Writers of Pro Football Prospectus 2008

Most Recent FO Features

JohnsonCal09.jpg

» Scramble Over/Unders: the Norths

The league's northern divisions pose a number of meaty questions, such as: "Is the Bears' offense due for a repeat performance?" "Why do the Lions have such pronounced splits?" and "Has Johnny Manziel made the Cleveland brass even crazier?"

18 Oct 2005

Week 7 DVOA Ratings

by Aaron Schatz

Hi all. I took a long, long, long time to write today's DVOA commentary, so the power rankings are not up at FOXSports.com yet. I decided to post the ratings so you could all begin discussing, and when the commentary goes up we'll go ahead and link it.

(Update Tuesday 7pm: Here's the commentary.)
(Update Wednesday 1pm: Estimated Wins numbers now fixed below.)

Yes, Miami moves up despite losing, because of the increased strength of opponent adjustments. Yes, Miami is ahead of Denver, in large part due to the Week 1 upset. Both of these things are discussed in the commentary.

Individual pages for offense, defense, and special teams are now online. Offensive line and defensive line also updated.

* * * * *

Here are the Football Outsiders team efficiency ratings through Week 6 of 2005, 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 averaged based on situation in order to determine value over average. (Explained further here.)

OFFENSE and DEFENSE DVOA are adjusted based on strength of opponent as well as to consider all fumbles, kept or lost, as equal value. Opponent adjustments are currently set at 60% and will increase each week until they are full strength after Week 10. SPECIAL DVOA is adjusted for type of stadium (warm, cold, dome, Denver/Mexico City) and week of season. NON-ADJ TOTAL VOA does not include these adjustments.

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


TEAM
TOTAL
DVOA
LAST
WEEK
NON-ADJ
TOTAL VOA
W-L OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
DVOA
S.T.
RANK
1 CIN 44.3% 1 64.8% 5-1 27.8% 3 -13.7% 8 2.8% 10
2 IND 40.6% 3 52.2% 6-0 26.1% 4 -17.8% 4 -3.3% 26
3 SD 35.1% 5 26.2% 3-3 33.3% 2 2.4% 18 4.2% 7
4 NYG 33.0% 2 39.4% 3-2 16.8% 6 0.7% 17 16.9% 1
5 JAC 28.6% 7 19.3% 4-2 -2.6% 18 -26.7% 2 4.5% 6
6 SEA 25.1% 6 33.9% 4-2 36.4% 1 11.0% 25 -0.3% 19
7 PIT 21.6% 4 30.8% 3-2 7.5% 13 -14.9% 6 -0.7% 21
8 DAL 21.4% 9 26.1% 4-2 8.1% 11 -10.4% 10 2.9% 9
9 TB 18.1% 8 34.1% 5-1 -2.7% 19 -18.9% 3 1.9% 12
10 MIA 15.1% 12 6.0% 2-3 -3.6% 20 -16.4% 5 2.3% 11
11 PHI 9.7% 11 14.6% 3-2 15.9% 7 -6.8% 13 -13.0% 32
12 ATL 9.6% 10 19.2% 4-2 19.1% 5 11.2% 26 1.7% 13
13 DEN 7.6% 16 12.5% 5-1 9.7% 9 0.0% 16 -2.0% 23
14 BUF 6.0% 20 17.5% 3-3 -15.9% 25 -9.3% 12 12.6% 2
15 CHI 4.0% 13 6.1% 2-3 -28.3% 30 -27.3% 1 5.0% 5
16 WAS 3.3% 21 -15.9% 3-2 7.6% 12 -0.6% 14 -4.8% 30
TEAM
TOTAL
DVOA
LAST
WEEK
NON-ADJ
TOTAL VOA
W-L OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
DVOA
S.T.
RANK
17 CAR 3.0% 14 14.0% 4-2 -10.0% 23 -12.6% 9 0.3% 18
18 KC 0.8% 17 1.6% 3-2 7.3% 14 10.3% 23 3.9% 8
19 CLE -2.4% 15 -20.0% 2-3 2.9% 15 7.0% 22 1.6% 14
20 OAK -4.3% 18 -4.7% 1-4 9.2% 10 10.4% 24 -3.0% 25
21 NE -4.4% 19 -12.5% 3-3 15.0% 8 20.9% 28 1.6% 15
22 GB -8.2% 22 0.3% 1-4 1.5% 16 5.9% 21 -3.9% 28
23 TEN -14.7% 23 -12.7% 2-4 0.1% 17 22.0% 29 7.2% 3
24 BAL -15.6% 25 -10.0% 2-3 -23.0% 27 -10.1% 11 -2.6% 24
25 ARI -18.3% 24 -11.3% 1-4 -13.7% 24 5.3% 20 0.7% 17
26 DET -21.6% 26 -27.6% 2-3 -34.7% 31 -14.3% 7 -1.3% 22
27 NO -27.8% 30 -35.5% 2-4 -4.3% 21 14.3% 27 -9.3% 31
28 STL -29.7% 27 -28.0% 2-4 -4.7% 22 24.4% 31 -0.5% 20
29 NYJ -30.9% 28 -26.8% 2-4 -24.6% 28 2.6% 19 -3.8% 27
30 MIN -31.4% 29 -50.0% 1-4 -27.2% 29 -0.1% 15 -4.2% 29
31 HOU -56.0% 31 -78.9% 0-5 -22.7% 26 34.4% 32 1.1% 16
32 SF -69.0% 32 -76.4% 1-4 -52.7% 32 23.0% 30 6.7% 4

  • FOX RANK represents the FOXSports.com Power Ratings which are 80% 2005 DVOA, 5% 2005 pre-season projection, and 15% a special weighted DVOA for 2004 that includes the playoffs.
  • 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 least consistent (#1, highest variance) to most consistent (#32, smallest variance).


TEAM TOTAL
DVOA
W-L FOX

RANK
ESTIM.
WINS
RANK PAST
SCHED
RANK FUTURE
SCHED
RANK VARIANCE RANK
1 CIN 44.3% 5-1 2 4.8 2 -12.0% 30 2.7% 11 25.8% 14
2 IND 40.6% 6-0 1 5.5 1 -17.1% 32 0.5% 16 7.2% 30
3 SD 35.1% 3-3 3 4.8 3 12.5% 6 4.9% 7 10.4% 27
4 NYG 33.0% 3-2 5 4.4 5 -3.9% 22 -2.2% 23 27.7% 12
5 JAC 28.6% 4-2 6 4.2 6 18.1% 1 -23.6% 32 33.5% 9
6 SEA 25.1% 4-2 7 4.7 4 -10.4% 27 -10.4% 29 14.9% 23
7 PIT 21.6% 3-2 4 3.6 10 -2.3% 19 3.3% 10 35.8% 8
8 DAL 21.4% 4-2 9 3.9 8 1.3% 13 1.3% 14 31.5% 10
9 TB 18.1% 5-1 8 3.9 7 -11.8% 29 -9.7% 28 16.5% 19
10 MIA 15.1% 2-3 11 3.9 9 0.8% 15 -3.4% 25 13.1% 24
11 PHI 9.7% 3-2 10 3.2 15 -8.3% 26 9.6% 2 30.8% 11
12 ATL 9.6% 4-2 13 3.3 12 -3.8% 21 -2.7% 24 18.3% 17
13 DEN 7.6% 5-1 12 3.3 14 13.1% 3 5.1% 6 22.5% 16
14 BUF 6.0% 3-3 14 3.1 18 -12.0% 31 6.2% 4 42.3% 5
15 CHI 4.0% 2-3 19 2.7 21 -1.6% 18 -11.8% 31 50.9% 3
16 WAS 3.3% 3-2 17 3.1 19 11.8% 8 3.5% 9 1.1% 32
TEAM TOTAL
DVOA
W-L FOX

