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

Week 4 DVOA Ratings

Week 4 DVOA Ratings
Photo: USA Today Sports Images

by Aaron Schatz

As we've noted many times in the last few years, a recent wave of "surprise" Super Bowl champions has obscured the fact that parity during the regular season has actually been getting weaker. And boy, are we feeling that this season. Opponent adjustments are applied to our DVOA ratings this week for the first time in 2013, but that doesn't do much to help the Jacksonville Jaguars, who currently register as the worst team Football Outsiders has ever tracked through four games.

THROUGH 4 GAMES, 1989-2013*
THROUGH 4 GAMES, 1989-2013*
Year Team DVOA x Year Team DVOA
2013 JAC -83.8% x 2013 JAC -67.1%
2001 WAS -71.6% x 2002 HOU -61.5%
2008 DET -69.8% x 2004 MIA -60.4%
2008 STL -65.1% x 2001 WAS -56.5%
1989 DAL -64.2% x 2009 OAK -45.1%
2002 CIN -60.3% x 2007 CHI -44.8%
2000 CIN -59.6% x 1999 PHI -41.8%
1999 CLE -57.4% x 2005 MIN -40.2%
1999 CIN -57.0% x 2002 CIN -40.2%
2005 HOU -55.1% x 1999 ARI -39.7%
2006 TEN -54.0% x 1996 STL -39.1%
2006 HOU -54.0% x 2006 OAK -39.0%
*For teams with Week 5 bye, we listed ratings with adjustments through Week 4 rather than Week 5.

How bad is it for the Jaguars' offense? In pass offense, the gap between Jacksonville (-74.2% DVOA) and the No. 31 New York Jets (-19.9% DVOA) is larger than the gap between the Jets and the Seattle Seahawks, who rank fifth in the league in passing offense (31.7%). That's just mind-boggling. In our latest playoff odds simulation, the Jaguars end up with the No. 1 overall pick more than two-thirds of the time.

And on the other side of things, we have the Denver Broncos. Remember when we wrote in Football Outsiders Almanac 2013 that the Broncos' defense would probably regress heavily towards the mean this year? That prediction has turned out absolutely right, although it certainly has been helped along by the injury to Champ Bailey and the suspension of Von Miller. The Broncos are currently just 19th in defense. Does it matter? Not so far, because the Broncos lead the league in special teams and have the second-best offense we've ever tracked through four weeks.

THROUGH 4 GAMES, 1989-2013
THROUGH 4 GAMES, 1989-2013
Year Team DVOA x Year Team DVOA
2007 NE 73.4% x 1999 WAS 54.9%
1991 WAS 71.8% x 2013 DEN 53.2%
2013 DEN 62.2% x 1998 DEN 52.4%
1999 STL 62.1% x 2005 SD 48.9%
1996 GB 60.8% x 2007 NE 46.6%
2009 PHI 57.1% x 2002 KC 46.1%
1998 DEN 54.3% x 2007 DAL 43.5%
1992 PHI 54.2% x 2005 SEA 42.1%
2009 NO 53.6% x 1997 DEN 41.0%
1992 BUF 51.5% x 2001 STL 40.9%
2007 DAL 51.1% x 2010 NE 40.8%
2001 STL 50.8% x 2005 PIT 38.6%

To answer your question: Yes, that really is Washington in 1999 listed with the best offensive DVOA through four games, led by not-even-close-to-Hall of Famer Brad Johnson. We don't think of this as one of the great offenses in history because a) they slowed down and b) the defense wasn't good --  it finished the year 24th in defensive DVOA -- so it wasn't like the team was winning big every week. Washington lost its first game in overtime, then won the next three. In those four games, they averaged 32.5 points per game, although they allowed 28. That was the season Stephen Davis went crazy, with 1,405 yards and 17 touchdowns (nine of which he got in the first four weeks). For the full season, that team also got over 1,100 yards with 18.3 yards per reception from both Michael Westbrook and Albert Connell.

Back to Peyton Manning and this year's Broncos... I got an e-mail in the mailbag yesterday about Denver's ridiculous pace so far this year.

Bill Herman: Peyton Manning and the Broncos are headed for breaking a fistful of major all-time-records, including (but not limited to): most points in a season, biggest point differential, most passing yards, most passing touchdowns, and best passer rating. I'm wondering if you and/or others at FO have opinions on which of these (or other major records) are mostly likely to be broken by the end of the season.

So I e-mailed Bill back and pointed out the problem with reporting that "Player X is on pace for..." There usually were other players in the past who also were on similarly ridiculous paces early on in the season. Those players mostly trailed off a bit, as you might expect. It's hard to keep throwing four touchdowns a game. So the best way to determine if a team/player is likely to break a record is not to look at their current pace compared to the record, but to look at their current pace compared to the best teams/players in the past at a similar point in the season. That we do above with the "best/worst teams in DVOA" tables. These tables only look at the first four weeks of the season, with opponent adjustments based only on the data we had at the time (and only at 40 percent strength, same as now).

So, let's look at where the Broncos sit compared to other teams through four weeks.

Most points for an NFL team, season: 589
1966 Cowboys: 183
Broncos, so far: 179 (on pace for 716)
2002 Raiders: 160
1968 Cowboys: 159

Biggest point differential, season: +315
Broncos, +88 (on pace for +352)

Amazingly, the Broncos don't even have one of the top dozen "through four games" point differentials since 1960. The 1966 Cowboys were +138. The 1968 Cowboys were +116. Other recent higher (or equal) numbers:
1992 Bills: +108
2007 Patriots: +100
1999 Rams: +95
1997 Patriots: +90
1991 Redskins: +88

Most passing TDs, season: 50
Manning, so far: 16 (on pace for 64)
Kurt Warner, 1999: 14
Tom Brady, 2007: 13
Daunte Culpepper, 2004: 13

Most passing yards, season: 5,476
Kurt Warner, 1999: 1,557
Tom Brady, 2007: 1,553
Manning, so far: 1,470 (on pace for 5,880)
Kyle Orton, 2010: 1,419

Yes, Kyle Orton really is fourth all-time in passing yardage through four games, in a season where the Broncos went 3-10 with him as the starter. He had 476 yards on 57 passes in a ridiculous Week 3 loss to Indianapolis that year.

Best passer rating, season: 122.5
Manning, so far: 138.0
Kurt Warner, 1999: 136.0
Tom Brady, 2007: 134.7

Looking at these numbers, I think it's clear that the record most in danger is Tom Brady's record for passing touchdowns in a season. The other records are certainly possibilities, but passing touchdowns is the most likely. Point differential is the least likely, in part because you really can't expect the Broncos to be scoring this many points off special teams for the entire season.

As long as we're looking at teams that rank among the best and worst through four weeks, let's look at one more table. We all know the Giants have no running game this year, their offensive line is a mess, and the defense isn't playing well. You may not know that the special teams have been even worse than the offense or defense. The biggest issue is giving up two punt returns for touchdowns, but even if we remove those two returns, the Giants' average of 9.5 yards allowed on ten other returns would rank 23rd in the NFL. In addition, Josh Brown has missed two makeable field goals (38 and 44 yards), and Giants punt returners have muffed two punts. As a result, the Giants end up with the second-worst special teams we've ever tracked through four weeks.

