is the starting TE, and has made some nice plays, as well as gotten more reliable in his blocking. Osi Umenyiora is the only major "recovery" to date, but his return this week corresponded with Justin Tuck missing a game due to a setback on the neck issue that kept him out of week 1 (this injury is my biggest worry about this team).

And some of their players are emerging and/or improving. You all saw the Victor Cruz highlights, and Jason Pierre-Paul is racking up a highlight reel of his own. Linval Joseph (a 2010 2nd-round pick who took Barry Cofield's old starting spot) looks like he belongs as a starting DT. Will Beatty isn't looking totally lost in his first year as the starting LT, although he's lucky that his matchups against premier pass-rushers (Orakpo, Cole) happen to have been games officiated by crews who don't call much offensive holding. Michael Boley has been having by far his best season as a Giant. Although it can be tough to judge DB play on the network TV film, the above-mentioned Aaron Ross seems to be improving with each game.

81 Re: Week 4 DVOA Ratings

I'll take your word for it, but I saw the Giants play poorly against a StL team more determined to lose. They also looked merely average against a pretty terrible Cardinal squad as well.

One good game out of four doesn't usually translate into such a strong rating.

82 Re: Week 4 DVOA Ratings

I watched their first three games. I don't disagree with a lot here. Ross though does look like a liability, and it's not like he's played really well in the past either. He got benched in the St. Louis game. His two INTs against the Eagles were a fortuitous bounce off Steve Smith's hands, and a ball that Kafka shouldn't have thrown. I'm not sure about Cruz either. He looked poor against Washington, and the team had such faith in him that they brought in Brandon Stokley. His really long TD against the Eagles looked more like the product of Kurt Coleman deciding not to tackle him. Beatty hasn't impressed me either, although I haven't focused that much on him. I agree about the holding calls. The long pass from Eli to Nicks in the 1st quarter of the Redskins game - Beatty held Orakpo so bad I was surprised the officials missed it.

28 Re: Week 4 DVOA Ratings

Seems so odd to see Buff, Det, SF, Oak and Cinn in the top half of the league and Indy, Philly, SD and Pitt in the bottom half.

73 Re: Week 4 DVOA Ratings

Well, the playoff odds report says they have an 11.4% chance of winning the conference and a 5.8% chance of winning the Super Bowl, so I'd say 5.6%.

33 Re: Week 4 DVOA Ratings

I'd love to see the one game DVOAs rather than the cumulative DVOAs in these charts. The correlations would be lower (obviously) but at last we'd know which week is the true Jump to Conclusions Week.

Is there a week where "getting the bugs out of the playbook" balances out "losing good players for the rest of the season" so that teams are playing "like themselves" more than any other time?

35 Re: Week 4 DVOA Ratings

I actually sent an email about that a few years ago and the response (if I remember correctly), was no, there isn't a noticeable spike/curve for dvoa by week

36 Re: Week 4 DVOA Ratings

I'm a bit surprised to see the Giants that high. Yes, they're 3-1, but they haven't impressed me at all when I've had the chance to watch them. I didn't see any of the Philly game, so maybe that's when they looked awesome.

38 Re: Week 4 DVOA Ratings

I have watched them, and no they have not looked impressive. Eagles outplayed them, the Giants just had a few big plays and stopped them on crucial red zone possessions and one 4th down. They didn't beat the Rams as much as the Rams beat themselves. They beat the Cardinals but only barely and somewhat due to a controversial call. They could just as easily be 0-4 as 3-1.

43 Re: Week 4 DVOA Ratings

Offensive line hasn't gelled, and the defense is having growing pains and depth issues at linebacker and cornerback. The linebacker issue coupled with the loss of Thomas (who was great in run support) has left us vulnerable to the run.

Thomas isn't coming back. But we can deal with that if Prince can contribute, and if the linebackers are limited right now by inexperience and not by ability, and if the offensive line can gel.

The G-men have shown enough everywhere else that if those three ifs hold true, and nothing catastrophic from an injury perspective happens further, then we're going to be in the mix right to the end.

If not? Well, .500 might be a reach.

45 Re: Week 4 DVOA Ratings

The thing about the Giants is that they're consistently inconsistent. When I checked the score on my phone and saw they were losing to ARI by 10 in the 4th, I wasn't the least bit surprised. When they scored two TDs in the final 5 mins, I still wasn't surprised in the least. They have a turnover prone offense and a great pass rush that causes turnovers; neither of those is very consistent from week to week, and it leads to a lot frustrating games for Giants fans.