RANK
ESTIM.
WINS
RANK PAST
SCHED
RANK FUTURE
SCHED
RANK VARIANCE RANK
17 CAR 3.0% 4-2 15 3.3 13 -10.9% 28 -0.3% 17 11.4% 25
18 KC 0.8% 3-2 18 3.2 16 -2.9% 20 12.1% 1 5.1% 31
19 CLE -2.4% 2-3 21 3.5 11 13.0% 4 -1.1% 19 7.8% 29
20 OAK -4.3% 1-4 20 2.8 20 12.5% 5 5.5% 5 11.4% 26
21 NE -4.4% 3-3 16 3.2 17 12.1% 7 1.2% 15 15.2% 22
22 GB -8.2% 1-4 22 1.9 26 -5.1% 23 1.8% 12 35.8% 7
23 TEN -14.7% 2-4 24 2.2 23 0.9% 14 -1.2% 20 40.6% 6
24 BAL -15.6% 2-3 23 2.4 22 -5.8% 24 6.7% 3 18.2% 18
25 ARI -18.3% 1-4 25 1.6 29 -7.5% 25 -5.7% 27 16.1% 20
26 DET -21.6% 2-3 26 2.2 24 0.2% 16 -1.7% 21 59.1% 1
27 NO -27.8% 2-4 28 1.9 25 2.0% 12 -1.9% 22 57.6% 2
28 STL -29.7% 2-4 30 1.6 28 -0.6% 17 -11.4% 30 15.3% 21
29 NYJ -30.9% 2-4 27 1.6 30 8.8% 10 3.5% 8 9.0% 28
30 MIN -31.4% 1-4 29 1.9 27 9.6% 9 -4.2% 26 26.7% 13
31 HOU -56.0% 0-5 31 0.2 31 16.5% 2 -1.0% 18 23.6% 15
32 SF -69.0% 1-4 32 0.1 32 4.7% 11 1.7% 13 48.4% 4

Posted by: Aaron Schatz on 18 Oct 2005

210 comments, Last at 17 Dec 2005, 9:43pm by Bonnie Smith

Comments

1
by Becephalus (not verified) :: Tue, 10/18/2005 - 4:52pm

I think DEN fans will be pissed again...as weel as some Phi and ALT ones perhaps. Personally, I lvoe the ratings. They confirm my suspicions that JAC and SEA are as good as they look. These rankings make more sense than any others i have seen.

2
by Tom (not verified) :: Tue, 10/18/2005 - 5:18pm

What amazes me is that the Chargers defense is ranked so highly. I've watched every game this year and the defensive backs are horrible. Not only are they poor in coverage but they can't catch at all either. They had 2 sure interceptions that they dropped against the Raiders last week (and that wasn't the only game where that happens). I'm always shocked when teams run on the Chargers -- those are just lost plays to pick on Jammer!

3
by tunesmith (not verified) :: Tue, 10/18/2005 - 5:18pm

Yeah, I defended Denver's low ranking last week, but I'm not sure anymore. I thought their weak point was red zone offense, but didn't they do really good on that this week? Plus, they've beaten teams with good records.

What does it mean if a team consistently outperforms its DVOA ranking? Does that mean they are just uncommonly good at exploiting matchups?

4
by Edgar Loo (not verified) :: Tue, 10/18/2005 - 5:34pm

I don't understand DVOA Ratings
the more we approach Week 10 the more data we have to adjust better the Strength of the opposing team.

Unfortunately until this week.
i see the trend that if a Top ranking team plays a low Ranking team, then his stats goes Up.
but a good team plays a good team,
they probably stay the same or go down.
and you can see it next week in two games. Colts-Texans and bengals-Steelers.

5
by Pat (not verified) :: Tue, 10/18/2005 - 5:54pm

The opponent adjustments aren't at 100% yet. So if a +40% team plays a -40% team, they might play like an 80% team, but the correction pulls them down to a 64% team. That's still higher than the 40% team, so they'll go up.

You might think there's sufficient evidence to believe that NYG is good and STL is bad, but last year in week 7, everyone still thought the Giants were a good team.

It sorts itself out in the next few weeks. It's just the system "figuring out" the lay of the land slowly.

6
by Andrew (not verified) :: Tue, 10/18/2005 - 5:54pm

I see only 5 complete teams on the list (significant positive offense, significant negative defense): Cinci, Indy, Pittsburgh, Dallas, Philly. And Philly has big special teams problems.

If Philly resolves its special teams problems, it moves way up the rankings.

San Diego, Giants, Washignton, Denver, just need to discover a defense.

Miami, Jacksonville, and Tampa need to discover an offense.

Atlanta and Seattle are frauds without a real defense, and Seattle just took another hit losing Hamlin.

In the 15 years since the 6 seed playoff system started, only one team has started 3-3 and made it to the Super Bowl - the flukie 2001 Tuck Rule Patriots. Super Bowl contenders start 4-2, 5-1, or 6-0 and finish between 11-5 and 14-2. Again excepting the 2001 Patriots 11-5 teams since 1990 that make the Super Bowl don't win the Super Bowl, because 11-5 is a show that you really just weren't good enough. Teams since 1990 that go 15-1 don't even make it through the Championship game, since this appears to be an indicator of an excess of luck, or an excessively easy schedule, rather than of overwhelming dominance like the 84 49ers or 85 Bears. In my mind, this says San Diego isn't going to make it, Giants, Pittsburgh, and Philly must win this Sunday to stay on the right side of history and minimize their total losses.

7
by Pat (not verified) :: Tue, 10/18/2005 - 6:01pm

By the way, any Philadelphia fan who's upset with these rankings should be shot. Look at them more closely: a good fraction of the reason that Philly's ranked so poorly are its special teams, at -13.0%, worst in the league. Do I disagree? God, no! Special teams have been horrible - missed extra points, short kickoffs, blown coverage, and horrible returns.

Make the special teams normal, and poof, Philly's the #7 team in the country.

Make the special teams what they were last year, and Philly's right in the top 5.

Yes, Philly's special teams are struggling now. But Harbaugh's special teams have been in the top 5 ever since 2000. Does anyone really doubt that the guy can't turn it around?

8
by Pat (not verified) :: Tue, 10/18/2005 - 6:02pm

psst, andrew: I tried that last year. Didn't work so well. Had the Giants looking like a great team. keep it on the down-low. :)

9
by Drew (not verified) :: Tue, 10/18/2005 - 6:18pm

Does anyone really doubt that the guy can’t turn it around?

I don't doubt that he can't turn it around.

10
by S (not verified) :: Tue, 10/18/2005 - 6:37pm

What DVOA seems to be saying right now (to me):

Cinci and Indy have been the 2 best teams in the NFL through this point, but neither team has played anybody decent.

SF and Hou have been terrible and it doesn't look like like that will be changing anytime soon.

SD and Sea look pretty solid (how about the Seahawks with the #1 offense acording to DVOA!), but it's too early to say for sure.

The Jets pretty much stink, but they're not as bad as SF or Hou.

After that, it's a big question mark, so don't panic just yet.

11
by Aaron (not verified) :: Tue, 10/18/2005 - 6:39pm

Offense, defense, and special teams pages now online.

Pass offense ranks of San Diego opponents: DAL (5), DEN (14), NYG (10), NE (8), PIT (6), OAK (11)

San Diego pass defense DVOA: 2.5%
San Diego pass defense VOA: 22.9%

12
by Pat (not verified) :: Tue, 10/18/2005 - 6:48pm

Why isn't there that big a correction for Indianapolis's pass defense?

Pass offense of Indy's opponents:

-76.7%, -38.4%, -21.1%, 3.4%, 17.1%, 0.0%
(sf, hou, bal, jac, cle, stl)
And yet Indy's pass VOA is -34.9%, and the DVOA is -32.4%.