THROUGH 4 GAMES, 1989-2013
Year Team DVOA
2010 SD -25.0%
2013 NYG -17.3%
1997 STL -17.0%
2011 CAR -17.0%
2005 PHI -16.9%
1999 CIN -16.8%
2000 DEN -16.6%
2012 WAS -16.5%
1990 DEN -16.4%
2007 CIN -15.3%
1990 SF -15.2%
2007 DEN -15.1%

Applying opponent adjustments this week means we can get our first good look at schedule strength. As you might expect this early in the season, the best teams come out with the easiest schedules and the worst teams with the hardest schedules. It is hard to tell at this point whether Kansas City is really this good, and that's why opponents like the Eagles and Giants have struggled, or whether the Eagles and Giants are awful and so Kansas City looks better than it truly is. (This is kind of kooky, but even though Denver has blown away the rest of the NFL this year, their playoff odds actually dropped by 0.2 percent this week because the Chiefs went to 4-0 and rank third overall.) However, it is interesting to note a pack of teams near the bottom of the DVOA rankings who have played easier than normal schedules so far. Washington, Oakland, and St. Louis rank between 28th and 30th in DVOA with schedules ranking between 21st and 26th.

We also can start to get a good idea of future schedules at this point. The Patriots may have started 4-0, but things are going to get a lot harder. Not only did they just lose their best defensive player -- you can argue for Aqib Talib, but I'll still say Vince Wilfork -- but they have the league's hardest remaining schedule. The issue is less that they have a lot of hard games left, and more that they have no "gimmes." The worst teams still on New England's schedule are the AFC North teams other than Cincinnati, which currently rank 23rd, 24th, and 25th. The Chiefs may look good now, but they've had the league's easiest schedule so far and rank third going forward. Meanwhile, if you are looking to figure out who is going to win the NFC East, be aware the Eagles come out with the easiest remaining schedule out of all 32 teams, although Dallas is only 24th. There's a pretty good chance the NFC ends up with two wild cards with double-digit wins, because Seattle, San Francisco, Green Bay, Chicago, and Detroit all rank among the seven easiest remaining schedules.

When comparing DVOA and VOA, also remember that the difference isn't entirely based on schedule strength; that difference also accounts for fumble recovery luck and the difference between special teams with (DVOA) and without (VOA) weather and altitude adjustments. Tennessee so far this year has recovered three of their own four fumbles on offense, and all four opponent fumbles on defense. On the other hand, the Steelers have lost five of their seven fumbles, and haven't recovered any of the three fumbles they've forced on defense.

* * * * *

Each week during the 2013 season, we'll be partnering with EA Sports to bring special Football Outsiders-branded items to Madden 25 Ultimate Team. Each week, we'll be picking out a handful of players who starred in that week's games. Some of them will be well-known players who stood out in DVOA and DYAR. Others will be under-the-radar players who only stood out with advanced stats. We'll announce the players each Tuesday in the DVOA commentary article, and the players will be available in Madden Ultimate Team packs the following weekend.

The Football Outsiders stars for Week 4 are:

  • QB Drew Brees, NO: Led all quarterbacks in DYAR in Week 4.
  • RB Danny Woodhead. SD: Second among running backs in DYAR; 7-for-7 on passes with two touchdowns.
  • LT Nate Solder, NE: Anchored a line that blocked for 5.2 yards per carry with no sacks.
  • ILB NaVorro Bowman, SF: Five Defeats, including two sacks, two TFL on runs, and a PD to prevent a third-down conversion.
  • CB Aqib Talib, NE: Only allowed one pass on eight targets for just one yard; also had interception and game-ending pass defensed.

For those who are curious about some other players, we try not to add players who were updated in recent weeks. For example, we considered both Alterraun Verner and Zack Brown of Tennessee, but they were both part of the Week 1 "Team of the Week." We also considered Seattle's Chris Clemons, but he was recently added in a player update, so the guys from Madden Ultimate Team want to wait a few weeks before they put a newer, boosted version of Clemons out there.

Also, the Talib stats listed here are based on a graphic from the NBC broadcast; we're doing our own charting to determine our take on how many passes Talib was targeted on/allowed completions on.

* * * * *

All 2013 stat pages will be updated in the next few minutes, including snap counts, playoff odds, and the FO Premium database.

* * * * *

[ad placeholder 3]

These are the Football Outsiders team efficiency ratings through four weeks of 2013, 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 for strength of schedule and 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.

Because it is early in the season, opponent adjustments are only at 40 percent strength; they will increase 10 percent every week through 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 45 percent of DAVE (55 percent for Carolina and Green Bay). The projection portion of DAVE has been adjusted for some teams with major injuries or personnel changes.