The biggest problem I see going forward is their #32 ranked schedule for the next 13 weeks.

44 Re: Week 4 DVOA Ratings

I've watched every game (bias warning, I am a fan) and while I wouldn't put them at 6th, I could easily see top ten. Eli manning had a disaster of a first game, but the other three were all solidly above average in my opinion (and by just about every metric used to rate QBs). The offensive line is a concern and the defense has been severely limited by injury.

Their strengths are very strong and their weaknesses are very weak. I'm expecting a roller coaster of a season.

Hakeem Nicks is a hell of a player, I would compare him with young TO but with much better hands. He is going to make Manning look a lot better than he is (he is above average, but not by a lot). The Giants pass rush, and Justin Tuck in particular, have shown the ability (admittedly inconsistent) to take over a game against even the leagues greatest offenses ('07 SuperBowl baby).

As for the weaknesses, the offensive line is a liability in pass protection and their run blocking isn't what it used to be. The back seven on Defense are a big concern. The LBers can't stop the run without safety help and the DBs can't cover without safety help. Ross is an absolute liability in coverage, rookie CBs rarely make a difference but Prince has to be better than Ross.

52 Re: Week 4 DVOA Ratings

Good point on Nicks. He is a beast and is the one guy that clearly stood out in the games I've watched. Their track record on drafting WRs (or finding undrafted free-agent WRs) has been quite impressive the past few years.

46 Re: Week 4 DVOA Ratings

Wow. Philadelphia with the 32nd ranked defense.

Has there ever been a team that has invested so much money on the defensive side of the ball (Asomugha, Jenkins, Babin contracts + bloated Samuel contract on the books) and had so little to show for it?

I mean, you could just field a defense full of undrafted free agents instead of all of those high-priced cornerbacks and have the same basic result - the 32nd ranked defense in the league.

Amazing. I didn't except them to be #1 after all of those hyped acquisitions, but I certainly didn't expect them to be the worst defense in the league, either.

88 Re: Week 4 DVOA Ratings

You need a defensive coordinator to field a defense. The Eagles simply don't have one.

Also, please note how outlandish it is that the Eagles defense leads the league in sack rate, are Top 10 in power run situations and middle of the pack in terms of runs stuffed... so they have a great defensive line! Asomugha and Samuel didn't just become awful. This is... gah.

Certainly, this is the worst Eagles defense since the late 70's.

100 Re: Week 4 DVOA Ratings

This is obviously worse than expected, but ever since Jim Johnson died, the defense is in a downward spiral. It went from 3rd in 2008 (last year of Jim Johnson) to 6th, to 14th and now to 32nd. I really doubt it will end up 32nd, but I think that downward trend will continue.

48 Re: Week 4 DVOA Ratings

Looks like the 49ers offense "explosion" was due more to Philly's ineptitude than actual improvement. (Though it did look like Harbaugh opened up the offense in the 2H.)

At least Tampa is also bad on defense. Excited to see if the Niners can go to 4-1 before facing the "juggarnaut" Detroit Lions.

49 Re: Week 4 DVOA Ratings

My hope for this season is to have the Vikings go winless while maintaining a positive DVOA.

53 Re: Week 4 DVOA Ratings

Obvious (to me at least) follow-up question after showing those correlation graphs.

What is the correlation between DAVE and the final DVOA of a team in each week? What about DAVE to wins?

Is DAVE actually better than DVOA at predicting final DVOA or wins early in the season? How far into the season?
Are the DAVE weights finely tuned to create the best correlation possible or just SWAG (scientific wild ass guess)?

59 Re: Week 4 DVOA Ratings

Good question, and I hope someone can give us a real answer. I would expect DAVE to be close to worthless as a predictive tool, though. I don't even look at it, because who cares?

62 Re: Week 4 DVOA Ratings

I think that DAVE's main function is to forestall the "How can you rank [team] so high/low after a fluky week or two?" questions.

63 Re: Week 4 DVOA Ratings

Pretty sure DAVE is vastly more predictive than early season DVOA. I can't remember when the crossover takes place, though - maybe around now or the next few weeks?

