Writers of Pro Football Prospectus 2008

Most Recent FO Features

ManningPey98.jpg

» Game Preview: DEN-CIN

Maybe the Bengals are a paper tiger, but are they really that bad in prime time games? Is Peyton Manning struggling in Denver's new offense? We detail the Monday night clash.

23 Sep 2008

Week 3 DVOA Ratings

by Aaron Schatz

Last year, the Patriots got off to such a ridiculous start that we had to create a weekly table called the "Best DVOA Ever Watch," to track how the Patriots (and Colts and Cowboys) stood among the top teams of the DVOA Era. We kept that going all through the season, and at the end the Patriots did, in fact, finish with the best DVOA ever.

If you live in the state of Missouri, you know where I'm going with this, don't you?

Worst DVOA Ever Watch

WORST TOTAL VOA
AFTER WEEK 3
  WORST DEFENSIVE VOA
AFTER WEEK 3
2008 STL -88.2%   2008 STL 55.8%
2001 WAS -88.1%   2008 DET 55.6%
1999 CLE -85.9%   2006 HOU 52.2%
2006 OAK -76.5%   2001 ARI 46.8%
2008 DET -70.4%   2000 SF 46.6%
2008 KC -68.9%   2007 NYJ 41.0%
2007 NO -68.6%   2007 BUF 35.1%
1996 TB -68.5%   2000 CHI 34.9%
2000 CIN -64.7%   2000 STL 34.7%
1996 SEA -64.1%   2007 NO 32.8%

Just like the Patriots were historically good with the Colts and Cowboys not far behind, so too are the Rams historically bad with the Lions and Chiefs not far behind. 

Detroit and St. Louis aren't just among the worst defensive teams ever after Week 3. They're up there with the worst defensive teams over any three-game stretch, period.

Detroit's average defensive VOA in its first three games is 58.5%. St. Louis' average defensive VOA is 56.3%. (These numbers are different from the defensive VOA numbers in the table because there are a different number of plays in each game, and therefore season-long DVOA does not weigh each game equally.) Prior to this season, only five other teams since 1995 had put together a three-game stretch with an average defensive VOA (no opponent adjustments) above 55%:

  • 2000 Chargers (63.1% in Weeks 3-5, 56.5% in Weeks 4-6)
  • 2000 Seahawks (57.4% in Weeks 5-7)
  • 2000 Vikings (58.9% in Weeks 15-17)
  • 2006 Texans (57.8% in Weeks 1-3)
  • 2007 Broncos (57.5% in Weeks 4-7, with a bye Week 6)

The Rams are even worse when you add in offense and special teams, with an average VOA of -88.9% over the first three games. St. Louis becomes just the 12th team to put up three straight games below -60% VOA -- interestingly, not one of these teams played from 1995-1999.

  • 2000 Bengals (Weeks 2-4)
  • 2000 Browns (Weeks 13-16)
  • 2000 Chargers (Weeks 3-6)
  • 2001 Redskins (Weeks 1-3)
  • 2001 Saints (Weeks 15-17)
  • 2002 Broncos (Weeks 15-17)
  • 2002 Cardinals (Weeks 10-13)
  • 2002 Jets (Weeks 2-4)
  • 2004 Giants (Weeks 12-14)
  • 2005 49ers (Weeks 4-7, with a bye Week 6)
  • 2005 Texans (Weeks 5-7)

We got to see the battle to be the best team ever play out on the field, as the Patriots faced both the Colts and Cowboys during the 2007 regular season. Unfortunately, this year we aren't so lucky -- the Rams, Lions, and Chiefs do not play each other at all season.

As for the Patriots, they plummet in the VOA and DAVE ratings after the huge loss to Miami. Overall, they drop from ninth to 22nd in VOA. They still have the highest DAVE rating in the AFC East, but the Bills are now listed with better playoff odds because of their one-game advantage over the Patriots in the standings. Buffalo's projected chance to make the postseason has gone up to 80 percent, while New England's chances have dropped to 56 percent. (Hey, at least the Patriots lead the league in something -- they're number one in special teams.) Thanks to strength of schedule, the team with the best chance of landing the number one overall pick isn't one of the teams challenging the all-time record for worst DVOA -- it's Cleveland. Hey, maybe that makes up for not having any picks in the first two rounds last year.

* * * * *

Before we get to the numbers, a little website housekeeping.

The midseason KUBIAK update will be available later this week, probably late Thursday. It takes me some time to put it all together.

All the individual stats pages are now updated, and you'll find the first editions of our offensive line and defensive line stats. Team stats are updated, including the first edition of "defense against types of receivers." Next week, we'll begin running the second DVOA table that includes variance, schedule strength, and "estimated wins." However, due to the Jewish holidays, next week's DVOA will go up on Monday and will not include the Ravens-Steelers game.

We're still working on getting past stats pages in, but hopefully that will be set by the end of this week. Unfortunately, we're not yet able to update the Premium pages with 2008 DVOA splits. Sean McCall, who handles that database, lives in Houston, and believe it or not he is still powerless and staying with friends because of Hurricane Ike. We'll try to get the 2008 numbers as soon as we can. We'll also be adding a couple of oft-requested new views, including DVOA through each week of each season as it developed from game to game (complete with changing opponent adjustments throughout each year) plus run/pass offense and defense on a per-game basis.

There is a brand new way to sign up for the Football Outsiders "New Articles" notification newsletter. If you have registered for the site, go to My Account, scroll down, then click on "my newsletters" and you will be able to check a box for the "New Articles" newsletter. We're working on something for non-registered users to still be able to subscribe to the newsletter, and we will be shutting down the old notification e-mail list on October 3.