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 DEN 62.2% 2 43.4% 1 4-0 53.2% 1 3.0% 19 12.0% 1
2 SEA 43.1% 1 33.4% 2 4-0 6.6% 9 -28.0% 1 8.5% 2
3 KC 32.7% 3 10.6% 8 4-0 5.3% 11 -18.9% 3 8.5% 3
4 NO 31.7% 5 21.1% 3 4-0 15.4% 7 -15.5% 6 0.8% 15
5 IND 25.4% 4 11.0% 7 3-1 19.1% 4 -5.5% 13 0.8% 13
6 CAR 21.8% 7 18.6% 4 1-2 13.7% 8 -9.4% 10 -1.2% 20
7 NE 13.4% 11 18.6% 5 4-0 2.6% 15 -5.0% 14 5.8% 4
8 TEN 11.5% 17 -1.3% 17 3-1 4.4% 14 -9.4% 9 -2.3% 23
9 GB 11.2% 9 18.2% 6 1-2 30.7% 3 21.1% 31 1.6% 12
10 BUF 9.4% 19 -1.3% 18 2-2 -4.7% 21 -13.3% 7 0.8% 14
11 CHI 7.7% 6 4.0% 12 3-1 -0.6% 19 -8.9% 11 -0.6% 18
12 SD 6.6% 18 0.5% 16 2-2 33.1% 2 25.5% 32 -1.0% 19
13 DET 4.8% 16 1.7% 15 3-1 4.7% 12 -3.1% 15 -2.9% 25
14 ATL 4.4% 13 4.5% 10 1-3 17.7% 6 14.9% 26 1.6% 11
15 CIN 3.2% 10 4.4% 11 2-2 -1.5% 20 -1.7% 16 3.0% 10
16 DAL 2.3% 14 -1.6% 19 2-2 6.0% 10 4.3% 20 0.6% 16
17 HOU 1.6% 20 3.8% 13 2-2 -4.8% 22 -16.1% 5 -9.7% 30
18 SF -1.2% 27 6.3% 9 2-2 0.6% 17 -1.2% 18 -3.0% 26
19 MIA -1.4% 8 -4.6% 21 3-1 4.5% 13 10.5% 24 4.6% 7
20 NYJ -1.5% 12 -3.8% 20 2-2 -24.3% 29 -17.9% 4 4.9% 6
21 MIN -3.4% 24 -8.3% 26 1-3 -0.6% 18 6.7% 23 3.9% 8
22 TB -6.4% 21 -5.9% 23 0-4 -24.6% 30 -19.6% 2 -1.4% 21
23 BAL -6.8% 22 2.1% 14 2-2 -18.0% 26 -7.6% 12 3.6% 9
24 CLE -12.1% 26 -11.0% 27 2-2 -18.3% 27 -10.3% 8 -4.1% 27
25 PIT -14.6% 23 -7.9% 24 0-4 -5.9% 23 6.2% 21 -2.5% 24
26 PHI -14.9% 15 -8.2% 25 1-3 18.7% 5 20.8% 30 -12.9% 31
27 ARI -17.4% 28 -14.4% 28 2-2 -17.6% 25 -1.6% 17 -1.4% 22
28 WAS -22.4% 30 -5.1% 22 1-3 0.8% 16 14.3% 25 -9.0% 29
29 OAK -31.8% 29 -24.8% 29 1-3 -9.1% 24 17.1% 28 -5.7% 28
30 STL -35.5% 25 -27.1% 31 1-3 -20.9% 28 19.7% 29 5.0% 5
31 NYG -48.7% 31 -25.5% 30 0-4 -24.7% 31 6.6% 22 -17.3% 32
32 JAC -83.8% 32 -55.8% 32 0-4 -67.1% 32 16.3% 27 -0.5% 17
  • 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 DEN 62.2% 4-0 73.3% 4.0 1 -25.6% 30 -0.4% 19 1.3% 5
2 SEA 43.1% 4-0 47.6% 3.4 3 -15.4% 29 -7.7% 31 18.7% 25
3 KC 32.7% 4-0 45.2% 3.5 2 -36.3% 32 7.3% 3 2.9% 7
4 NO 31.7% 4-0 37.0% 3.1 4 -5.2% 24 6.6% 5 7.4% 18
5 IND 25.4% 3-1 38.5% 3.0 5 -29.6% 31 3.1% 10 6.3% 13
6 CAR 21.8% 1-2 25.9% 2.2 15 1.3% 14 0.9% 15 23.1% 28
7 NE 13.4% 4-0 11.8% 2.9 6 1.5% 13 7.7% 1 0.9% 3
8 TEN 11.5% 3-1 24.0% 2.2 13 -2.0% 18 -5.3% 25 14.7% 22
9 GB 11.2% 1-2 16.4% 2.2 17 -6.8% 25 -5.6% 26 6.6% 14
10 BUF 9.4% 2-2 12.2% 2.3 10 6.7% 9 -3.0% 22 16.8% 23
11 CHI 7.7% 3-1 8.9% 2.6 7 -2.5% 20 -6.9% 29 4.5% 11
12 SD 6.6% 2-2 0.3% 2.4 9 0.1% 15 -0.1% 17 1.6% 6
13 DET 4.8% 3-1 9.9% 2.3 12 -8.9% 27 -6.0% 28 3.6% 9
14 ATL 4.4% 1-3 10.3% 2.4 8 2.0% 12 7.0% 4 0.8% 2
15 CIN 3.2% 2-2 5.0% 2.2 14 -2.0% 17 1.1% 13 4.5% 10
16 DAL 2.3% 2-2 12.2% 1.8 22 -11.2% 28 -3.4% 24 3.0% 8
17 HOU 1.6% 2-2 -0.3% 2.2 16 13.6% 5 -6.9% 30 6.8% 16
18 SF -1.2% 2-2 -4.7% 1.8 21 11.0% 7 -5.7% 27 35.4% 32
19 MIA -1.4% 3-1 -5.1% 2.3 11 12.3% 6 3.9% 9 24.9% 30
20 NYJ -1.5% 2-2 -7.6% 2.1 18 7.0% 8 1.3% 11 22.3% 26
21 MIN -3.4% 1-3 -0.7% 2.0 20 -3.6% 22 1.0% 14 8.1% 19
22 TB -6.4% 0-4 -8.9% 1.4 25 6.5% 10 7.4% 2 0.6% 1
23 BAL -6.8% 2-2 -12.0% 2.0 19 15.3% 4 -0.4% 18 24.8% 29
24 CLE -12.1% 2-2 -0.5% 1.3 26 -2.1% 19 -3.3% 23 13.0% 21
25 PIT -14.6% 0-4 -22.7% 0.9 28 4.7% 11 -2.5% 20 4.8% 12
26 PHI -14.9% 1-3 -21.8% 1.8 23 19.8% 2 -12.5% 32 26.2% 31
27 ARI -17.4% 2-2 -14.1% 1.7 24 -1.3% 16 1.2% 12 6.7% 15
28 WAS -22.4% 1-3 -18.6% 1.3 27 -7.7% 26 0.1% 16 7.1% 17
29 OAK -31.8% 1-3 -23.9% 0.4 30 -4.7% 23 6.4% 7 1.2% 4
30 STL -35.5% 1-3 -36.8% 0.5 29 -3.0% 21 6.4% 6 22.4% 27
31 NYG -48.7% 0-4 -66.3% 0.0 31 29.8% 1 -2.9% 21 17.3% 24
32 JAC -83.8% 0-4 -87.6% 0.0 31 17.3% 3 5.3% 8 12.2% 20


137 comments, Last at 06 Nov 2014, 3:22pm

1 Re: Week 4 DVOA Ratings

You should add a new item to the Playoff Odds Report for this season: Chance the NFCE has a sub-.500 champion.

2 Re: Week 4 DVOA Ratings

How much of the Chiefs' tough remaining schedule is having to play Denver twice and how much is the rest of the schedule?

4 Re: Week 4 DVOA Ratings

San Diego has a positive DVOA as well with a terrific offense so far. Other than that, the AFC South looks like a decent division minus the Jags who they've already played.

9 Re: Week 4 DVOA Ratings

Yeah the Chiefs so far have beaten up on three NFC East patsies and the woeful Jaguars. The Cowboys are probably the best of that lot and they hover around mediocre (I called an 8-8 finish for them at the beginning of this year happy they could finally fire Jason Garrett for missing the playoffs. Of course little did I know that 8-8 looks like a good bet to win the division).

3 Re: Week 4 DVOA Ratings

Small correction: I think you meant "AFC North teams other than Cincinnati," not AFC East.

5 Re: Week 4 DVOA Ratings

Ah, Tampa, 0-4, ahead of 2-2 teams. Why? Because DVOA can't account for "stupid at the wrong time".

7 Re: Week 4 DVOA Ratings

Tampa's offensive futility is at least somewhat offset by their very good defense. The problem is their defense isn't quite at the old Bears "able to win games on its own" level yet.

21 Re: Week 4 DVOA Ratings

"The defense isn't quite at the old Bears "able to win games on its own" level yet."

The 1997-2001 Buccaneers won a lot of games without a competent offense, but there's a big difference between having Tony Dungy as your coach vs Greg Schiano.

6 Re: Week 4 DVOA Ratings

I threw out some Jacksonville stats on Twitter earlier, while I was reading some articles on them. This is how bad Jacksonville is:

- They've scored 31 points this year. Only the 2009 Rams, who finished 1-15, have scored fewer in their opening four games since 2002.
- 21 out of 32 teams, or roughly 2/3 of the league, have scored as many or more points in ONE GAME as Jacksonville has in four. Twice already, BOTH teams have scored 31+ (Chicago @ Detroit and NY Giants @ Dallas).
- Denver has scored more points in every single game they've played (49, 41, 37, and 52) than Jacksonville has all season.