54 Re: Week 4 DVOA Ratings

I find it hard to believe that the Falcons defense is ranked as high as tenth. I mean, they just made Tavaris Jackson look like Tom Brady on Sunday.

56 Re: Week 4 DVOA Ratings

I'm having trouble wrapping my head around the Lions' no. 1 DVOA ranking on STs, given that they've given up 3 long kick returns and haven't busted many long ones themselves. I can only guess that it's all about the K and P.

67 Re: Week 4 DVOA Ratings

Something is wrong here. Vegas has Philly a 3 point (1.5 on some books) favorite on the road against the Bills. Either we believe the professional oddsmakers or we think the DVOA has some merit and the Bills should win in a blowout. C'mon, we should re-mortgage the house and put it all on Buffalo if we had any faith in this system...

71 Re: Week 4 DVOA Ratings

Well, first of all, the Eagles really need a win and that usually makes a team the favorite in Vegas. Plus, unlike DVOA, Vegas factors in swagger. Any team that loses to the Bengals obviously loses their swagger, ergo, Vegas thinks they'll lose.

91 Re: Week 4 DVOA Ratings

Its not what "vegas thinks" or what the "oddsmakers" decide.

They simply attempt to get as close to 50% of the bets on either side as they can. So its what the "betting public" thinks.

95 Re: Week 4 DVOA Ratings

That often gets trotted out, and it is a fallacy - or at least it should be. The line should not be determined by the volume of bets, but more by 'who' is betting. The vast majority of the betting public are fundamentally stupid, irrational and partisan, and will lose money to the house in the long run. So why would I set my line according to what they think? The skill of the bookmaker is to set the line as accurately as possible, then identify who the clever punters are who will win money in the long run and move the line if they bet. If one very clever punter bets one side of a line, and 1000 stupid punters bet the other, that's a job well done. And if no clever punters bet your line, it's an indication that the line is correct and there is no need to move it. That is a far more profitable long term strategy than simply balancing your book at all costs.

96 Re: Week 4 DVOA Ratings

That strategy also maximizes risk on a very volatile transaction. If, by chance, those 1000 stupid punters happen to get it right, you lose everything. If you balance it 50/50, you win no matter what.

109 Re: Week 4 DVOA Ratings

Well, yes, oddsmakers do want to set a line that makes sense. But in the end, all they really want is even money on both sides. You can see why if you imagine that you were a bookie and your clients were, say, 50 of your high school class mates, or college friends, or guys at the bar or whatever. It doesn't matter have savvy you are at setting a line, if 40 of your stupid friends take the home team to cover on what you think is a sucker bet-- but end up getting it right-- then you're in trouble.

Sure, you might make the right call sometimes and collect a lot of cash, but all it takes is one screw-up to put you out of business (and then some). To protect yourself, your strategy has to be to get even betting on both sides of the line.

111 Re: Week 4 DVOA Ratings

Yeah, clearly you don't want to be risking making yourself bankrupt because of one (or two, or three) bad result(s). But presumably if you are a professional oddsmaker (e.g. a casino) you have capital reserves large enough to cover the occasional heavy loss, provided your business model is sound and you make money in the long run.

122 Re: Week 4 DVOA Ratings

Well, look: clearly, these casinos with their large capital reserves know what they're doing. I think the major flaw if your argument is that bookmakers can truly gain a significant edge in their superior knowledge. It sounds like you're imagining that sports books are significantly better than the average bettor-- but is that really the case? Most professional gamblers make a living being right just a little over 50% of the time-- and they do so having the luxury of having access to multiple sports books and finding that one book with the spread that's an extra half point the the direction they want.

If you're a book though, how do you imagine your system works? , Do you really want to toss out a line based on your "superior knowledge" and see 60+% of your gamblers line up on one side of that line? Why risk a sure thing for the chance at a greater sum of money with no guarantee of success? The reality is that sports books aren't in the business of gambling. They aren't interested in risking lots of money; they're interested in making lots of money.

You seem to think that you've got a foolproof system. Why no pitch your plan to a Vegas Wise Guy and see how it goes?

74 Re: Week 4 DVOA Ratings

In addition to the above, Vegas also considers match-ups. Philly's D is weakest up the middle, Buffalo's O is strongest on the perimeter. However, logic like this often doesn't take into account magnitude.