Finally, for all those asking -- yes, we are working on trying to solve the problems with getting the site to look correct in IE 6.0.

* * * * *

These are the Football Outsiders team efficiency ratings through three weeks of 2008, 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 VOA are adjusted to consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS VOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season.

There are no opponent adjustments in VOA until the fourth week of the season, which is why it is VOA right now rather than DVOA. 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 VOA to get a more accurate forecast of how a team will play the rest of the season. Right now, the preseason projection makes up 55 percent of DAVE (75 percent for Baltimore and Houston).

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>

 

  TEAM TOTAL
VOA
LAST
WEEK
TOTAL
DAVE
RANK W-L OFFENSE
VOA
OFF.
RANK
DEFENSE
VOA
DEF.
RANK
S.T.
DVOA
S.T.
RANK
1 BAL 73.5% 2 24.2% 5 2-0 10.9% 13 -64.3% 1 -1.8% 23
2 NYG 52.7% 1 27.1% 3 3-0 35.2% 2 -12.9% 6 4.6% 7
3 PHI 45.4% 8 38.8% 1 2-1 21.4% 7 -20.6% 5 3.4% 10
4 TEN 44.2% 11 18.5% 7 3-0 4.5% 15 -48.5% 2 -8.7% 31
5 BUF 35.8% 3 8.1% 14 3-0 11.9% 11 -21.2% 4 2.7% 14
6 ATL 35.0% 17 -6.6% 23 2-1 22.1% 6 -8.3% 10 4.6% 8
7 DAL 31.4% 4 20.5% 6 3-0 31.4% 4 3.0% 17 3.0% 11
8 GB 29.9% 6 33.6% 2 2-1 18.9% 8 -4.4% 13 6.6% 4
9 SD 23.3% 13 25.0% 4 1-2 35.1% 3 14.2% 23 2.5% 15
10 PIT 20.0% 5 12.6% 10 2-1 -7.0% 25 -24.2% 3 2.8% 13
11 SF 20.0% 21 -6.3% 22 2-1 7.8% 14 -12.5% 7 -0.2% 19
12 DEN 16.5% 12 9.0% 13 3-0 45.0% 1 22.1% 26 -6.4% 28
13 TB 14.4% 10 17.4% 8 2-1 1.9% 18 -11.2% 9 1.3% 17
14 ARI 13.9% 7 5.5% 15 2-1 15.7% 10 0.9% 15 -0.9% 21
15 CHI 12.6% 14 4.5% 16 1-2 1.4% 19 -7.1% 11 4.1% 9
16 WAS 12.5% 15 0.5% 18 2-1 24.4% 5 2.9% 16 -8.9% 32
  TEAM TOTAL
VOA
LAST
WEEK
TOTAL
DAVE
RANK W-L OFFENSE
VOA
OFF.
RANK
DEFENSE
VOA
DEF.
RANK
S.T.
DVOA
S.T.
RANK
17 MIN 7.6% 18 16.4% 9 1-2 3.7% 16 -12.1% 8 -8.3% 30
18 MIA 4.2% 26 -9.8% 24 1-2 11.5% 12 -0.6% 14 -7.8% 29
19 SEA 3.4% 25 10.1% 11 1-2 2.2% 17 -6.5% 12 -5.3% 26
20 OAK 0.8% 16 -12.7% 26 1-2 -2.7% 21 6.1% 19 9.6% 2
21 CAR -7.5% 19 -1.0% 20 2-1 -6.2% 23 6.1% 18 4.8% 5
22 NE -8.8% 9 9.4% 12 2-1 -3.7% 22 16.5% 24 11.4% 1
23 NO -11.5% 23 -4.7% 21 1-2 16.3% 9 26.9% 30 -0.9% 20
24 IND -13.9% 22 4.4% 17 1-2 0.4% 20 10.3% 21 -4.1% 24
25 JAC -14.5% 24 -0.8% 19 1-2 -10.5% 26 8.6% 20 4.6% 6
26 NYJ -19.8% 20 -11.7% 25 1-2 -6.5% 24 13.8% 22 0.6% 18
27 CIN -52.9% 31 -24.9% 28 0-3 -29.8% 28 18.6% 25 -4.5% 25
28 CLE -58.6% 27 -33.6% 29 0-3 -39.2% 31 26.5% 29 7.1% 3
29 HOU -63.5% 29 -21.0% 27 0-2 -37.6% 30 24.8% 28 -1.1% 22
30 KC -68.9% 28 -40.2% 30 0-3 -39.6% 32 23.8% 27 -5.5% 27
31 DET -70.4% 30 -43.6% 31 0-3 -16.4% 27 55.6% 31 1.6% 16
32 STL -88.2% 32 -54.8% 32 0-3 -35.2% 29 55.8% 32 2.8% 12

Posted by: Aaron Schatz on 23 Sep 2008

98 comments, Last at 20 Nov 2009, 11:08am by Dana DeArmond

Comments

1
by asg (not verified) :: Tue, 09/23/2008 - 5:10pm

Wow, the Ravens' managing a near-11% offensive DVOA with a rookie QB, a retired Jonathan Ogden, no star WR, and a basically untested RB committee is pretty impressive. Admittedly that's based on two games rather than three, but still.

5
by CincyFan (not verified) :: Tue, 09/23/2008 - 5:19pm

Of course, most importantly: they've faced the Bengals' and Browns' defenses, which are probably 2 of the worst in the NFL.

After opponent adjustments are brought in, I bet that +11% will be negative.