12 Re: Week 4 DVOA Ratings

Denver's scored 31 in three different halves, if I remember correctly.

At some point I or someone needs to go through and figure out just how little yardage Jacksonville has amassed when they're within three touchdowns - I'd guesstimate that almost half of their yards and points have been scored when behind by more than 20 (I think almost every score against Seattle was of the down 30+ category).

Or I could just poke my eyes out before Sunday. Tempting.

47 Re: Week 4 DVOA Ratings

As much as we like to make fun of JAC, if DEN played SEA, KC and INDY 3 of 1st 4 games I don't think they would have scored 31 in a game, let alone a half.

8 Re: Week 4 DVOA Ratings

Do we think AFC West's strength of schedule is actually tougher seeing, as there are 3 teams in the top 12 and OAKs offensive isn't as bad as we thought it could be due to QB play?

10 Re: Week 4 DVOA Ratings

All stats pages should now be updated, except drive and pace stats, which I assume Jim Armstrong will get to sometime tonight. Thanks for waiting while I cooked family dinner.

11 Re: Week 4 DVOA Ratings

Good for you. We can always wait for a home cooked meal.

Extra points for comparing apples to apples in the four week DVOA tables. And I think your guess that the TD record is the most at risk is spot on. A few years ago I'd have said it would take a miracle or major rule changes. It may have only needed Welker and Manning to move to a team that already had a very good line and quality receivers.

16 Re: Week 4 DVOA Ratings

I think total points is about as easily topped as TD passes. Both are heavily tied to whether Manning stays in these games for 4 quarters (what doomed him in '04 from not having ~55 TDs - he only threw 492 passes that year).

It would help their cause if they keep getting Special Teams and Defensive TDs too.

31 Re: Week 4 DVOA Ratings

My money is on TD passes and Passer Rating. The only thing Manning truly controls is his efficiency. If they get up by 4 TDs every time, he'll be pulled in the 3rd qtr. a bunch. I think he sat out 8-10 quarters in 2004, as dmstorm22 alluded to above. So he can have a crazy high rating with many fewer yards than Brees. Similar thing for TDs, but that's less under his control, if D and ST score two in a game and the O scores 2, Peyton might find himself holding the clipboard for 30 minutes of game time....

Of course to get up by 4 TDs a game, that means he's probably producing 3 of them each game. And for a few times that the D/ST lets up a bit and the opponents have 14-21 at halftime, Manning will be unleashed for the whole second half by Fox like The Lord Humongous unleashing Wez in "Road Warrior."

Wow, did I just say that?

44 Re: Week 4 DVOA Ratings

Looking back to 2004 Manning had 46TDs after 13 games, then they played Baltimore who limited him to 1TD and the following week he threw 2TDs against San Diego in an overtime game. Clearly in neither of those games did he sit out. But then Jim Sorgi played the final game leaving Manning with *only* 49TDs. That's very illustrative of how record pace sometimes slows down. (I'm sure there were games early in the season where he also sat out the 4th quarter).

30 Re: Week 4 DVOA Ratings

Damn kids, always eating, needing to eat, be cleaned up after... and then they have the nerve to complain about dinner! Oh wait, wasn't this a football thread? My mistake.

13 Re: Week 4 DVOA Ratings

Jacksonville is clearly ranked too high because Blaine Gabbert will never be the 32nd best QB in the league. The BCS is way better than this. David Garrard is on the path to sainthood after accomplishing the miracle of dragging this team to a playoff win.

14 Re: Week 4 DVOA Ratings

Any hope that Blackmon coming back will turn things around?
Yeah, didn't think so....

20 Re: Week 4 DVOA Ratings

Unfortunately, Blackmon neither plays quarterback nor offensive tackle.

I do, however, think that having him and a healthy Lewis might help. Not to the level of turning things around, but help.

Oh, and they apparently just agreed to trade Eugene Monroe to Baltimore. That definitely won't help.

15 Re: Week 4 DVOA Ratings

So the bottom 3 Defenses all belong to top 5 Offensive teams? Is that unusual?

18 Re: Week 4 DVOA Ratings

Well, passing is usually more efficient than running (i.e. pass well to get ahead, run well to stay ahead). So teams with excellent offenses tend to force their opponents to pass in order to keep up. But there are just a ton of complicating factors, such as:

1) When does the team consider the game in hand? When is garbage time? Has the defense switched early to a prevent formation that will make the opponent DVOA look better?
2) Does your opponent have a competent passing game, such that forcing them to pass more is actually a benefit?
3) Is your defense weak enough in the secondary (and with the pass rush) such that your opponent is better off passing instead of running?

I've watched all of San Diego's games this year and their secondary and pass rush is beyond awful. I'm not surprised at all by their last place ranking. However, if San Diego wasn't scoring points than opponents would be running more on them, so the ranking wouldn't be as bad.

Conclusion: not unexpected, depending upon schedule and defensive strengths.

51 Re: Week 4 DVOA Ratings

Ah! I should have thought of it that way, since when I play Madden, I play backwards so I WON'T have the League worst Defense. I would get ahead by so much that the other team was throwing all day long, and in the end put up lots of yards.

Now when I play against the console, I defer the coin toss to keep points lower, and run as much as I can. If I can keep it to 17-0 at the half, the AI is still running the ball and trying short passes.

22 Re: Week 4 DVOA Ratings

It seems to be an artifact of the fact that teams up by a lot tend to play a lot of prevent defense, which gives up a lot of yards/points. Does FO Compensate for this?

23 Re: Week 4 DVOA Ratings

All plays account for scoring margin. If the Broncos is on the field up by 40 points, then they are compared to how other defenses fared when they were ahead by 40 points.

25 Re: Week 4 DVOA Ratings

Hmm. Then it could be a philosophical difference between elite defensive teams and elite offensive teams. The elite defenses pride themselves on not giving up points, and will probably continue to give it their all throughout the game, whereas defenses of elite offensive teams are probably more likely to coast and play prevent, knowing that their offense can keep pace.

26 Re: Week 4 DVOA Ratings

Does DVOA account for the fact that teams who are ahead by 40 points tend to be above-average teams? The average level of performance by a team that is good enough to be ahead by 40 points, while they are ahead by 40 points, is better than the average level of performance that a random team would have if they somehow got ahead by 40 points. (Because you're averaging together a bunch of plays by the Broncos, Seahawks, Saints, etc., with very few plays by the Jaguars, Raiders, etc.)

27 Re: Week 4 DVOA Ratings

We've had this conversation before.

Until you guys are willing to tell us what the actual buckets are for what teams are compared against, saying "[we] account for scoring margin is meaningless".

What exactly is a 40 point lead bucketed with? Only other 40 point leads? Leads of greater than 20? Greater than 17? I'd argue that those are all very different situations, and we have no way to tell if you're lumping them together.

I seem to remember Aaron saying at one point that blowouts were leads of more than 2 scores in the 4th quarter, which is very different than leads of 40.