83 Re: Week 4 DVOA Ratings

Vegas is setting a line to maximize profits, not to make a prediction about the game. If the bookmakers think that gamblers think the line should be Eagles -3, that's where the line will be.

Yes, there is some aspect of risk minimization. But if that's where the line is, it's because there's a lot of money on the Eagles relative to the Bills.

DVOA can only measure how well the Eagles and Bills have played. Past performance isn't always the best predictor for future performance (even though, all other things being equal, it is a pretty good one.) Presumably a lot of gamblers think the Bills are due for a fall and that the Eagles are likely to start playing better. (I would agree with at least one of those two predictions.)

97 Re: Week 4 DVOA Ratings

It would be interesting to hear from someone who has actually worked in the sports book of a major Vegas concern, because I hear both opinions stated as gospel over and over. The one thing that makes me think that "Vegas" can't be 100% concerned with minimizing risk (by having an equal share of money on both sides of a game) is that they employ lots of people who do nothing but analyze the NFL and are way better at it than the rest of the world. That seems like a major financial outlay if the only benefit is setting a more accurate opening line.

98 Re: Week 4 DVOA Ratings

And AIG had a full time staff of mathematicians who did nothing but analyze risk, and were a lot better at it than the rest of the world.

118 Re: Week 4 DVOA Ratings

Well...apparently not since AIG simply ignored the possibility of risk correlation. If they had good mathematicians, they sure as hell weren't listening to them! Either that or their business model hinged upon "too big too fail."

But I digress...

120 Re: Week 4 DVOA Ratings

After reading the Big Short, it seems like the derivatives divisions simply didn't tell the risk analysts what they were doing.

76 Re: Week 4 DVOA Ratings

Baltimore with negative numbers across the board and still comes in number 3.

They're pretty much saying "Our offense isn't very good, our STs aren't very good, but our defense will so thoroughly ruin your attempts to mount an offense that you won't be able to do anything about it."

Something makes me think that model won't be sustainable. This isn't the 2000 defense.

101 Re: Week 4 DVOA Ratings

Their rank isn't different, but their overall DVOA is quite a bit better so far in 2011 than it was in 2000 (-1.1 vs -9.7).

I think that offense number will rise also (I have no basis to think this other than the fact I can't see Flacco's cmp% being bel How 50 for too much longer, and it might be skewed coming off of a lousy game). That defense number is extraordinarily low (as in good), and I believe it would be the 2nd best defense in the DVOA era if the season ended today. However, getting 8 fumbles on defense in just 4 games probably isn't a sustainable pace.

79 Re: Week 4 DVOA Ratings

Dallas is clearly ranked too high because Tony Romo is the most clutch QB in the history of the NFL. With two clutch victories against Dallas, a clutch win against the 49ers and a heroic performance against Washington with bad ribs, there's no way they're not better than the 17th ranked offense in the league. Romo just plain wins games!

84 Re: Week 4 DVOA Ratings

NE's defense is justifiable terrible, but it does give me hope when I see that they've only faced offenses with positive DVOAs.

(Of course, that could be because those teams had the luxury of facing NE, but I'll ignore that for the time being, TYVM.

89 Re: Week 4 DVOA Ratings

In reply to by Anonymous1 (not verified)

NE-Jets will be interesting this weekend. #29 offense vs. #25 (I think, doing from memory) defense on one side; #1 offense vs. #2 d on the other.

(Before anyone gets their back up, I don't mean to imply that the game is therefore a "toss up." NE obviously looks like the better team at this point, even presuming Mangold returns.)

90 Re: Week 4 DVOA Ratings

Considering the trouble NE has had covering tall receivers, I'm a little worried about Plax and Keller always scares me. I guess being able to take Gates out of the game is a good sign. I really don't want to see LDT carve them up all game though, his first down celebrations are so over the top.

93 Re: Week 4 DVOA Ratings

In reply to by Newjamarcus (not verified)

NE was the better team last January as well, but that didn't help them too much between the lines.

The Jets' season is in real trouble if they lose this game, so I expect a much better performance than last week.

85 Obligatory Bucs Comment

A few notes on my Buccaneers-

1. I'm not going to quibble with the overall rankings. First, because this is not the forum for it (how dare you not have a clutchness category!). Second, because I'm not certain that it's wrong.