38
by Thump (not verified) :: Tue, 09/23/2008 - 7:01pm

I agree to a point -- their competition has been weak. But so far this year, very few teams have really stepped up with the defense needed to stop an offense. A good defensive team could very well go a long way -- even with an average offense, it will mean a lot.

It will be interesting to see where all the teams are after 8 weeks.

2
by Cosmos (not verified) :: Tue, 09/23/2008 - 5:13pm

Wow, VOA truly hates Dallas.

6
by anachronist (not verified) :: Tue, 09/23/2008 - 5:20pm

I don't quite understand how stopping a good defense on 3rd down a lot and mostly keeping them from scoring TDs didn't really shift Dallas's defensive ranking. Oh well. It's early in the season, still.

27
by Pat (filler) (not verified) :: Tue, 09/23/2008 - 6:19pm

You mean "good offense," not "defense." But there's no adjustment for strength of opponent yet, so Green Bay is just your everyday generic offense. And holding a team to 3 FGs and a TD is just average-ish, especially when two of those field goals were really short. Holding a team to a field goal when they're inside the 10 is one of those "high leverage" plays - stop them, and it's a 4-point difference on the scoreboard, but because it's only a single play (or a few plays), it's not nearly as repeatable a performance as the rest of the drive.

Again, though, I'm not trying to say that Dallas's defense is 'bad' or they were 'lucky' or anything like that. They had average defensive production against a team that will probably be very good, which is a very good defensive performance.

41
by Temo :: Tue, 09/23/2008 - 7:25pm

Well, opponent adjustments aren't in yet (which will raise the Philly/GB games and depress the Cleveland game).

Also, I'm sure VOA doesn't like the ultra-efficient drives the Packers had when down 18 with 5 and a half minutes left. It doesn't know it was a soft protect-the-sidelines defense that did its job.

Also, it REALLY hates Romo's inconsistent performance Sunday Night.

3
by Cosmos (not verified) :: Tue, 09/23/2008 - 5:16pm

Wow, VOA truly hates Dallas. Why not just put them at 32?

30
by In_Belichick_We... :: Tue, 09/23/2008 - 6:26pm

I'm sure they would have put them at 32 if they had more room on the chat board for your brilliant comments.

57
by The Ninjalectual :: Tue, 09/23/2008 - 11:10pm

Please use the proper template:
[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]

[/Redskins homerism]

95
by MassEagleFan (not verified) :: Fri, 09/26/2008 - 6:33pm

Ok, I'll do it for you..

1. Eagles
2.....

32. Dallas

Sorry, but Westy doesn't fumble. Your win was a fluke.

4
by Cosmos (not verified) :: Tue, 09/23/2008 - 5:19pm

Wow, VOA truly hates Dallas. Why not just put them at 32 and be done with it?

7
by Cosmos (not verified) :: Tue, 09/23/2008 - 5:21pm

P.S. CAPTCHA is Crap

8
by BucNasty :: Tue, 09/23/2008 - 5:23pm

Look at the split between Atlanta's VOA and DAVE.

VOA ATL: 35.0% 6th
DAVE ATL: -6.6% 23th

14
by tuluse :: Tue, 09/23/2008 - 5:38pm

I would bet Atlanta ends up in the lower middle of the pack, but they have an extremely high variance, beating up weak teams, and losing to good teams.

32
by the silent speaker (not verified) :: Tue, 09/23/2008 - 6:34pm

Actually, I'm surprised that that would be statistically a bad thing. If you beat up on bad teams, that suggests that you're in a tier above them. If good teams beat up on you, that suggests that they're in a tier above you. When the Giants last year were 5-0 against teams wih 6 wins or fewer, 5-2 against teams with 7-9 wins, and 0-4 against teams with 13-16 wins, I took that to mean that 10-6 was exactly where based on skill they deserved to be.

81
by SamoanRob_ (not verified) :: Wed, 09/24/2008 - 5:23pm

and the Giants were 1-0 against teams with 18 wins

15
by Hummingbird Cyborg :: Tue, 09/23/2008 - 5:43pm

They played KC and Detroit in two of the first three games.

They got blown out against TB.

Actually, it will be a little bit interesting to see if those first three games look even worse for KC and Detroit once Atlanta's adjustments have been factored in.

For that matter, KC played Atlanta and Oakland in two of its first three games. Maybe Oakland is a bit better than we thought, but it would surprise me if they have a positive DVOA at the end of the season.

And, to be fair, St Louis has played Philly and the Giants, so they may have a boost.

Anyhow, we'll just have to wait and see.

50
by Just Another Falcon Fan (not verified) :: Tue, 09/23/2008 - 7:58pm

"Blown out" is an overstatement with respect to the Tampa Bay game. They fell behind early, struggled back, and were within 8 with 4 minutes to go before Earnest Graham broke a long run to put the game away. The game was closer than the final score.

79
by English Atlanta Fan (not verified) :: Wed, 09/24/2008 - 4:41pm

Yeah, the Tampa Bay game - as I'm presuming VOA knows - was more competitive than it looked. Our offense was pretty bad, but our defense was quite good.

9
by Hummingbird Cyborg :: Tue, 09/23/2008 - 5:26pm

It's always interesting to check out what the VOA of your opponents is so that you might try to gauge just a bit what they'll look like after adjustments.

Oakland was defeated pretty soundly by Denver and it put Denver near the top of VOA after week one. Still, Oakland was always suspected of being a bad team, so that should come as no huge surprise. They blew out KC, but that clearly doesn't mean a lot, but they hung close to Buffalo who is ranked very well. They're defense is playing just a bit below average, but it shouldn't hurt Denver that much when the opponent adjustments kick in.