13
by putnamp (not verified) :: Tue, 10/18/2005 - 6:51pm

I'm absolutely blown away by the possibility that the Seahawks easy schedule could lead to home-field advantage for the playoffs. They're a different team at home, and if they can maintain home-field advantage, um.. I'm going to stop talking now.

14
by Aaron (not verified) :: Tue, 10/18/2005 - 6:56pm

Hmmm, good question Pat. I'll look at that for Friday's mailbag.

15
by Fnor (not verified) :: Tue, 10/18/2005 - 6:56pm

#12:
Did StL have a decently high pass offense voa up until this week? Because if it did and last week's number was used to calculate Indy's pass defense voa, that might have something to do with it.

Or not. Probably not. But a guess.

16
by Matthew Furtek (not verified) :: Tue, 10/18/2005 - 7:03pm

Last week Kansas City was ranked ahead of Washingon by 1 spot, 16 to 17. After beating Washington, they dropped to 18 and Washington went up to 16...

Kind've funny if you read them as "Team A is ranked higher, thus DVOA predicts Team A would win on a neutral field."

The past 2 games have made up for Washington's "lucky wins"... dominating in all facets except turnovers.

17
by princeton73 (not verified) :: Tue, 10/18/2005 - 7:04pm

DVOA best bets this week (simplest version,i.e. underdog has higher DVOA):

NO +3 Rams
GB pick Minn
SD +4 Phila
Buff+3 Oak
Tenn +3 Ariz

18
by SomeGuy (not verified) :: Tue, 10/18/2005 - 7:08pm

I won't even touch the Denver situation (too easy), but who can honestly think that Miami should be higher ranked than Philly?

19
by Robert (not verified) :: Tue, 10/18/2005 - 7:11pm

Why does Miami's week 1 upset of Denver (at home) hurt Denver so much but Denver's domination of Jacksonville on the road not matter so much?

I think the rankings need tweaking. It's not a matter of being upset at the rankings, but clearly if you need a bunch of time to write an explanation each week your system just isn't very good.

20
by EorrFU (not verified) :: Tue, 10/18/2005 - 7:11pm

I think the ratings make perfect sense so far.

DEN has been outplayed by some of their opponents in terms of DVOA. That is why their score is lowere than it seems.

The Skins offense is better than anyone would have imagined, but their defense has acctually gotten worse a creating turnovers this year and is the main reason their defense is ranked so low. This team could go any wich direction depending on luck alone because their weaknesses can be masked or enhanced by that luck. a true wildcard of a team. That crazy low varience makes me believe that they are consistant and Gibbs might not be so out of touch as some say. They are a tough matchup for teams like KC and DEN with good run offense and bead pass defense.

In fact, the whole NFC East is such a jumble I don' think anybody can state with any conviction who is coming out of it unscathed.

I really don't get Cincy. They are a confusing team who is obviously not the best in the league. They have that amazing pass offense but it all seems so illusionary to me. Yet, they are definitely better than the general punditocracy may believe.

Jacksonville has a real creampuff schedule coming up so watch the talking heads to praise them for really turning it around, when in fact they have some serious flaws.

Ultimitely, and I hate to say this, Indy looks like the class of the league. They have that nasty pass rush, and that offense finally figured out that you only need to score more than the other team. They are using Edge a lot more and are giving that defense a rest from time to time. They are the team to beat.

21
by SomeGuy (not verified) :: Tue, 10/18/2005 - 7:20pm

You just said the rating make perfect sense, then you went on to say that Indy is the class of the league (which they clearly are), however the rating say that Cincy is #1. This whole thing reminds me of the BCS, I'm ot even gonna get started on that.

22
by Becephalus (not verified) :: Tue, 10/18/2005 - 7:29pm

Yeah but unlike the BCS this isn't important. This is just some peoples attempt to understand what goes on in games and what makes a team good. And it seems to be getting better as time goes on. So what is the big deal? Rankings look pretty good. The rigid statistical framework punishes certain teams unfairly (as any rigid statisical framework you could create would). But that is the nature of the beast. Those comments about IND and SD pass DEF are nteresting....hmmm....

23
by Countertorque (not verified) :: Tue, 10/18/2005 - 7:29pm

Does the VOA after 6 weeks for a specific team simply equal the sum of the VOA after week 5 and the VOA for their game in week 6?

If not, how do I figure out my team's VOA for a specific week?

24
by Aaron (not verified) :: Tue, 10/18/2005 - 7:29pm

The commentary is taking a long time to write because

a) Last year I didn't have to write something about every single team.

b) I'm not just writing it for you folks anymore. I'm writing it for a lot of new readers who need to have statistical concepts like sample size explained to them.

As far as the rankings needing tweaking, this is what I say in the commentary that will be showing up on FOX:

a) I'm not going to toss out a system that works better than anything else available for 112 games from 1998 to 2004 just because it seems a little kooky for one team over a six-game period in 2005.

b) I'm not interested in "If Denver isn't ranked higher, something is wrong." I'm interested in "If Denver isn't ranked higher, something is wrong, and here are a couple of specific ideas about what it might be that could also improve the accuracy of DVOA for the league as a whole." Heck, I'll give you an example: when I eventually have a chance to create run/pass adjustments for weather, it might recognize that road teams have problems in Florida in September and adjust those teams slightly upwards.

25
by Fnor (not verified) :: Tue, 10/18/2005 - 7:46pm

Does my idea about determining strength of schedule based on pass d/o and run d/o comparisons instead of just overall dvoa count as specific or stupid? :)

26
by Justus (not verified) :: Tue, 10/18/2005 - 7:51pm

Allow me to disagree strongly with #19. Calculus takes a long time to explain but that doesn't mean there is anything wrong with it. DVOA may need improvements but that's a terrible line of argument to take. Plus, not everyone thinks the DVOA rankings need to be explained.

27
by Aaron (not verified) :: Tue, 10/18/2005 - 7:57pm

But gosh darn golly, I'll explain 'em anyway. Fox commentary now up (click link on my name). I'm off to go wash blueberries off my daughter's face.

28
by Robert (not verified) :: Tue, 10/18/2005 - 8:15pm

Great points on your response to me Aaron. I honestly don't have the time or the brainpower to suggest how they could be tweaked because I don't understand them (obviously) nearly as well as you.

One vague thought that I have that may have nothing to do with your system and may already be accounted for: something that bugs me about passing and passing statistics is quarterbacks that put up huge numbers simply because they are coming from behind.

Case in point-- Brady against the Broncos this past weekend. He wouldn't have had to heave the ball up 40+ times and thrown for 300+ yards had he not been down 28-3. Heck, I think Plummer threw for 499 yards last year in a bad Broncos loss simply because they were trying to come from behind. Brady's stat sheet looks great and more impressive then Plummer's, but I'd rather have my QB throw for 250+ yards in a week and put it up < 30 times any day.

What does this have to do with your rankings? I don't know, but maybe stats when the score differential is >=17 should be counted less than stats when it's a closer game. (Again, maybe you're already doing this.)

29
by Robert (not verified) :: Tue, 10/18/2005 - 8:20pm

My last post got truncated.

I said I'd rather have my QB put it up < 30 times / game and throw for 250 yards than have to put it up > 40 times and throw for over 300.
Then I suggested (if you are not already doing it) perhaps weighting stats more heavily when the score differential is <= 14 (or 17) than when it's greater than 14 (or 17). I also realize this might be unrealistic and overly-complicated, but it might be a more realistic gauge.

30
by Robert (not verified) :: Tue, 10/18/2005 - 8:24pm

Sorry, I can't use less than signs I guess... Grrr.. (HTML! The preview is coming out ok but the post gets truncated!)

For the last time (hopefully), I'd rather have my QB put it up less than 30 times than have to put it up over 40 times and have more impressive numbers.