2. Random slight surprise- that the computer ranks their offense so well (10th) and their defense so poorly (25th). I know computers don't lie, and their defense has had some rough spots, but I've been pleasantly surprised by their defense. In fact, in the last two games especially, I would go so far as to say they've been driven by their defense. I'm guessing it's because the Colts are bad.

3. Totally off-topic comment. I've always thought flags even out. I've never had a problem with officiating. I hate it when people whine that *their* team is getting screwed by the officials. But I've had this weird thing about the Bucs this year- for the first time, I've been thinking they have been getting a little screwed. I thought it was just me, but Mrs. Loki13 (who, god bless her, has been learning football this year) made the same comment, unprompted, during the Atlanta game. And spent the whole Monday expecting to get screwed. And was flabbergasted by MNF's calls. Well, that and Gruden's package. Anyway- I'm just saying. Don't know what to make of this. Either my eyes are deceiving me, I've become one of those fans, my wife is one of those fans and she's influenced how I'm watching the game, or the Bucs have morphed into Raiders East.

117 Re: Obligatory Bucs Comment

The flag difference in Monday night's game was justified for the most part. From the other side, I'm a Colts fan for what its worth, but all of the Bucs penalties were well deserved and pretty obvious. In fact the refs missed a rather obvious hold on Freeney on the game winning drive.

You could argue a few instances where the Colts should have been called for a penalty and weren't so I'll give you that it was slightly imbalanced, but the Bucs did their best to shoot themselves in the foot over and over again, but Curtis Painter and the Colts secondary outsucked them all.

87 Re: Week 4 DVOA Ratings

I just realzed whether you played an opponent at home or on the road it doesn't factor into schedule strenth. Why is that?

92 Re: Week 4 DVOA Ratings

DVOA does not yet have home/road schedule strength adjustments. I have heard no talk from FO about ever adding it.

94 Re: Week 4 DVOA Ratings

Cannot read all comments nkw but will say Raisers deserve 12 rankung. Wil imptove though. Pates were good test for them. Raiders wi ll get them back in playkffs just don't you worty about that. Juat like ravens did in 09 and Jets last season. Raiders personnel really coming togeyher and going to only het better seapsn voes on.

102 Re: Week 4 DVOA Ratings

If Campbell didn't throw that awful interception, it is probably some variation of 17-13, 17-17 or even 14-17 NE-OAK at halftime. They were in that game for a half. I take quite a few positives out of that game for the Raiders. Also, I've read in many places (not here) that Brady "carved up the Raiders" or something similar. No, Wes Welker did. Brady barely completed 50% of his throws, even though he got good protection. The Raiders really locked down Gronk, and Ocho continues to do nothing.

103 Re: Week 4 DVOA Ratings

The NE defense is clearly ranked too high because there is no possible way that five teams could actually be worse on defense than they are. Some system that doesn't overvalue turnovers is way better than this. J-E-T-S suck!

105 Re: Week 4 DVOA Ratings

I could say the same thing for Buffalo. They get a couple flukey picks (along with a couple solid ones) and suddenly they are 90% better per DVOA on the game?

Personally, I think Buffalo's defense is the worse of the two and DVOA will agree by the end of the year.

107 Re: Week 4 DVOA Ratings

I actually think there's some validity to your templated argument. Turnovers are tremendously valuable, and they're predictive as well. But they are also relatively low-occurrence events. It seems reasonable that their small sample size could lead to more substantial swings than other variables do, especially early in the season.

119 Re: Week 4 DVOA Ratings

Do you want DVOA to measure what's happened or predict the future? Those are two different tasks.

Just ask anybody who's ever been hit by lightning.

125 Re: Week 4 DVOA Ratings

I wonder what the Texans defensive DVOA would be if you threw out the fourth quarter against the Saints. Seems like they have been better than where they are valued for all but that one quarter. I suppose if that is true and trends continue, we will find out as more games are played.

126 Re: Week 4 DVOA Ratings

Hmm. They have had a probably unsustainable habit of allowing long drives and then ending them with takeaways, though . . .

I think the Texans defense is well equipped to annihilate sub-par offenses but not good enough to trouble really good ones. That will probably manifest as high variance but in a predictable way - almost as if they had a kind of strength of opponent multiplier.

Save 10%
& Support Aaron
Support Football Outsiders' independent media and . Use promo code SCHATZ to save 10% on any FO+ membership and give half the cost of your membership to tip Aaron.