San Diego clearly has a terrific offense and Denver played pretty evenly with them. That San Diego ranks third in offensive VOA is nice to see because it lends hope that in the future Denver will be able to slow down some offenses that aren't quite as good. On the other hand, San Diego ranks pretty below average on Defense so far. So, this means that the offense probably will come down a bit in opponent adjustments.

New Orleans looks similar in that they have a strong offense, so the horrid looking defense will probably look somewhat less horrid with adjustments, but their defense is very weak and this will make Denver's ranking come down somewhat.

Now, it's hard to imagine Denver not being 4-0 next week against KC, but I still wish that they weren't quite that terrible. That means that Denver has to really blow them out to keep a decent ranking.

On the other hand, it will be a good chance to inflate the old VOA.

29
by Richard :: Tue, 09/23/2008 - 6:24pm

I think Denver's rating for that game will look better as the Chargers' defense improves actually.

36
by Bronco Jeff :: Tue, 09/23/2008 - 6:53pm

I'm not making any excuses for NO, or SD for that matter, but IMO a large part of their poor VOA ratings have to do with the fact that the Broncos are so good on offense. And make no mistake, the Broncos offense is awesome!

Conversely, the Broncos' defensive VOA is lower because they faced two excellent offensive clubs--not that the Broncos have been good on defense...just good enough to win, apparently.

Whatever it is, I'll certainly take it!

Eschew Obfuscation!

10
by Escoot (not verified) :: Tue, 09/23/2008 - 5:27pm

Once the teams are adjusted next week you'll see Dallas go way up since they played Philly and Green Bay.

11
by Dales :: Tue, 09/23/2008 - 5:28pm

It is so strange seeing Eli Manning near the top of the DVOA and DYAR rankings after three weeks. His unusual career path continues to be... unusual?

12
by tuluse :: Tue, 09/23/2008 - 5:30pm

This feels much more like an NFL season, all the top teams have flaws that can be exploited. The DVOA curve is much less steep, and I think every team from Jacksonville up can believe they can beat any other team, with a little luck.

Also, Chicago has managed to pull it's DAVE into the top half of the the league, I hope Lovie can keep the team together, so we can win some of the games we're due to.

In the words of Greg Easterbrook, "Don't look now, but" the Bears have a positive offensive DVOA, and Kyle Orton basically average with -0.5%.

65
by DB (not verified) :: Wed, 09/24/2008 - 9:45am

Actually, doesn't a -0.5 VOA mean that he's just about replacement level. And isn't that quite a bit below "average"?

73
by tuluse :: Wed, 09/24/2008 - 1:49pm

No, the A in DVOA is average, Orton is has been worth 66 yards above replacement. Also, Orton is ranked 17th among QBs, pretty much average.

13
by tuluse :: Tue, 09/23/2008 - 5:33pm

double post

16
by Yaguar :: Tue, 09/23/2008 - 5:48pm

The DAVE projection for the Rams is astounding. -54.8%. I remember at this time last year, DAVE had the Patriots projected at around 56%. I asked Aaron whether they'd really finish with a DVOA like that, and Aaron thought they wouldn't, but then they did. (52%.) So, there's a very real possibility that the Rams end the season with a DVOA of around -55%, or, slightly worse than the regular season 2007 Patriots were good.

Question: Why don't opponent adjustments start now? Sure, they shouldn't be full strength or anything, but I think we can safely say that the Falcons have had an easy schedule so far, for example.

43
by Temo :: Tue, 09/23/2008 - 7:30pm

Question: Why don't opponent adjustments start now? Sure, they shouldn't be full strength or anything, but I think we can safely say that the Falcons have had an easy schedule so far, for example.

Sure it SEEMS safe, but it also seemed safe to bet that the Pats wouldn't lose by multiple scores against the Dolphins. Point being, as sure as you think you are, more sample size is always good.

17
by doktarr (not verified) :: Tue, 09/23/2008 - 5:49pm

I'd like to see DAVE include opponent-adjustments (based on offensive and defensive DAVE). That would be a better projection IMO. We would certainly see clearer pictures of things like Baltimore's or Atlanta's offenses.

60
by Becephalus :: Tue, 09/23/2008 - 11:43pm

It doesn't matter if you think it would make a better prediction IYO. It matters whether adding in such adjustments would have made DAVE more predictive in the past.

The Wire should win the Nobel prize for literature.

61
by Dales :: Wed, 09/24/2008 - 12:12am

Italics...off?

18
by Giovanni Carmazzi (not verified) :: Tue, 09/23/2008 - 5:52pm

And for one week, the niners rock VOA.

31
by sfckoski :: Tue, 09/23/2008 - 6:34pm

Yeah! Take that DAVE!

Someone take a screen shot, this may be the highest the Niners get all season.

39
by Thok (not verified) :: Tue, 09/23/2008 - 7:04pm

Honestly, as a Niners fan I find it more important that Dave has the Niners as a mediocre 6-8 win team as opposed to the worse team ever. This is significant progress.

19
by Tom Gower :: Tue, 09/23/2008 - 5:59pm

TEN's two main RBs makes for an interesting comparison: Johnson is 29th in Success Rate but 9th in DVOA and 8th in DYAR, while White is 10th in Success Rate but 27th in DVOA and 26th in DYAR. Can you guess which one of those two guys may not be faster than me? I wonder just where White would be without those 4 TDs.

20
by Fan in Exile :: Tue, 09/23/2008 - 6:00pm

Okay I was all excited to come in and see a time when the Bronco's special teams rating wasn't horrible. What did I miss?