Finally (for the last time hopefully) I suggested if you could weight the stats more heavily when the score differential was less than 14 (or 17) than when it was greater than that maybe that would be better. You may already be doing that. That may also be overcomplicated. I think it would make for a more accurate gauge, however

My apologies for 3 posts-- don't try to use "less than" unless you use "ampersandlt;" I guess.

31
by Tim Kirk, York(UK) (not verified) :: Tue, 10/18/2005 - 8:28pm

I'm still a Denver fan who isn't unhappy with their ranking... I still can't work out if they are good and relaxing to let people back in the game to have to hang on more than they should. Or riding some luck and big plays and clinging on depserately to win...

I guess the Giants away is a good chance to see - they could easily come out of it like Washington did after their game at Denver - a good loss might do more for their credibility than a another slightly dodgy win (though I'd prefer a win, however dodgy).

They do seem to switch to prevent style defence rather easily, and also the normally slightly too gambling offence (I like the PFP comments about Shanahan being a gambler of a coach, and this not being a good point - right back to the predictable 'go for the jugular' interception in the GB Superbowl when simply running for a FG would have been so much more sane) seems to be playing more sensibly this year, but getting too predictable whe holding a lead - the succession of 3&outs when holding a second half lead sound (I listen to radio) to me as though they are playing it very safe and conting on the defence not to leak points too quickly as long as the offence at least finishes drives with a decent punt...

Denver feel a better team this year than last year to me. But I'd like to be convinced that they can play for a whole game if they need to, and so far I'm not... but if they keep doing just enough to win the rating will climb (and Miami & Washington & NYG are all better than last year - so will their schedule will lok tougher in hindsight than it did at the start of the season).

32
by EorrFU (not verified) :: Tue, 10/18/2005 - 8:44pm

You answered your own question. Their ranking is based on what Happened and not why. Whether Denver was lucky or they relaxed in the second half means that they preformed differently. DVOA says WAS outplayed DEN yet still lost a close game. The point is wins and losses are too ephemeral to determine true ability. That is why over 16 games much of this stuff evens out. If Denver is good than future performances will improve their DVOA, but an argument can be made that the teams ranked higher have outplayed the Broncos on Average.

DVOA can't make value distinctions based on game situations. They played bad in all of their second halves, just not bad enough to lose. Therfore they get punished for bad play even though they didn't completely implode.

33
by Bj&ouml;rn (not verified) :: Tue, 10/18/2005 - 8:52pm

I am a Broncos fan who is perfectly satisfied with their rating. There's alot of things about Denver that I am very uncomfortable with right now.

34
by CaffeineMan (not verified) :: Tue, 10/18/2005 - 9:08pm

Sadly, I have to say I think NE's rating is dead on target. Duane Starks... I will now light myself on fire.

35
by doktarr (not verified) :: Tue, 10/18/2005 - 9:24pm

Re #21:

Cincy above Indy is due almost completely to Cincy having better special team's numbers. This, in turn, is due almost entirely to two things: Vanderjagt having an unusually inaccurate year, and Indy having a terrible return game.

Both of those things are definitely true through six games. It's not hard to see why the algorithm would think these things, because they are true.

If you think those things are flukes and we will see a rebound, then fine, Indy should be #1, but that's just your prediction. If, on the other hand, you think those things are real actual problems in Indy, then it's reasonable to imagine those weaknesses could cost the Colts a close game at some point. Early on, yesterday sure looked like that game, with the muffed return and the missed FG.

Re: Denver just "letting teams hang around" - great teams don't do that. They could have lost a couple of those games pretty easily. Great teams run away and hide.

36
by james (not verified) :: Tue, 10/18/2005 - 9:24pm

Hmmm Redskins offense now 12th in DVOA. I think someone's drive stat analysis pointed out that they were very close to being a top 10 team offensively and DVOA wasn't showing that yet back in the week 3 "redskins fraud" thread.

Just wondering if you some of you guys like apples?

37
by mactbone (not verified) :: Tue, 10/18/2005 - 9:30pm

Anyone bothered to look through the archives and find out what the lowest (ie easiest) future and past schedule DVOAs were? Right now JAX is almost doubling the next easiest schedule with -23.6% and I'm wondering if that's unusual, a one week aberration, or just having Houston, SF and some other teams that have looked weak so far.

By the time the Bears are finished, all those easy teams on the schedule will look a lot tougher after a win over Da Bears. I'm a Bears fan but 8-8 is the best I can expect.

38
by james (not verified) :: Tue, 10/18/2005 - 9:33pm

next question about to be answered by skins...

Can they cause turnovers? They have not taken the ball away in 4 games.
Evidence so far: inconclusive

Why?

3 of 5 games on the road.

4 of 5 opponents ranked in top third in giveaways.

Whats the redskins offensive production gonna look like when they dont have the league worst starting field position once they get to play some more mistake prone teams.

We know they can play tough on the road without getting any turnovers. What does this team look like when they figure out how to be better at taking the ball away?

39
by Bj&ouml;rn (not verified) :: Tue, 10/18/2005 - 9:43pm

Here is something I checked out in terms of power ranking quality. The only measure I can think of for quality if ranking is the number of places a team moves from week to week. I know that I have difficulty giving respect to a ranking system that bounces teams around like crazy. For instance, did you know that espn.com moved the Ravens 9 spots up last week. Anyways, check this out.

CBS Power Rankings - The 32 teams moved a combined 106 spots from last week (that counts moving up and moving down). CBS ranked teams averaged movement of 3.31 places last week.

NFL.com power rankings - Vic Carucci only ranks the top 12 teams, so you'd figure that would be easier to figure out, right? Only a little bit easier, apparently. His 12 teams move 28 spots, or 2.33 per team. The average could actually be greater, because he includes two previously unranked teams. The 2.33 per team assumes that these teams were ranked 13th and 14th. If you include the fact that 2 previously ranked teams fell out of the list, (to at least 13th and 14th again) the average goes up to 2.43 per team, or 34/14.

ESPN.com - The aforementioned blastoff by the Baltimore Ravens (from 30 to 21) helps the average espn-ranked team move 3.16 places. How can a team jump 9 spots for beating Cleveland? How???

Football Outsiders - You probably all know how this is going to turn out, right? Well, for the 32 NFL teams, the FO system moves everyone a total of 54 spots. That is an average of 1.69 per team.

What those numbers indicate to me is that, while not necessarialy more accurate, the FO rankings are far more consistent.

40
by calig23 (not verified) :: Tue, 10/18/2005 - 9:44pm

Re:#6

In the 15 years since the 6 seed playoff system started, only one team has started 3-3 and made it to the Super Bowl - the flukie 2001 Tuck Rule Patriots.

That is incorrect. The 1995 Steelers not only started 3-3, they actually fell to 3-4, before finishing 11-5 and reaching the Super Bowl.

41
by Fnor (not verified) :: Tue, 10/18/2005 - 9:47pm

re: james and "apples"
Your drive stats argument was ridiculous because I rebutted it with 3rd down stats (which were much improved in the last 3 games), and no one could make heads or tails of your argument.

I'm still not sold on the skins, in any case.

42
by thad (not verified) :: Tue, 10/18/2005 - 10:07pm

re some guy
Yes I absolutly believe the eagles are ranked too high. They pass too often, which leads to incompletions, which leads to getting zero yards on 25% of their plays so far. I do not believe their has been a NFL team in the last 10 years to make the playoffs and run lesss than 30 percent of the time. Its a hard way to live in October, wait till December. So, either they run more, a lot more, or they are in pretty big trouble.

43
by dedkrikit (not verified) :: Tue, 10/18/2005 - 10:32pm

By the end of the year you will be an expert on the 2005 Denver Broncos, Aaron. Now if only the rest of the teams can get as much attention! ;)

44
by James (not verified) :: Tue, 10/18/2005 - 10:53pm

re 41 Fnor
my drive stats argument was not the greatest argument of all time but it was sure better than the conventional "52% completion pct. can't be continued, so the Skins offense sucks".