91 Re: Week 4 DVOA Ratings

So teams that are willing to run up the score will receive a huge bonus compared to those that aren't? It seems to me that performance in massive blowouts has as much to do with tactics as team ability.

19 Re: Week 4 DVOA Ratings

As bad as the Vikings have been (and it is sobering to Ponder that their production at cb is about the same as qb), it is crazy to think that they are two long td drives, yielded in the closing moments, from being 3-1, and they were tied in the other loss fairly late. If it gets them to give up on Mr. Steele, however, these losses will not have been in vain. Never was so little hoped by so many from so few.

24 Re: Week 4 DVOA Ratings

I haven't seen any Jacksonville or Giants games this season, but I have seen each Steelers game. If there's anything I thought the Steelers would be good at this season it would be being terrible. How can Jacksonville be beating them so thoroughly at being terrible? By which I guess I mean Pittsburgh is clearly ranked too low because they have been awesome at being terrible. Start to Crate is way better than this. The Nigerien Sooper Eagles are jolley good.

32 Re: Week 4 DVOA Ratings

The Steelers have plenty of flaws and weaknesses right now, but they've still got Ben Roethlisberger at QB. He's leagues ahead of anything that Jaguars can use.

I'm still wondering how bad the Jax defense has to become before punting on 3rd down becomes the optimal strategy. Gotta figure the sneaky 3rd down punt can often go for an extra 20 yards. And all the Jaguars would be losing would be one offensive play. Subtracting a negative might be the right way to go.

33 Re: Week 4 DVOA Ratings

The Colts have me pleasantly confused. This #4 offense I am familiar with but don't quite understand--Luck has been efficient but not a world-beater. They're down 4 starters (C, OG, TE, RB), and although the RBs (and Luck) seem to compile a lot of yardage, the traditional YPC numbers are fairly pedestrian. Maybe it's just that they are balanced and both run and pass games are "pretty good" but not great. Maybe that's worth a 4.

The D has been surprisingly stout (shades of 2005?) and what's with this typo for ST? I am used to seeing 31 there, not 13. Surely that's wrong! I know McAfee had a great game punting against the Niners, but the returns have been modest, no 60 YD FGs... I guess playing competent football has its rewards. Just happy and slightly confused. I don't think they quite pass the eye test to be this highly ranked, but I guess time will tell. And time starts this Sunday.

48 Re: Week 4 DVOA Ratings

It is quite strange and not quite what I was used to for this decade. As much as we like to make fun of the "power rushing" concept, they actually do seem to be going toward a more balanced offense that has been pretty productive. Fleener has replaced Allen's production and they still have RB depth after the trade. McAfee has been booming kickoffs as well and the special teams has not been allowing the other team to get to the 30 every time as in the past.

A note on injuries as well. Being down the starting C is a boon for the offense. Satele is awful and I wish they'd cut him. The offense was better last year when he was hurt and Shipley played. The offense is better this year with him not in. Also Delano Howell seems to be a beast. The defense has improved a lot since he took over for Landry.

76 Re: Week 4 DVOA Ratings

Agree 100% with all of that. Honestly I have a problem with some of our player evaluation given that injuries to Satele and Landry have seemed to be a good thing so far. I am a big Satele hater (him and Donald Brown are my least favorite players on my favorite team) so that's not a surprise, but Howell has seemed awesome out of nowhere so far.

I really don't know how the rush offense is 2nd in the league though. It's obviously the best rushing offense the Colts have had in quite a while, but there still seems to be an awful lot of running into a pile for 2 or 3 yards.

78 Re: Week 4 DVOA Ratings

Looks like it is mostly the line, ALY is high, the 2nd level and open field ranks are lower. I think they've finally got a line that can do one thing pretty good. Last year they had a few guys that could run block well (Castonzo, McGlynn) and then some guys that could pass block well, and then Satele who just seems bad. Cherilus seems good at run blocking and so-so passing, Thomas seems good in the run game, and Linkenbach looks really good when he's a puller. So they now have a line that's kinda leaky against the pass in general but pretty solid running the ball I guess instead of a mishmash. I guess that makes sense, but honestly having watched all the Colts games I still wouldn't have expected 2nd in the league.

36 Re: Week 4 DVOA Ratings

Perhaps there should be a DVOA cap in the schedule calculations?

I see the Bills sitting with the 3rd hardest past schedule, which is arguable, but I'll accept that (even though I still think CAR is way overrated by beating the Giants) and the 22nd hardest future schedule, which I found, well, bizarre.

The Bills still have to play NO/KC/NE/NYJ/ATL/CIN/MIAx2 - that's a minimum of 8 tough games remaining (the worst, DVOA-wise, being the NYJ at a decidedly average -1.5%), how could they possibly have the 22nd hardest schedule left? Then I remembered - they also have to play JAC. Their -83.3% can drag the schedule estimate from that entire group of competitive-to-good teams to approximately zero, but if the Bills win, it still only counts as one win at the end of the day and really has no bearing on how the Bills will fare against their remaining schedule.

I'm thinking two better ways to estimate SoS:

1) Clip any single opponent to a -20% to 20% (we can quibble about exact numbers) range for SoS purposes. This should keep any extremely bad (or good) team from unduly influencing SoS for W/L purposes. The same logic should apply for teams that have lots of NFCE left + DEN left on their schedule.

2) Perhaps a separate SoS estimate based on extrapolated W/L for the season, maybe? This would probably be similar to the playoff odds report, but I'm not quite sure what goes into that.

40 Re: Week 4 DVOA Ratings

The playoff odds report is how a team with their DAVE rating performs on average (simulated) with their schedule, so what you'd want instead is how a team with 0% DAVE performs. That seems preferable to your first option, because that can't distinguish between, say, Jacksonville and Oakland, while the second method would see something like .82 expected wins against Jacksonville versus .67 against Oakland.

By the way, the .83 was derived from 1 - (2.2 expected wins/12 remaining games), but... 2.2? That seems awfully high.

101 Re: Week 4 DVOA Ratings

0% DAVE team--that makes a lot of sense.
Simulate a 0% team against each team's respective future sched., and rank schedule strength based on those expected wins.

39 Re: Week 4 DVOA Ratings

Looking at seattle's past schedule doesn't it seem unfair to allow the pathetic Jacksonville team pull them down so much (could apply to other teams Jax has played as well)? The jaguars suck but even if they are the worst team ever you still only get credit for one victory and one games worth of play. There are plenty of teams above Seattle that I would've traded the first four games with since the majority of the games would've been easier.

54 Re: Week 4 DVOA Ratings

Jacksonville's not pulling them down as much as you thing. Remember that opponent adjustments are nowhere near full strength yet. More than likely, playing them was a net positive, because they got to generate a large positive VOA by beating up on them. I don't think it's a coincidence that 3 of the top 5 teams have played Jacksonville (I'm not saying they're not good teams, just that their DVOA probabably got a big boost)

97 Re: Week 4 DVOA Ratings

If you meant to say "as much as you think", I'd agree with you. If you meant to say "as much as your thing", I'd have to ask you exactly how far you think my thing is pulling them down?

42 Re: Week 4 DVOA Ratings

To use a statistical term: NFL dvoa week to week is a nonstationary series. Ultimately, that implies that future results are likely not going to give us a better indicator of the true value of these teams. Which ultimately makes sense, considering each game has its own meaning and circumstance.