47
by Temo :: Tue, 09/23/2008 - 7:44pm

Darren Sproles?

58
by Fan in Exile :: Tue, 09/23/2008 - 11:25pm

They played the Saints in week 3, so I'm guessing that Sproles didn't have an impact.

63
by Fan in Exile :: Wed, 09/24/2008 - 8:07am

So does no one understand why the Broncos are so low or is it that this post is buried in the middle of the thread now and no one reads it?

76
by Johnsobs (not verified) :: Wed, 09/24/2008 - 2:37pm

Darren Sproles

68
by Temo :: Wed, 09/24/2008 - 10:31am

You understand that the Ratings are for the entire season and not just one game, right?

And that Sproles had a 103 yard TD return in week 2?

They also allowed quite a bit of KO return yardage against the Raiders, if my memory serves.

71
by Fan in Exile :: Wed, 09/24/2008 - 12:19pm

They had one bad kick off return against the Raiders. Neither that nor Sproles return explain them ranking 28th. Certainly they are a part of it, but a very small part. What does explain it though is the 30th ranking that they have from DAVE. It's not what they've done this year it's how bad they were last year.

75
by Temo :: Wed, 09/24/2008 - 2:22pm

The offensive, defensive, and Special Teams VOA are not affected by DAVE. Only the total VOA is affected DAVE and that number is listed in the DAVE column.

And I do think that a couple of bad returns in coverage while doing nothing on your own returns would warrant a -6.8% Special teams VOA, considering the small sample size in question.

EDIT: Special Teams DAVE and more detailed information is found here: http://footballoutsiders.com/stats/teamst

Note that they have a terrible Kickoff record, and they are merely below average in every other facet of special teams. But that adds up to being quite bad overall.

82
by Fan in Exile :: Wed, 09/24/2008 - 9:35pm

If you look at the numbers none of them add up to -6.8%, even the -4.7 for kick offs is way below -6.8%. Of course it isn't terribly what relationship they have with each other. It seems the adjustments that they put in which lower it to -6.8%.

Of course you would think that the two bad returns would sink their VOA people always hold on to their explanations.

85
by Temo :: Thu, 09/25/2008 - 8:14am

Ok look... lets say over the course of the season you give up 2 big Kick off returns every 3 games... that would be pretty bad, no? Well that's ALL they have thus far-- the sample size is SMALL. Also, it's not just the kickoff coverage, it's that their returns haven't done much at all. Only San Diego has had a worse kick off coverage team, and they gain lots of points on returns to make up for it. Denver has done nothing on returns.

And the points for the individual parts of special teams are expressed as points gained/lost compared to league average. So if the league average starting field position after a kick off is the 30 yard line and your team consistently starts from the 25 yard line, you lose points based on lower field position. Based on that, the VOA is constructed.

86
by Fan in Exile :: Thu, 09/25/2008 - 9:05am

Temo,

I understand your argument I just disagree with it, you don't have to keep trying to explain what's pretty obvious.

Did you not look at their non-adjusted VOA? It's only -2.1% how does it go from there to -6.4%? The numbers only seem to be the 28th worst when you don't compare them to the teams that are around them in the rankings.

You have to go all the way up to Baltimore at 23 to find a non-adjusted VOA that's better. That and the other teams up there all have some really bad numbers as well. Look at Seattle's -9.1, or K.C.'s terrible numbers.

It looks to me like it's the adjustment that's killing them and that doesn't make any sense to me.

88
by Temo :: Thu, 09/25/2008 - 10:32am

I understand what you're saying now; I didn't know you were talking about the weather adjustments. There are weather adjustments for special teams for cold, warm, dome, and Denver. Since the broncos, well, play in Denver, they get a downward adjustment for special teams.

Their kicking and punting is below average for the league anyway, and should be normally above average because of the thin air in Denver (where they've played 2 games). Also, they played in Oakland, where it was warm.

By the way, these aren't my arguements, I'm merely trying to explain why DAVE had nothing to do with the low special team ranking.

21
by joenamath :: Tue, 09/23/2008 - 6:00pm

rookie quarterback and wins over browns and bengals
number 5 dave?

28
by In_Belichick_We... :: Tue, 09/23/2008 - 6:23pm

I'm sure you already know this, but....
DAVE does not compensate for opponent or the age of the QB.

52
by Eddo :: Tue, 09/23/2008 - 8:16pm

Actually, DAVE might adjust for QB age. It is a projection system, after all, so you have to think it would look at a first year starter and project that he will commit more turnovers, thus affecting the team projection.

Likewise, a very old player would probably project to a higher likelihood of injury, thus affecting the team projection.

40
by BucNasty :: Tue, 09/23/2008 - 7:14pm

DAVE is a formula which combines our preseason projection with current VOA...

And there you have it. They had a high preseason projection, if we're talking about what was in the book. And obviously, they've been playing very well on a per play basis thus far as evidenced by their #1 ranking and insanely high VOA after two games. Combine the two and I think it's probably their preseason projection, and not their play this year, that is keeping them out of the number 1 spot. They'll start dropping pretty soon, but Aaron's been emphatic that they'll be a good team this year. Don't be surprised to see them in the top 10 at year's end, which would make their 5th ranked DAVE spot after two games make a lot of sense.

22
by joenamath :: Tue, 09/23/2008 - 6:00pm

rookie quarterback and wins over browns and bengals
number 5 dave?

23
by Mikey :: Tue, 09/23/2008 - 6:05pm

Anyone else see this at the top of hte page?

"You can't request more than 20 challenges without solving them. Your previous challenges were flushed."