While my drive stats argument correctly stated that the Skins DVOA would improve your 3rd argument incorrectly stated that the offense would stay negative.

My argument is supported by the skins offense being ranked 12th DVOA after having a negative DVOA after the first three games. I believe I predicted DVOA would start to show a ranking of around 10.

My point wasn't to sell you on the redskins but to point out a subset of information that sometimes predicts a little better than DVOA.

Most on the board were stating the Redskins offense would prove to be worthless. With 444 yards against Denver and 432 yards against KC on the road, I think it's time you admit that Gibbs and co. are at the very least an "unfraudulent" offense.

45
by Fnor (not verified) :: Tue, 10/18/2005 - 11:02pm

James: the point is that you were searching for anything that would indicate your presupposed idea that the skins offense was great without really having any idea of what they defense was doing. Witness when you said, paraphrasing "getting long and short," which wasn't just not true, but almost entirely the opposite of what was actually happening (3rd and 10+ as often as 3rd and 5-).

Plus, drive stats INCLUDE this bizzare 3rd down performance, which was why I brought it up.

So, we have two different situations, here. I had a stat that was out of line with what was normal. You had a stat, which was affected by the one I had, saying the offense was great. And now Washington has shown that they can play offense without bizzare 3rd down performance. That doesn't mean that what you said was more predictive than what I said. It's essentially like calling a coin toss.

46
by Israel (not verified) :: Tue, 10/18/2005 - 11:03pm

When we talk about Pittsburgh's stats for the rest of the season, we're going to have to give two numbers: full-season, and "not counting the Jacksonville game."

Not just "for the rest of the season." More like for eternity. And I needn't add "if Ben stays healthy" because Batch would be adequate if necessary.

47
by Fnor (not verified) :: Tue, 10/18/2005 - 11:17pm

ugh, another typo. "Getting third and short" in my paraphrasing.

48
by Aaron (not verified) :: Tue, 10/18/2005 - 11:19pm

Offensive and Defensive Line pages now updated. Oh, and Denver fans, if you want to e-mail me, the contraction "they are" as in "your ratings, they are crap" is spelled they're, not their. Thanks.

49
by james (not verified) :: Tue, 10/18/2005 - 11:24pm

My stats did not say the offense was great. It said it was moving the ball better than most teams and failing to score.

I've been using drive stats for two years. I tried to share what they usually show me and was flamed beyond compare.

Now the skins offensive performance was a coin toss? Before the benefit of hindsight noone but me was willing to go out on a limb and actually state with some evidence that skins offense would actually improve. It had nothing to do with fandom and everything to do with evidence that I have seen show its face before. If you want to see the next example of DVOA about to change dramatically with drive stats showing it earlier you only need to look at the New York Giants.

I put out the evidence, told everyone they would see I was right and to wait, and then the results came in. I was right. Everyone else was wrong.

You may not like drive stats but they helped me see what noone else could.

Look forward to arguing with you more. At least you cause me to check my work and bring up good points to the dispute.

50
by Fnor (not verified) :: Tue, 10/18/2005 - 11:25pm

Pointing out a weakness in your argument isn't flaming you. If I recall, you pissed Pat off to the point that he stopped responding. Thanks for the snark, though.

51
by Bj&ouml;rn (not verified) :: Tue, 10/18/2005 - 11:46pm

Children, calm down.

52
by Fnor (not verified) :: Tue, 10/18/2005 - 11:49pm

Sorry Bjorn :(.
Does this mean no snack time...?

53
by james (not verified) :: Tue, 10/18/2005 - 11:55pm

Fnor,

As I have stated before I do not intend for anything I write to be taken personal. The same way that Carl can argue about TMQ's points being wrong with vigor and still show respect is the same way that you and I can argue with respect. I was basically being called dumb and am very glad to be able to throw it all back in some of those faces now. I don't include you in that group and stated my respect for the way you presented your arguments in my last post to this thread.

I will be wrong at some point but it won't be after failing to even try to understand someone else's argument and then call them stupid because I don't understand what they are saying. Some may be offended but that's exactly what alot of posters were doing. I found that to be ignorant by alot of people, again that group doesn't include you.

Yet, noone is saying maybe the crackpot is on to something. Instead I get it was a lucky prediction. I guess I expected more from FO posters. Again, I've always been very quick to admit when I'm wrong. Hopefully, noone will get to see that trait becasue I would rather not be wrong.

thanks for reading and I look forward to corresponding with you more

54
by Andrew (not verified) :: Tue, 10/18/2005 - 11:57pm

calig23 re #40:

Thanks for the correction. I miswrote what I intended. Only one team has gone 3-3 and won the Super Bowl since 1990 - the 2001 Patriots. The 1995 Steelers and 1996 Patriots both went 3-3 and made the Super Bowl at 11-5, and of course lost, like every other 11-5 team has besides the 2001 Patriots.

Among winners, 1 went 3-3, 3 went 4-2, 5 went 5-1, 6 went 6-0. Among Super Bowl Losers, 2 went 3-3, 4 went 4-2, 6 went 5-1, 3 went 6-0. The typical Super Bowl winner has been 13-3. The typical loser has been 12-4. The divisional playoffs and championship playoffs are filled with the carcasses of frauds that started off 3-3 and somehow managed to turn it around and get a winning record.

Similarly, teams that end up 10-6 or 9-7 almost always lose in the divisional round. 93 49ers, 95 Colts and 96 Jaguars are the only exceptions.

55
by Pat (not verified) :: Wed, 10/19/2005 - 1:02am

By the way, a previous long comment regarding Vanderjagt got eaten.

This is not an uncommonly bad year for Vanderjagt in terms of accuracy. Look at the special teams for Indianapolis for the past years:

(field goals/extra points only)
2000: 9.7
2001: 6.1
2002: -3.9
2003: 18.1
2004: -3.6
2005: -0.6

He's not a consistent kicker by any stretch of the imagination. He just managed to have one really good string.

If you prefer conventional statistics, just look at Vanderjagt's year by year statistics side by side with Vinatieri or Akers. Both of their FG percentages barely fluctuate at all, whereas Vanderjagt's are all over the map.

And anyway, Vanderjagt's only missed one kick his year so far, so it's not like he's hurting them at all. Except by taking up a roster spot and forcing them to carry three kickers on a team.

56
by RowdyRoddyPiper (not verified) :: Wed, 10/19/2005 - 1:15am

I watched every ball crushing moment of both PSU/MICH and PIT/JAX. I've just sold a kidney to pay my therapist but I feel much better.

I can't really believe that the Steelers didn't fall further. Maddox is taking a lot of heat, as he should (blaming the best Mennonite in pro football doesn't help his case), but the Steelers couldn't get a damn thing moving on the ground. Getting gashed on the other side of the ball by the third or fourth string RB didn't really inspire too much confidence. Hopefully they get it together, I know Cincy was living off of picks early on.

Aaron, as you are busy fending off pissed off Bronco fans you say,

" because of their great running game, Denver faces an average of just 6.2 yards to go on third downs, the third-lowest number in the NFL."

3rd and 6.2 seems pretty high, even for a guy who saw a lot of 3rd and 11.5 on Sunday. Is that a type-o?? Have I totally missed something?

For Washington to start creating turnovers they've gotta uncork LaVar, having seen the leap in person, you just can't have a guy like that on the sideline (okay I may be biased).