Take for instance, last years matches between STL and SF. The first one: Both teams had a positive offensive dvoa and a positive(bad) defensive dvoa. In their second meeting, the results were flipped, despite both games going into overtime.

49 Re: Week 4 DVOA Ratings

I just looked at JAC's remaining schedule. There easiest game is this week against the Rams. All the rest are against teams with above average defenses.

50 Re: Week 4 DVOA Ratings

Something lost in Rivers' season is the fact that his #1 and #2 WRs are both out for the season and the last two games have had at least 3 starting Offensive Linemen injured during the game (last week one of the second stringers got injured as well

52 Re: Week 4 DVOA Ratings

Anyone else find it interesting that 1999 Washington is in the top 10 list for best offense after 4 games and only 2 years later they made it into the top 10 for worst?

55 Re: Week 4 DVOA Ratings

I suspect a lot of that has to do with strength of schedule. At week four, the opponent adjustments are only weighted at 40%. So if you get an easy schedule to start the season, your DVOA gets an unearned boost. You "look" better to DVOA than your "actual quality" - assuming as always that there is such a thing as "actual quality" for DVOA to measure.

Aaron had to use this form of DVOA rather than using the final, fully weighted opponent adjustments, because the point was to include this year's Denver Broncos in the lists. We don't know what the final, fully weighted adjustments will be for Denver's first four games. Aaron could have projected those adjustments, but it would have been speculative.

Given the problems with week four DVOA opponent adjustments, it's better to see this as Washington going from a top ten "apparently" fastest starting offenses, to a bottom ten "apparently" slowest starting one.

136 Re: Week 4 DVOA Ratings

I'm pretty sure that the "Through Week X" numbers are retroactively updated to include full opponent adjustments after the season is over. That is to say, while this year's teams only have 40% adjustment, if we were to look at this same table next year the adjustment would be at the full 100%.

70 Re: Week 4 DVOA Ratings

Much appreciated, but I have to disagree.

Radioactive Dan Marino is both shorter, and more versatile in conversation. Any time a QB seems to suddenly, inexplicably improve his play (Brian Hoyer & Jake Locker this season; Ryan Fitzgerald and Derek Anderson in years past), he has been "bitten by a radioactive Dan Marino". I think it belongs in the FO lexicon alongside ROBO-PUNTER.

It's like the way a bad personal matchup is someone's "Kryptonite" (Ronde Barber to Donovan McNabb, Bill Belichick to Peyton Manning, etc.).

71 Re: Week 4 DVOA Ratings

Hmmmmm....."Paranormal Albert Haynesworth", for any time an owner gets his team in cap hell with guys who like eating Doritos more than they like playing football?

72 Re: Week 4 DVOA Ratings

In reply to by Independent George

I think you've unintentionally coined another term. "Ryan Fitzgerald" -- a player who hangs around for a decade, yet remains so forgettable that it's not even worth learning his name.

82 Re: Week 4 DVOA Ratings

In reply to by Revenge of the NURBS (not verified)

Ha! I would be embarrassed, but I enjoy that too much. How about we go one better and call it a Bryan Fitzgerald?

94 Re: Week 4 DVOA Ratings

It's like the way a bad personal matchup is someone's "Kryptonite" (Ronde Barber to Donovan McNabb, Bill Belichick to Peyton Manning, etc.).

Unless you have some behind-the-scenes info that we don't, I think you mean "personnel matchup."

100 Re: Week 4 DVOA Ratings

Manning has a fairly good record against Belichick, especially recently. His Kryptonite is clearly Antonio Cromartie:

November, 2007 (Manning throws six picks (three to Cromartie) in a driving rainstorm and loses by two):

January, 2008 (Manning throws two picks, one to Cromartie, and loses by four):

November, 2008 (5.8 YPA, Manning wins by three):

January, 2009 (Indianapolis loses, Manning has an okay day):

The Colts did not play San Diego in the regular 2009 season. After the 2009 season, Cromartie left for the New York Jets.

January, 2010 (Indianapolis wins 30-17, Manning throws for 377 yards)

The Colts did not play the Jets in the regular 2010 season. In the playoffs, the addition of Cromartie caused a halving of the Colts score from the year before, and the Jets won 17-16.

Obviously, some of this is tongue in cheek.

59 Re: Week 4 DVOA Ratings

There should be a special notice on this site about the upcoming Football Life: Matt Millen. I laughed out loud when I saw that one. How could they possibly do that? That's like "Franklin Pierece, a political Life", or "Ambrose Burnside, a military Life"... I mean, can you get any more incompetent??

65 Re: Week 4 DVOA Ratings

Millen was a hell of a linebacker when he still played - his being a horrible GM does not change that any more than Isaiah Thomas' stint as the Knicks GM does, or Crazy Al Davis vs Young Al Davis.

Bart Starr and Norm Van Brocklin were likewise horrible coaches, but I'd love to see "A Football Life" devoted to them.

ETA: Come to think of it, I'd love to see a few episodes devoted to figures who mix failure in with success. Marty Schottenheimer, Bud Grant, and Sam Wyche come to mind. Lyle Alzado would be fascinating. ESPN already did one on Todd Marinovich, but I think NFL Films could do better.

68 Re: Week 4 DVOA Ratings

As much as Millen's reign (of terror) as GM has scarred my psyche, I still concede that he was a heck of a linebacker.

But I'll definitely be DVR'ing that show to metaphorically watch a replay of the train wreck that was his Lions' tenure. Sometimes it's nice to think about how, even if things aren't perfect now, they used to be much, much worse.

73 Re: Week 4 DVOA Ratings

Kwame Kilpatrick murdered what little hope was left in the city of Detroit. His career as mayor failed so spectacularly it led to his mother's ouster from office and left him in federal prison.

He's still more popular in Detroit than Matt Millen. People in Detroit hate Millen so much even panhandlers won't talk to him.

62 Re: Week 4 DVOA Ratings

I know it isn't a "special" matchup, but it would be interesting to see the odds of a Denver-Seattle Superbowl included with the others at the bottom of the Playoff Odds page.

67 Re: Week 4 DVOA Ratings

In reply to by CM (not verified)

I would have thought it would be as simple as multiplying their respective odds of winning their conference championship (.445*.429 = .191). Checking that against the "special" matchups in the table seems to indicate that's not exactly the case, but it's really close. There maybe some rounding involved.

In theory, you could approximate the odds of any "special" matchup you want in this way. PIT-ARI (Bruce Arians Revenge Bowl) = .0001%.

74 Re: Week 4 DVOA Ratings

In reply to by Revenge of the NURBS (not verified)

It won't exactly come out to respective odds of winning conference championships multiplied because of interconference play. For example, the Seattle game against Indianapolis this week will affect both Seattle's playoff odds and Denver's playoff odds (because a Colts win would lower the Broncos' chances of getting the best playoff seed, while a Colts loss would increase those chances).