25
by In_Belichick_We... :: Tue, 09/23/2008 - 6:16pm

Um, no.
"You can't request more than 20 challenges without solving them. Your previous challenges were flushed."

24
by In_Belichick_We... :: Tue, 09/23/2008 - 6:12pm

Hmmm, why did the Patriot's coaching staff throw a divisional game? Hmmm. Why would you move your linebackers from the center of the field against a team that loves to run and has a weak armed QB who makes a living of the underneath stuff? Hmmmm, maybe someday Bill will explain to us why he threw this game, much like the game at home against SD a few years back (also before a bye week?).
Hmmm...

35
by cowfez (not verified) :: Tue, 09/23/2008 - 6:52pm

Get to a doctor, quickly.

26
by Bowl Game Anomaly :: Tue, 09/23/2008 - 6:18pm

Opponent adjustments will show that the Redskins D is a lot better than their current league-average ranking. Each of their first 5 games features an opponent currently in the top 10 in offensive VOA.

(Formerly "The McNabb Bowl Game Anomaly")

33
by pressrow (not verified) :: Tue, 09/23/2008 - 6:46pm

Wow, 49ers. Wow. I know their last win came against this epically awful Detroit team, but they did beat Seattle on the road in OT. Remember, there were a few weeks last year where their offense was featured in the "Worst offense of the DVOA Era" section. Now, they are 14th. And a 7th-ranked defense! With a minimal pass rush and (I'm sure) one of the worst 3rd down percentages in the NFL.

Amazingly, their only strong point last year--special teams--is now sitting in the second half of the rankings, with virtually no change in personnel or coaching.

44
by Temo :: Tue, 09/23/2008 - 7:34pm

Amazingly, their only strong point last year--special teams--is now sitting in the second half of the rankings, with virtually no change in personnel or coaching.

I think that is best explained by "Andy Lee can't boom punts from his own 10 yard line anymore"

55
by Jimmy :: Tue, 09/23/2008 - 9:10pm

The Takeo Spikes fumbled kickoff won't be helping matters either.

62
by Karl Cuba :: Wed, 09/24/2008 - 7:36am

I think they pulled the same trick against arizona as well, though I could be mistaken.

Lee has not quite been the star he was last year but not all of it is his fault. Seattle got a decent return on a punt that had to be re-kicked because of a penalty and re-kicks have a tendency to get punished as the coverage unit is tired.

70
by 49ers billy (not verified) :: Wed, 09/24/2008 - 10:58am

woo hoo niners made the top 11! (Arbitrary cutoff point). I dont think I have ever seen them above 20 except in a current time DVOA rating.

34
by Telamon :: Tue, 09/23/2008 - 6:51pm

Congratulations Detroit on eliminating yourself (by DAVE's estimation) from playoff contention in a scant 3 weeks. That's impressive. I can already see Michael Crabtree in a Lions' uniform.

Also, my mind refuses to accept the strong statistical performance thus far by the 49ers. It just doesn't seem possible.

64
by zlionsfan :: Wed, 09/24/2008 - 8:10am

Detroit eliminated itself from playoff contention some time between the 2000 and 2001 seasons. Not surprisingly, even the Lions' official site refuses to divulge the exact date on which the franchise chose its current path to ruin.

I would guess that the vast majority of fans of teams not in Missouri or Michigan are very, very, very happy that the Rams, Chiefs, and Lions have 0 games between them this year. As for the rest of us, well, I for one am glad none of them play. One more game against a weak team is one more chance to get a win that Mr. Ford will see as the "turning point" for this franchise.

I am not looking forward to DVOA, not unless San Francisco turns out to be a solid team ... maybe I should pick up a customizable Lions jersey and have them put "EPIC FAIL" on the back.

37
by Dales :: Tue, 09/23/2008 - 6:57pm

I just noticed-- the Bengals have played the top team, the second team, and the fourth team in total VOA to this point.

ETA-- and 1, 2, and 6 in defensive VOA.

42
by Bobsaget (not verified) :: Tue, 09/23/2008 - 7:27pm

"DAVE is a formula which combines our preseason projection with current VOA to get a more accurate forecast of how a team will play the rest of the season."

Wait... so they're using subjective predictions here, right? Not any kind of preseason VOA? I guess that explains why Dallas remains so low on the list despite being at the top of everyone else's power rankings.

What if, instead, we factored in tabulated power rankings from some well-respected sites rather than holding so strongly to weeks-old predictions? That might make DAVE a bit more accurate until strength of schedule starts to come into focus. I know that the idea is to keep from allowing a knee-jerk reaction here, but haven't all of our perceptions of teams like the Bills changed in the past few weeks to the point that they shouldn't be punished for the mistakes analysts made in the preseason?

45
by Dales :: Tue, 09/23/2008 - 7:38pm

Dave uses less and less of the preseason projection each week. And they do it the way they do it because it is the way that has worked most accurately so far of the different methods they have tried.

46
by Pat (filler) (not verified) :: Tue, 09/23/2008 - 7:43pm

It's not subjective. The preseason projection is statistical. Dallas is "so low" (jeez, they still have a 20% DAVE, f'crying out loud) because 1) over the past 2 years Dallas consistently has faded at the end of the year, which tends to be a marker for decline, and 2) Dallas has had extremely good health over the past few years, so if you project them based on average health, you get a much weaker team.

Note that health is one of those factors that'd show up at the end of a season, not the beginning - so it's entirely possible that while the DAVE projection for Dallas seems low for them now, it'll start looking a lot more reasonable as the season progresses.

Put another, simpler way: for the past two years, Dallas has looked great in September, and a lot more mortal by December. DAVE's taking a longer view than most fans.