57
by bobman (not verified) :: Wed, 10/19/2005 - 1:59am

RE 35 and 55:
Stop slamming Vanderjagt already; he's missed one kick so far (Monday night). Boo-hoo. Try to keep the persona and the FG/EP kicker separate, please. I noticed Saint Vinatieri missed an XP earlier this season--an XP for Pete's sake, from the man who has won every important game in the final seconds since Knute Rockne retired. (Vandy's kickoffs, alas, are truly subpar, but as a team they have addressed that, so it's out of Vandy's hands, er, feet for now and the stats should reflect KOs that are five yards deeper on average than Vandy's efforts the past few years, at least the ones I've seen in 3 of 6 games this season)

The Colts' return game "sucks" because they are playing like the offense--ball-control and cautious. Many more fair catches than I remember seeing in the past and I suspect it comes straight from the top. Also, Brad Pyatt, who was leading the league as a KR halfway through his rookie year before getting injured (remember his long return that helped turn around the MNF game against the Bucs?), and then got hurt again last year, and then again this year (sense a trend?) hasn't been there. If he ever makes it back, their return numbers will improve.

So the kicking game is fine; Hunter Smith would rather have a punt land anywhere inside the 20 than a TB, regardless of what it does to his average, and the return game is playing it safe.

Where they are actually improved over previous years is coverage--I'm stunned to see them tackling like they do, and no big returns against them so far (that is none that weren't reversed due to penalties).

So statistically, their Special Teams are not so special but they are doing what is needed. I would not be at all surprised to see improvement in the ST ranking by the end of the season.

58
by NF (not verified) :: Wed, 10/19/2005 - 2:00am

Looking at the Fox Sports article, I was intrigued and decided to check out the 2003 Week 6 DVOA data.

Basically, take all the predictive stuff that worked for 2004, and predict the opposite. Baltimore went from worst-rated team with a winning record, to a top 5 team that made the playoffs. 12th-ranked New York Jets with a 1-4 record ended up with a 6-10 record and negative DVOA.

Bizarre.

59
by Kibbles (not verified) :: Wed, 10/19/2005 - 3:14am

Here's another Denver fan right here who has absolutely no problem with the current ranking. Why? Because they're 5-1 against the 3rd hardest schedule in the NFL (according to DVOA). I'll take that any day of the week.

If Tags announces that he's going to start factoring DVOA rank in when deciding who gets home field advantage in the playoffs, then I'll be a little bit upset that Denver's getting "shafted". In the meantime, I choose to be thankful that we're 5-1 after a positively BRUTAL stretch of games.

60
by Bruce Dickinson (not verified) :: Wed, 10/19/2005 - 3:31am

didn't the 2003 Panthers go 11-5 and only lose the Super Bowl by a field goal? i know they're a bit of an anomaly around here, but that's what they did.

61
by Sean (not verified) :: Wed, 10/19/2005 - 4:01am

It's interesting to see where some teams are vis-a-vis what their predicted Pythagorean production suggested. Jacksonville was singled out as a team that wasn't as good as their record and would likely fall off, but instead they are sitting at #5 in the rankings. This despite not making any substantial personnel additions (although Matt Jones has in fact contributed much more than most people would have expected). Tampa Bay was expected to make a jump based on their Pythagorean and DVOA stats, and finally they seem to have, although it remains to be seen whether or not they can keep up the pace with Chris Simms under center. Atlanta was expected to be exposed as a fraud, and the jury is still out on them. They are winning games, but they certainly don't look like a Super Bowl contender, and with the competition from Carolina, Tampa, and all the NFC East teams, it's very likely they miss the playoffs altogether. And Philly, who was expected to roll through everyone and easily hit 12-4, is struggling thanks to a collapse in the one area where they've been most consistent the last five years, namely their special teams. They're already halfway to four losses, and they'll be one game closer after they get rolled on Sunday.

62
by james (not verified) :: Wed, 10/19/2005 - 4:03am

Sean,
San Diego that good?

I think they are the most overrated team out there.

Be interesting to hear why SD is so good.

I don't have them as being more than 8-8 team.

63
by 2 cents (not verified) :: Wed, 10/19/2005 - 4:32am

I would think rankings should represent teams in a way that a higher placed team should beat a lower placed team more often than lose to it when playing in a nuetral field. However, just because team A beats team B, it is not necassary that A has to be placed higher than B. For ex: houston may beat colts next week but that wouldnt place houston higher than colts. Keyword: UPSET. Like last years 29-28 Mia win over Pat.

Having said all that, I cannot see how the DVOA presents a true picture of the NFL teams at all. A lot of folks have pointed out specifics. But intuitively (or common sense wise), having watched most of the teams, the DVOA doesnt seem to get a hang of the true picture.

My ranking if teams played next week on a nuetral field (disregarding bye week):

Broncos
Colts*
Chargers
Bengals
Philly
Pats*
Bucs
Jags
Seahawks
Steelers
Falcons
Redskins

As far as the colts-pats are concerned, the colts should win hands down on PAPER. But they have a huge mental mountain to climb. So if they were to play next week with current teams, its anybody's game.

And last, phylli will beat falcons on a nuetral field.

64
by Sean (not verified) :: Wed, 10/19/2005 - 4:51am

Well, they've got the #3 DVOA ranking in the league through six games, and I don't think that is particularly inaccurate. They have an offense that is very difficult for most teams to match up with. They have top five players on offense at quarterback, running back and tight end according to DVOA. Those positions work together to make for a very efficient offense that is able to consistently generate favorable down and distance situations. Their defense hasn't played as well as they did last year, but they are actually better personnel-wise, and they have talent that is equal or better to the other offensive powerhouse teams like Cincy and Indy.

I would be very worried about San Diego if I was a Raiders or a Broncos fan.

65
by tunesmith (not verified) :: Wed, 10/19/2005 - 5:52am

#62... however, San Diego will be the BEST 8-8 team in years...

66
by James Gibson (not verified) :: Wed, 10/19/2005 - 8:52am

Random comments on a bunch of these:

I don't think San Diego is overrated at all. You're talking about a team that went to Foxboro and thrashed New England and also destroyed the N.Y. Giants at home. They lost a narrow game to a good Pittsburgh team and a narrow game at Denver (a place they lost last year when they had a good record as well). The only loss looking not that good is the opener to Dallas, but 1) Dallas is somehwat good this year and 2) they were playing that game without Antonio Gates. I'd be really surprised if they finished 8-8.

I think Washington has done what was expected of them in that article - lose. It's funny to start saying "I told you so" after the team has gone from 3-0 to 3-2 when the article basically was suggesting that Washingon is about a 9-7 team and not a 12-4 team.

As for drive ratings being better than DVOA, I don't think so. First, I said during that thread, you'd need D-Drive ratings for a comparison. Second, if you look at the chart on page 4 in Pro Football Prospectus, you see why VOA and drive stats won't have a linear correlation even if you added a starting field position to the drive stats. The non-linearity inside your own 20 and then again inside the opponents' 15 seems to be what really diverges VOA from drive stats.

What's not clear to me is how well the defense adjustment works after game 3. Clearly, the more connections between teams the better the D on DVOA works. After week 3, there just may not be enough connections even if all teams are connected.

I watched every minute of the Penn State game, too. I was so mad afterwards I refused to watch the end of the USC-Michigan game. If we had DVOA in college, I wonder where Penn State's special teams would be. That was brutal. Fortunately, I'm not also a Steelers fan.

67
by James Gibson (not verified) :: Wed, 10/19/2005 - 9:10am

One comparison that might be interesting is DVOA compared to a different power rating, such as Sagarin's. I think the one that would be appropriate for comparison would be pure points, as opposed to his joint or ELO-Chess ratings which are more W/L based. My list, based on above with Sagarin Pure Points ranking in parentheses:

1. Cincinnati (2)
2. Indianapolis (3)
3. San Diego (1)
4. N.Y. Giants (12)
5. Jacksonville (4)
6. Seattle (8)
7. Pittsburgh (5)
8. Dallas (17)
9. Tampa Bay (15)
10.Miami (24)
11.Philadelphia (10)
12.Atlanta (16)
13.Denver (7)
14.Buffalo (27)
15.Chicago (6)
16.Washington (14)
17.Carolina (21)
18.Kansas City (9)
19.Cleveland (13)
20.Oakland (20)
21.New England (11)
22.Green Bay (19)
23.Tennessee (22)
24.Baltimore (28)
25.Arizona (30)
26.Detroit (18)
27.New Orelans (26)
28.St. Louis (23)
29.N.Y. Jets (25)
30.Minnesota (29)
31.Houston (32)
32.San Francisco (31)

I'm guessing the primary difference between DOVA and this rating is the actual reason that the teams score the points which is what VOA works at determining - that is VOA likes sustained drives more than single big plays, where as a system like Sagarin wouldn't differentiate.