79 Re: Week 4 DVOA Ratings

In reply to by Revenge of the NURBS (not verified)

This is basically right. There was some discussion in the not-too-distant past about this, and it's not exact, but it's close enough as usually makes little to no difference. It would only be exact if (a) the season were already over, or (b) the AFC and NFC never played each other until the SB, ala the World Series before they started interleague play. Due to common opponents & non-common opponents who nonetheless have some potential impact on the other team's playoff seeding & likely playoff opponents, the projections at this point aren't 100% independent. The "not yet independent" effect is probably a bit stronger if they are scheduled to play each other, but even then, I assume it's still well within a few percentage points.

85 Re: Week 4 DVOA Ratings

Yeah, I saw the "Smoke-a-Bowl" somewhere a while back (can't remember if it was here) and I still laugh every time I see it. Although Super Bowl 420 is pretty good too.

86 Re: Week 4 DVOA Ratings

I find it interesting that Denver won by 32 and their playoff odds went down a hair

88 Re: Week 4 DVOA Ratings

I'm sure there's a good explanation, but how do you figure the Brownies finish only 27th in DAVE, but 7.3 projected wins despite not *that* easy a schedule and estimated wins underperforming reality so far?

96 Re: Week 4 DVOA Ratings

I think the idea is that the NFL is still very balanced, and that an average team, say Detroit or Buffalo (their next two opponents) still breaks even when playing in Cleveland. They do have the seventh-worst expected wins going forward.

92 Re: Week 4 DVOA Ratings

How much of Denver's defensive DVOA is affected by the fact they are playing with such huge leads and giving up loads of passing yards to teams playing catch-up?

95 Re: Week 4 DVOA Ratings

This was only really true against Oakland. In their first halves, Denver gave up:

32 yards per drive against NY, versus 28.0, 28.3 and -2.8 for NY's other opponents
30.5 yards per drive against Baltimore, versus 30.6, 23 and 10.0 for Baltimore's other opponents
49 yards per drive against Philadelphia, versus 39.0, 40.5 and 25.6 for Philadelphia's other opponents

So they haven't been good in the first half either. They did a good job holding many of those long drives to field goals, but might run into trouble against better red-zone teams; Baltimore, NY and Philadelphia are 22nd, 27th and 26th in points scored per red-zone trip.

98 Re: Week 4 DVOA Ratings

I just noticed that Jacksonville's playoff odds (at 0.0%) did not change this week, meaning they were listed as having zero chance to make the playoffs after week 3. Has any other team ever been that bad? It seems crazy to think that at 0-3, the numbers already said that they were completely hopeless.

(I mean, looking at what they put on the field, it's not crazy...in fact, you could have said before a single game was played that they had no chance. I'm just surprised that the prediction went down to 0.0% after only three weeks).

To flip it around, you'd think Denver and Seattle would be 100% to make the playoffs but after week 3 they were 97.7% and 99.7% respectively, and actually went down a bit this week (due to the performance of other teams moreso than theirs, I'd imagine). The only way it makes sense to think that there's a chance they might miss the playoffs is if you consider the possibility of multiple injuries to key players, which the playoff odds don't take into account.

123 Re: Week 4 DVOA Ratings

Look at it this way: If a team has a 1% chance of winning according to DVOA, and that game is simulated 10,000 times, then that team will win 100 times. Statistically speaking, some of those low odd wins will happen in the same simulation, and those will lead to your .3% chance of not making the playoffs.

There is no such thing as a 100% chance of making the playoffs, because there is no universe in which the chances of winning any single game are 100%. There is always a chance, however infinitesimally small, that a lightning bolt will strike Peyton Manning on the field and the Broncos will lose every game thereafter.

Just like the chances of two rookie QBs going out of the same game with knee injuries are quite small, yet last night it happened.

104 Re: Week 4 DVOA Ratings

Some antiquated notions on display. Rushing is as important as passing, defense is as important as offense to name two. Until you change your underlying assumptions, you'll end up with nonsense like the colts and chiefs as top 5 teams through 4 games this year, or the bears as the the 6th best team last year.
You can never hope to turn all this wonderful analysis into anything useful for prediction of future performance until you admit the truth. Having a great offense correlates to wins more than a great defense, and having a great passing offense correlates to more points than a great rushing offense.

105 Re: Week 4 DVOA Ratings

So, Seattle and San Francisco got to where they were last year on the back of great pass-heavy offenses without the help of their defenses, is that it?

Oh, and let's not forget SF 2011. A pathetic offense with only magnificent defense and special team play holding it up. 6th in the NFL? How antiquated.

106 Re: Week 4 DVOA Ratings

I don't think DVOA works the way you think it does.

Rushing and passing offense are not treated as 50% of the offense each. Total offense is based simply on the plays run and how successful they were. So if you rush on 50% of your plays it will be 50% of the total, but if you rush on 40% of your plays, it will only be 40% of the total.

Except for penalties and such that prevent rushing + passing from adding up to 100%.

107 Re: Week 4 DVOA Ratings

Or you could just look at something like drive stats for offense.

The Colts are second at yards (per drive), third at getting points, second at avoiding turnovers, and second at extending drives. It doesn't matter how they do it, through the air or on the ground, so long as they do it.

Whatever DVOA is liking, it's showing up in the results. (WARNING: drive stats are not adjusted for opponents, so take this with a grain of salt that has nothing to do with the run-pass play call mix.)

True, most teams use a pass-heavy offense to get their best results. But isn't it great that there is more than one way to have a good offense in the NFL?

108 Re: Week 4 DVOA Ratings

In reply to by mike jones (not verified)

It would be a simple exercise to see whether offensive or defensive or toal dvoa correlate better with wins. I didn't do it because the answer seems obvious. And, who cares about wins? Isn't winning the Super Bowl the goal? The playoffs, and the SB this year are a winter sport.

111 Re: Week 4 DVOA Ratings

It is easily shown that offense and defense are equally important in contributing to wins. On any given play barring special teams there is exactly one offense and one defense on the field. Anything that happens on that play happens equally to the two. Since the sum of those plays equals the game, offense and defense must necessarily contribute equally.

109 Re: Week 4 DVOA Ratings

What might be interesting on the "Denver's going to set records" comparisons would be if any of those past teams at the four-game mark wound up setting the records you cite.

And where did the record-setters sit at four games in?

114 Re: Week 4 DVOA Ratings

Tuluse: "I don't think DVOA works the way you think it does. Rushing and passing offense are not treated as 50% of the offense each. Total offense is based simply on the plays run and how successful they were. So if you rush on 50% of your plays it will be 50% of the total, but if you rush on 40% of your plays, it will only be 40% of the total."

I think you may be missing the point. Over the long term, putting up passing yards that are consistently above the rest of the league in a given situation results in more points than equally outpacing the league in a given situation with a rushing attack; so while in the past 4 games being great a running team has served the colts well as evidenced by their points, if continued it likely will not prove as efficacious as a high powered passing attack that outpaced the league in a similar manner as their run game has.

Simply put, passing competency well above the mean is worth significantly more than rushing competency the same amount above the mean because it creates more points, and thus results in more wins. The Football Outsiders treat a running play that is 30% better than league in the same situation the same as a passing play that is 30% better than the league as equally impressive.

Smart people that make their living predicting future results in the NFL know this simply isn’t the truth. That’s why they have Detroit and New Orleans higher than the FO and the Colts lower. That’s why this FO tool while wonderful analysis, can lead you to very wrong conclusions if you don’t account for the distortions that their flawed assumptions underlying everything create.