48
by Bowl Game Anomaly :: Tue, 09/23/2008 - 7:50pm

Nope, not subjective. The preseason rankings are based on statistical projections. The only subjective element is determining which players are starters and which are subs.

Dallas had a low-ish statistical projection. I'm not sure why. It had something to do with expected injuries, but that's all I can say, and I didn't agree with it. If you look more closely at the chart you'll see that Dallas is ranked one spot lower in VOA than DAVE, so the preseason projections are actually bumping them up slightly. (They will go higher when opponent adjustments are added next week, possibly all the way to #1.) In any case, it may not be accurate for Dallas specifically but DAVE has been shown to be a better predictor than early season VOA of teams' season end DVOA. I know you held up Buffalo as a counterexample, but look at Atl, SF, NE, and Ind and try to tell me they are described better by VOA than DAVE. I think we should temper our expectations for Buf, and by the same token consider that Sea and Min might actually be a bit better than they seem.

EDIT: Pat beat me to it, so here's a link to the projections so my post isn't a waste.
http://www.footballoutsiders.com/dvoa-ratings/2008-dvoa-projections

(Formerly "The McNabb Bowl Game Anomaly")

51
by Dan :: Tue, 09/23/2008 - 8:06pm

DAVE combines offseason projections with regular season performance. The offseason projections are not subjective - they're based on equations that take into account things like last year's performance, injuries, and personnel changes. I believe that they've run regressions to figure out how much weight to give to the offseason projections and how much to give to this year's performance after each week.

67
by vinyltoupee (not verified) :: Wed, 09/24/2008 - 10:09am

Say what you want about Dallas. They're low. They're under two teams that they've beaten. It presents an anomaly to be ignored in the formula or an imperfection to be addressed.

72
by zip (not verified) :: Wed, 09/24/2008 - 1:09pm

correct, the NFL is an unpredictable place, DVOA/DAVE/subjective columnists are going be "imperfect" on a regular basis.

What makes this site interesting is the lack of subjectivity. There's a million sites with power rankings putting the Cowboys at #1 based on whatever subjective criteria the author uses... if all you want is to see the Cowboys at #1, then go there.

FO doesn't make snap adjustments to their math based on one team looking better than their stats say after three weeks. That's a good thing. If they go 16-0 and are still rated low -- THEN you might see some adjustments.

74
by Pat (filler) (not verified) :: Wed, 09/24/2008 - 2:00pm

Ah yes, it's the other DVOA/DAVE rating criticism template. Use long words, sound authoritative, and use an appeal to the majority.

It's 3 games into the season. Everyone else has Dallas #1 because Dallas is a big-market team that had high-profile wins. There's not really any information that Dallas is better than Baltimore or the Giants, for instance, and the information that Dallas is better than Green Bay and Philadelphia is shaky, because it's based on only one game in both cases.

77
by vinyltoupee (not verified) :: Wed, 09/24/2008 - 3:41pm

Pat,

My apologies if long words intimidate you. Let me use shorter words. and sentences.

What other sites' power rankings say don't matter.

I am not talking about Baltimore or NYG.

Small samples are all you get in the NFL. The information is not shaky. Dallas won. Philly lost. Dallas won. Green Bay lost. Therefore, Dallas is a better team than Philly and GB.

Any other conclusion is illogical.

78
by tuluse :: Wed, 09/24/2008 - 4:32pm

First of all, Dallas is ranked ahead of Greenbay in VOA, secondly, don't the other two games that Philly played mean anything? Sure Dallas out played Philly, but if the Eagles significantly out played Dallas in their other two games, they could conceivable be ranked ahead.

80
by Pat (filler) (not verified) :: Wed, 09/24/2008 - 5:22pm

The information is not shaky.

And therein is the problem in your analysis.

One game is shaky information. They won. So what? Is beating another team a guarantee you will beat them in the future? No - far from it. 4-point victors in the NFL win repeat games approximately 50% of the time - in other words, a 4-point victory against the Eagles is, statistically, a tie. Bully for the Cowboys, as they get a victory, but this isn't college football - you're ranking teams based on whether or not you think they're better than the other teams.

Is there any other 'direct' information that the Cowboys are better than the Eagles? No. So, it's shaky.

The idea that "team A beat team B, so they should be put above them!" can't work without some way to resolve the resulting loops. And, if you need a mechanism to discount "team A beats team B, therefore is better," why restrict the mechanism to other games?

The idea that "team A beats team B, therefore, they're better" is sometimes wrong. That's obvious - Miami beat New England who beat the Jets who beat Miami. And you can easily show that beating a team, at home, by 4 points, does not mean you're better than them. It has absolutely no predictive value for the next game whatsoever.

83
by KyleW :: Thu, 09/25/2008 - 4:02am

Speaking of "Team A should be above Team B because they beat them" does anybody know if the site that does (did?) the NFL beatpaths is still in existence?

87
by Eddo :: Thu, 09/25/2008 - 9:09am

Yeah, it's still around at www.beatpaths.com.

90
by Pat (filler) (not verified) :: Thu, 09/25/2008 - 10:08pm

I just want to point out, as I've pointed out elsewhere (and as I mentioned above) - the idea behind beatpaths is a bit flawed. Well, it's not flawed - it is what it is, and it's interesting, but claiming that beatpaths Best Represents the idea that the thing which should Drive All Rankings is "if team A beat team B, team A is above team B" is wrong.