68
by karl (not verified) :: Wed, 10/19/2005 - 9:24am

I love the comment in the FoxSports commentary that says:

"I'll be happy to respond to any e-mails that use correct spelling, grammar, and capitalization rules and ask about the methodology in general rather than accusing me of "disrespect" or other such nonsense."

The ignorant garbage emails you must be receiving would drive me nuts.

69
by Aaron (not verified) :: Wed, 10/19/2005 - 9:51am

Re: Week 6 of 2003, if I remember correctly, 2003 has the worst correlation between DVOA and wins of any year from 1998-2004. Unfortunately, you can't go back to Week 6 of 2002, etc., because the website didn't exist then.

If I get one more e-mail accusing me of trying to copy the BCS, I'm going to have to hurt somebody. Does anybody out there think that the problem with the BCS is the computer rankings? The problem with the BCS is that if three teams are on top, ONE OF THEM GETS SCREWED. This doesn't happen in the NFL, because the NFL does not depend on my ratings to decide who plays in the Super Bowl. They have playoffs where the teams have to actually face each other.

Karl, you wouldn't know the half of it.

70
by Oz (not verified) :: Wed, 10/19/2005 - 10:02am

Re: #12

The Colts haven't played Houston yet, but they have played Tennessee.

So replacing Houston's -38.4% value with the Tenn 8.3% pass offense rating probably explains the negligible difference in Indy's pass defense VOA and DVOA.

71
by james (not verified) :: Wed, 10/19/2005 - 10:09am

james gibson,
I'm only saying I told you so about the DVOA of the Redskins offense. Nothing else. They've now won 2 games they should have lost and lost two games they should have won. To me, that's exactly the way it should be. I'm only pointing out that the Redskins offense moves the ball up and down the field with ease just like "someone" said they would continue to be.

72
by james (not verified) :: Wed, 10/19/2005 - 10:15am

re also no. 66

The redskins had a negative dvoa for the first three games and the 2nd best 3rd down percentage. 99% of the posters on that thread said they couldn't keep up this "luck" and that their dvoa would stay negative.

Never said that drive stats were "better" than DVOA. I said that when they paint two different pictures sometimes DVOA will start to look more like Drive stats rankings.

Now I think DVOA more clearly illustrates what the skins can do on offense now. Thats what I said would happen. I was laughed at. Now they are 12 DVOA. Right around the 10 I said they would end up at. They have SF this week.

73
by james (not verified) :: Wed, 10/19/2005 - 10:52am

re san diego,

I think too much is going into their win against NE. It was totally dominating. However, it goes against what they have done the rest of the season. If they controlled the ball for a 2 to 1 advantage in any other game I would believe in them alot more.(not that it matter to anyone if I believe in them)

At the beginning of the year I wrote somewhere that San Diego's would be this year's last year's KC(mouthful). Some dominating offensive performance but just don't have the defense to do it consistently. KC had some very impressive wins last year but ultimately couldn't win with just offense two years in a row.

74
by James, London (not verified) :: Wed, 10/19/2005 - 10:55am

Aaron,

Why not give the best/worst of the "you suck"/"you're ratings are crap" emails "the're" very own thread?

A wall of shame if you will. Gives the rest of us a cheap laugh and may just deter the terminally angry from posting. It might also provide therapy for your soul...

75
by Pat (not verified) :: Wed, 10/19/2005 - 11:00am

Re: the BCS computer rankings:

I think that some people out there do think the problem is the computer rankings. Which, to me, is scary - the computers are used because of one very, very important fact - they aren't biased. They don't care if you're a storied program and you had a Heisman candidate at quarterback at the start of the season. If you're 11-1 at the end of the season, and all 11 wins came against teams with losing records, and the 1 loss came against a team with a winning record, it's going to say you're not so good.

For those who haven't seen it, Wes Colley has a very detailed explanation of how the rankings work here.

76
by Ted (not verified) :: Wed, 10/19/2005 - 11:05am

Aaron, I thought 2001 had the lowest correlation between wins and DVOA. The Pats winning the Super Bowl (#16 in DVOA), Tampa finishing third in DVOA with a 9-7 record, San Diego finishing ninth with a 5-11 record ahead of the 13-3 Bears. Some strange stuff happened that year.

77
by mawbrew (not verified) :: Wed, 10/19/2005 - 11:06am

Until last week, I still thought the Vikings would win the NFC North. They've now convinced me that they won't.

Since the development of DVOA, what is the record for lowest DVOA by a division winner? Whatever it is, I think it will fall this year.

78
by james (not verified) :: Wed, 10/19/2005 - 11:15am

Mawbrew,
The Bears have a positive DVOA and will be playing a weak schedule(31st future schedule rank).

I think they have the best chance to win the Central because of their defense and with such a weak future schedule should at least stay the same.

79
by Karl (not verified) :: Wed, 10/19/2005 - 11:21am

Aaron, I think the problem with the people sending nasty emails, is that they do not read. The average person doesn't like to read in general. You've done a great job explaining your methodologies here and on FoxSports, but i'm sure that there's a good majority of people that will simply scroll down to find their team and decide "this is wrong". They won't bother reading your explanations. They won't bother trying to understand why. It's all there in plain view however.

I think the more public attention FO receives (and I hope it's much more), the more you'll have to deal with enemies of progressive football analysis. It must be frustrating. Unfortunately, the world is filled with stubborn innumerates and narrow-minded know-it-alls.

80
by wrmjr (not verified) :: Wed, 10/19/2005 - 11:23am

I'm relatively new here, so maybe someone can explain this. Aaron notes that for approximately the first half of the season, some games from last year are used to determine the rankings. Are any adjustments made for loss and/or gain of new personnel due to free agency? I could see that skewing the numbers (though not in Denver's case). Could that explain some of the variance in New England and Chicago's rank vs. 2005 only?

81
by James Gibson (not verified) :: Wed, 10/19/2005 - 11:32am

Re: San Diego. It's not just their win over New England. They also pummeled the Giants, who currently rank #4 in DVOA, and their losses are the opposite of Washington's wins that sparked the Washington article from a few weeks ago - they were all close and could have gone the other way. The Chargers are only a few plays away from being 6-0. I would note that the Chargers started last year at 3-3 as well and then finished 11-5.

82
by Aaron (not verified) :: Wed, 10/19/2005 - 11:36am

Re: 76, you are correct, it was 2001 that was the worst year for correlation between DVOA and wins. Not 2003. My mistake.

Re: 80. The advantage of running a table of numbers rather than just a single list is that I can offer more than one number for each team. If people prefer the idea of only looking at 2005 data, look in that column, or the numbers above marked "DVOA." If you prefer the idea of considering the previous year in order to eliminate some of the power of early fluke wins, look in the first column of the FOX ratings, or the column in the table above marked "FOX RANK."

The method that includes 2004 numbers does not make adjustments for free agency. Read more in the link on my name. The plan was to do a few days of work to make the early-season ratings even more accurate before I started doing them for FOX, but fate stepped in. I'm not quite sure how to say to all the new readers at FOX, "These numbers aren't going to be as accurate as we would like early in the year, I could not improve the method because my father died and I had to take a couple weeks off from work."