Intropy: “It is easily shown that offense and defense are equally important in contributing to wins. On any given play barring special teams there is exactly one offense and one defense on the field. Anything that happens on that play happens equally to the two. Since the sum of those plays equals the game, offense and defense must necessarily contribute equally.”

It is easily shown that great offense is more important than having a great defense. If the history of the nfl since the rules were changed in 2005 isn’t enough for you, look at this.


There is a reason the Bears were rated sixth by the FO in total efficiency last year based mostly on their impressive defense yet didn’t even make the playoffs. It simply is not as important as the having the best offense in the nfl by the same margin. Could you imagine a set of circumstances where the best offense in the nfl that was 30% more efficient than the next best offense couldn’t even make the playoffs? That’s what happened with Bears last year. The football outsiders called them one of the best defenses they’ve ever seen in their database. http://www.windycitygridiron.com/2013/8/5/4589294/football-outsiders-chicago-bears-windy-city-gridiron-marc-trestman-jay-cutler

And the Bears didn’t even make the playoffs. Yet the Football Outsiders continue to weigh offense and defense evenly, both worth 3 parts of the total. Want a couple of more examples? The 2005 Bears and 2006 Ravens were both dispatched in the first round of the playoffs even though they had been allowing about 12 points a game all season.

Tuluse: “One could argue that either the defense or the offense has less effect on the success of the play and thus contributes less to a team winning.”

This is actually what is going on I think. I found an article last year that did a study about team total points scored that found that the offense teams performance prior performance was twice as indicative in the matchup as the opposing teams defense past performance. For the life of me, I can’t find that article.

In short, passing matters more than rushing, and offense matters more than defense if you are talking about winning real nfl games.

115 Re: Week 4 DVOA Ratings

Passing and rushing offenses are compared to the same baseline I'm pretty sure. So if a team gains on average 4 yards on 1st and 10 if you throw or run for 6 yards, it's the same DVOA, even if the average rush is lower than the average pass.

"Could you imagine a set of circumstances where the best offense in the nfl that was 30% more efficient than the next best offense couldn’t even make the playoffs?"

The 8-8 2002 Kansas City Chiefs.

117 Re: Week 4 DVOA Ratings

"Passing and rushing offenses are compared to the same baseline I'm pretty sure. So if a team gains on average 4 yards on 1st and 10 if you throw or run for 6 yards, it's the same DVOA, even if the average rush is lower than the average pass."

If that's true how are the Colts the #4 ranked offense? they are 9th in yards per play and more importantly 14th in yards per pass attempt. SOS isn't even fully accounted for yet. So that jump comes

"The 8-8 2002 Kansas City Chiefs.

I'm willing to concede defense was much more important pre-2005, before Irsay had the rules changed because the patriots kept beating down those high powered Colts offenses by hand checking and beating down their receivers. In fact, that is an ssumption underlying my arguments. Back in those pre post-modern NFL days the Ravens won a superbowl with a guy named trent dilfer throwing the ball.

118 Re: Week 4 DVOA Ratings

You are confusing concepts.

Passing offense isn't more important; it just varies by a greater magnitude. A passing offense that is one standard deviation more efficient than the mean is much more effective than a one-standard-deviation-above rushing offense. The standard deviation is higher. Likewise, bad passing offenses are more putrid than bad rushing offenses. Therefore, in looking at what causes wins, good passing will be more valued.

Likewise, offense is more consistent than defense. It has a lower variance, and therefore is better correlated with wins. It is hard to argue with the logic; the offense has the ability to set the pace and call the plays, and the defense is reactive.

Now, this doesn't mean that I dispute that passing is more important; but it isn't anything about the game that has changed.

[i.e. - saying that 2005 rule changes created the passing monster is pretty silly considering the massive passing numbers of 2004.]

120 Re: Week 4 DVOA Ratings

You're both wrong and right about the rule changes. Polian complained after the 2003 Title Game (Jan. 2004), before the 2004 season.

Also, what's more important than the 2004 rule changes was whatever the hell happened to league-wide passing after the lockout. 2011 as ridiculous, 2012 was worse, 2013 is even more to the point that teams are averaging 250 ypg (essentially, if every QB stayed healthy, they would average 4,000 yards).

Something happened there, whether it be better schemes, people becoming smarter and throwing more, or who knows what, but passing has gone up more in the past three years than what the rule changes did (2005, for instance, was a particularly run heavy year).

131 Re: Week 4 DVOA Ratings

Oh I see, didn't quite follow the conversation there.

I don't think the "emphasis" on hits to the head has had a large impact on league wide defense personally.

132 Re: Week 4 DVOA Ratings

I don't think so either. I just don't know how the numbers have risen so quickly right after the lockout. Many people wrote it off to defenses being unprepared early in 2011, and it did calm down late in that season, but numbers are still up the last two years. Are things like Off-Holding down and Def-Holding, Def-Illegal Contact up the last 2+ years?

Maybe it is just getting a lot of great QBs and smarter coaches who throw more now.

119 Re: Week 4 DVOA Ratings

"If that's true how are the Colts the #4 ranked offense? they are 9th in yards per play"

Because DVOA uses success points, not yards. A 1 yard gain on 3rd and 1 is more valuable than a 1 yard gain on 1st and 10.

116 Re: Week 4 DVOA Ratings

Hey smart guy,
Yes everyone knows that great passing offenses result in more wins. If you looked further into the stats you're criticizing you'd find that FO stats back that idea up.

Look at the top values for passing offenses last year:

NE: 53.9%
DEN 49.7%
GB 40.7%

Now look at top rushing values:

SEA: 16.5%
WAS: 16.4%

Now look at top defensive DVOA:

CHI: -26.7%
SEA: -14.5%
SF: -14.4%

Obviously being a great passing offense is the best way to be the most efficient team on a per play basis, because the upper limit for how good a team can be is so much higher when passing compared to when rushing or playing defense.

124 Re: Week 4 DVOA Ratings

This is exactly it. He was arguing a moot point presuming that DVOA doesn't bare out the value of a powerful passing offense when it clearly does. Nobody at FO to my knowledge has ever suggested that powerful passing offenses aren't superior to powerful rushing offenses. Of course, being good at running is valuable, especially if you aren't able to build that top notch passing attack.

FO writers have long defended the passing attack against people who suggest that teams pass too much because running correlates with winning. It's actually the superior efficiency of passing that causes teams to pass the ball when behind creating this correlation.

In particular, I recall reading analysis defending Andy Reid's Eagles for passing the ball so often when people pointed out this correlation without examining it's cause.

133 Re: Week 4 DVOA Ratings

I'm a complete newbie to this site and to the DVOA rating. What is the rating? How is it calculated?

Thanks for any help you can give.

Anonymous Dave

135 Re: Week 4 DVOA Ratings

Thinking in the Superbowl rematch section, you could add something regarding Peyton Manning facing his hometown team again... Call it the Peyton Manning hometown rematch?

Saints vs Colts 2009

Saints vs Broncos 2013???

Also, for the national media readers... enjoy:


137 Re: Week 4 DVOA Ratings