You can't do rankings based on that. Beatpaths doesn't do it, either - the "beatpower rankings" are relatively arbitrary. Plus, the idea that third-and-fourth order information can invalidate first-order information (team A beats B beats C beats D, and team D beats A - the fourth-order information that team A is better than team D invalidates the first-order information that team D beat team A) is questionable, at best.

But it's really great to look at those graphs, because then you can realize how completely arbitrary it is to try to turn that twisted, very horizontal graph into a straight and completely vertical one.

94
by Eddo :: Fri, 09/26/2008 - 3:34pm

Pat, every single thing you just said is true. However, I will defend Beatpaths a little. In everything I've read over there, it seems to me that the creator was just more interested to see if he could create rankings based on W/L data only, and if so, how accurate could he get them. He often points out flaws in the rankings.

As to your point about third-order information invalidating first-order information: doesn't DVOA do this a little bit? Hypothetically, using the 2007 quality of teams, the Packers (a very good team) could outplay and beat the Seahawks (another very good team) by an average amount (neither a blowout or a close game) in week n. This would suggest the Packers should be ranked ahead of the Seahwaks. Now, let's say, in week n+1, the Seahawks dominate the Eagles (a good team), but the Packers struggle against the Lions (a bad team), and nearly lose. Couldn't you conceivably see DVOA rating the Seahawks ahead of the Packers after week n+1? The third-order information (the Eagles being good and the Lions being bad) is effectively negating the first-order information (the Packers being better than the Seahawks).

Or am I classifying third-order information improperly?

I think all Beatpaths shows is that rating teams discretely is a futile action.

49
by Aaron M. (not verified) :: Tue, 09/23/2008 - 7:57pm

After watching the Chiefs play, they shouldn't be ahead of any teams. If they played STL or DET they would get manhandled.

54
by Temo :: Tue, 09/23/2008 - 9:04pm

Unless the Chiefs are starting high school players now, I don't see the Rams manhandling them.

53
by Reinhard (not verified) :: Tue, 09/23/2008 - 8:23pm

Last year I think there was a serious problem with the O coordinator... didn't they bring in some guy to "advise" the qb coach and then that very same week had their best offensive showing of the year? You have to think going from that to mike martz must be extremely helpful. And beating Seattle and Detroit... definitely will hurt a lot with opponent adjustments. And they have J Smith and Lawson, which probably helps out their defense a lot over last year.

84
by David :: Thu, 09/25/2008 - 5:01am

Yup - Hostler was expected to be the QBs coach, and got elevated to OC when Norv (somewhat unexpectedly) got offered the SD HC job. He was a bit out of his depth, and the offense picked up a little when Ted Tollner was brought in, around week ten

56
by The Ninjalectual :: Tue, 09/23/2008 - 11:04pm

The Redksins are clearly ranked too low because they are young and still learning how to play football. Last week's games are way better than this. Go Redskins woot woot WOOOOOOOOTTTTT!!!!!!

[/Redskins homerism]

59
by buzz (not verified) :: Tue, 09/23/2008 - 11:42pm

So the playoff odds say that the Titans are 80% to win the division. I wish in this case FO was setting the odds. I would take the field and 4-1 odds that someone else wins that division despite the 2 game lead.

66
by panthersnbraves :: Wed, 09/24/2008 - 9:53am

Well, everyone scoffed at using a roster spot for a KO Specialist, but the ST has moved from consistantly on the bottom to number 5.

Now if the rest of the team can get back on track....

"Just say no" to yellow hankies.

69
by Marcumzilla :: Wed, 09/24/2008 - 10:50am

From the "It's about time" department - Millen's out.
http://sports.espn.go.com/nfl/news/story?id=3606294

89
by Anonymous (not verified) :: Thu, 09/25/2008 - 4:59pm

Wow. Dallas at number 7? what are yall on? How can you leave philly at number 3 when dallas beat them while beating themselves? ill match the boys against any team in the nfl. especially them damn ravens....

91
by Raiderjoe :: Thu, 09/25/2008 - 10:43pm

Now way should Raiders be 20th lace. Raiders almost ebat Bills in week 3, have dominating win over chiefs, and 1st game was rough but when you add it all up, Raiders should be 15 or 16. But who really cares anyway, season is 17 weeks (16 games) and Raiders have 13 more games to go in regular season, will move up charts, and make playoofs where they will battle with Cowboys and Chargers probably

92
by Raiderjoe :: Thu, 09/25/2008 - 10:47pm

sorry double post

93
by Raiderjoe :: Thu, 09/25/2008 - 10:52pm

sorry triple post

use space to write about week 4

Raiders likely ro beat Chargers good chance score is Raiders 23, chargers 17

Raiders goignt o run all over Chargers running attack is greta
Raiders much better team then Jets. No way Chargers score 48 points again (bu t good chance they give up 29 points or more points)

GaME BRINGS rAIDERS BACK TO .500 AND lOSS mAKES Chagers go to 1-3 and in bad shape.

96
by Jay Gibbons (not verified) :: Sat, 09/27/2008 - 11:35pm

Dave seems to not be adding up correctly.
.55 * 40% (NE's preseason dvoa) + .45 * -8.8% (NE's current voa) = 18.04
,but NE's Dave here is 9.4. What am I doing wrong? I am using these preseason projections - http://www.footballoutsiders.com/dvoa-ratings/2008-dvoa-projections

97
by Yaguar :: Sun, 09/28/2008 - 3:18am

The DAVE projection has since been revised. DAVE is adjusted for lineup changes.

The Patriots might not have had many lineup changes from the preseason projection, but they've had one very, very big one.

98
by Dana DeArmond (not verified) :: Fri, 11/20/2009 - 11:08am

Miami is going to the superbowl! You bet on it.