83
by Pat (not verified) :: Wed, 10/19/2005 - 11:41am

#80:

In previous weeks, there was a component which was a projection from 2004 which included gains/losses in free agency. Now it's just 2004 performance (20%) and 2005 performance (80%).

This makes sense - early on, the 2005 performance isn't solid enough to say "this team is bad" or "this team is good", and so the 2005 projection tries to shore up the 2005 performance. But by now, the 2005 performance is pretty solid - the only reason you keep the 2004 performance stuff in there is to give a little bit of inertia to the rankings. If a team was good in 2004, you're much more confident if they've looked good in 2005 that they are good. If a team sucked in 2004, at this point you're still wondering if they're an aberration. Like the Redskins.

I also think it helps to add weight to the teams which finish stronger than others (due to good coaching/better roster depth).

84
by B (not verified) :: Wed, 10/19/2005 - 11:43am

Re #12 & 70: That's the second time Pat's confused Houston and Tennessee, which is understandable, cause the Titans used to play in Tennessee. Pat, I want you to write this down:
"The team formerly known as the Houston Oilers now play in Tennessee as the Tennesse Titans. The new team in Houston is the Houston Texans."

85
by james (not verified) :: Wed, 10/19/2005 - 12:07pm

re 81 SD

SD is gonna score 30 plus alot this season. Their problem is gonna be the 8 games when they can't do that the defense is not gonna be their to help them.

There is a strong correlation between SD and KC of the past few years.

no defense

top qb
top rb
top te

The defense still gave up over 20 points to New York Giants. I crude measure I use is games giving up under 20 points. I team needs a defense capable of doing so consistently to compliment a great offense. Because a good defense makes a good offense average. Pitt giving up 22 points on the road against SD is an example of a great offense becoming average.

NYG being 4 for DVOA will not last. That was not an impressive win for the Chargers. I predict they will be alternating wins and losses all season. Unlike Indy of the past few years who had no defense, SD is not dynamic enough to put up big numbers against good defenses.

86
by Pat (not verified) :: Wed, 10/19/2005 - 12:19pm

#84:

DANGIT!

87
by MJK (not verified) :: Wed, 10/19/2005 - 12:28pm

Re: the BCS computer rankings:

Pat,

I agree that some people irrationally have a problem with computer ratings (hence some of the hate mail that Aaron gets), and that that is scary. I'm actually getting pretty interested in methods of rating teams based only on W-L that factors in strength of schedule. Thanks for the link about Colley's model--I had never really understood it before (I had been under the impression that it was a maximum liklihood model, which I brought up with you once before). I had been playing with my own maximum liklihood methods for rankings in the NFL, but this gives me something new to think about. Question: Colley's model neglects non-division 1A opponents, right? Do you see this as a deficiency? I was reading about another article (it's linked from my site) that addresses how to factor this in...

88
by Karl (not verified) :: Wed, 10/19/2005 - 12:31pm

RE: #77

The lowest DVOA of any division winner in FO memory (1998-2004) was the 21st ranked Minnesota Vikings of 2000. They finished the season with an 11-5 recorded despite a -5.4% DVOA. Their Estimated Wins for the season was 8.6 however (Went 5-11 in 2001).

89
by Independent George (not verified) :: Wed, 10/19/2005 - 12:32pm

#84 & #86:

But I thought the Texans played in Dallas?

90
by Jamie T. (not verified) :: Wed, 10/19/2005 - 12:41pm

No, No.

Dallas plays in San Antonio, Houston plays in Austin.

91
by mawbrew (not verified) :: Wed, 10/19/2005 - 12:54pm

Re: 78

I thought one of the beauties of DVOA is that a weak schedule is 'unmasked'. That is, strong performance against weak teams doesn't help that much. In fact, Chicago's two wins will actually begin to be worth less if Minn and Detroit continue to play poorly.

I tend to agree that Chicago has the best chance of winning the division, I just don't think they are in the top half of the league (overall) and I suspect DVOA will ultimately reflect that.

Re: 88

Wow. That's really surprising. Thanks for the info. The Bears may be able to beat that standard after all.

92
by j austin (not verified) :: Wed, 10/19/2005 - 1:08pm

ok first off power rankings should not be based on your little dvoa stats. that is rediculous. games are not won by stats they are won by the better team. anyone who has watched the patriots over the past 4 seasons knows that. You dont have to win between the 20 yard lines you have to win by putting more points on the board than the other team. and so far denver has done that to some of the best teams in the nfl. why aren't wins over qauility opponents factored in. how many teams can say they beat 5 straight .500 or better teams. not even the colts. 3-3 san diego chargers who are ranked #3 in the power rankings, the 3-2 chiefs who were predicted to win the afc west with ease. the 4-2 jax, jaguars on the road, the 3-2 washington redskins with a top notch defense that held the high flying chiefs to 6 first half points and 28 overall. and last but not least the 3-3 defending super bowl champ new england patriots who no matter how battered up always figure out how to get it done. who else do we have to beat. the 3-2 giants or the 3-2 philly eagles. that would be 7 straight .500 or above opponents. in this day and age of the nfl most top teams dont even play that many .500 or above teams in a row. look at indy feasting on the division cellars. All i'm saying is we may not be the best statistically but that shouldn't matter if you know how to get it done. and we're still #2 in the league in rushing and plummer is playing mistake free. when champ bailey is healthy the passing defense will improve greatly. and how do people keep saying we need to find a defense. hello 88 yards rushing a game allowed. that's better than the steel curtain and 4th overall in the league. that's all till at least next week. signed frustrated denver fan.

93
by Pat (not verified) :: Wed, 10/19/2005 - 1:38pm

MJK:

No - that's a strength, not a deficiency, in my opinion. You've probably heard something like this already, but...

Think of college football as a set of islands, connected by bridges. If you drew lines between teams that had played each other, that's what it would look like. Each conference is a tight, tight web, because they "basically" all play each other. By the end of the season, the Big Ten will have played 88 games between 11 opponents. The distance between any one team in the Big Ten and any other team in the Big Ten is maximally "two" (where distance is the number of hops to get to a team - so Iowa is two hops from Penn State because you have to go through Ohio State) by the end of the season.

So if you tried to do a ranking of the Big Ten by the end of the season, it'd be easy, because they're so interconnected.

But when you try to compare those internal Big Ten rankings to other conferences, those rankings suddenly become predicated on weak links. The total number of links from the Big Ten to the ACC, for instance, is one - Wisconsin vs. UNC.

This the one big problem with using rankings like this in both college football and the NFL (AFC teams are separated from all divisions save 1 in the NFC by at least a distance of 2, and are separated from 9 teams by a distance of 3), and it gets a lot worse when you add in non-Division IAA teams.

Why is this a problem? There's a statistical way to estimate the error in a calculation - remove one game, recalculate the rankings. If you do this for all games, you can estimate the error in the calculation. But you can also use this to see how dependent your rankings are on one game.

If you do that in college football, the interconference games hugely influence games. Removing the Texas-OSU game, for instance, changes almost all other teams in the Big Ten by 0.01 - enough to move a team 3 places in the poll. It moves teams not in the Big Ten by something like 0.0004. Virtually nothing.

On the other hand, removing the Minnesota-Michigan game moves other teams in the Big Ten by 0.002 or so. So you can see how huge that Ohio State-Texas game is.

One thing I'd love to see the computer rankings do is list which games not played by the team in question are giving the largest weight to the ranking. If you did that, you'd see that the non-Division IA games are usually the games that influence the rankings the most.

What would be really cool is showing rankings determined by only interconference play, and then showing how those conferences "line up" against each other.

Wow, that was really long and technical

94
by Pat (not verified) :: Wed, 10/19/2005 - 1:43pm

#92: Not sure if you're aware of this, but sentences are supposed to start with capital letters. Thanks!

95
by Fnor (not verified) :: Wed, 10/19/2005 - 1:59pm

#92:
You're talking about rushing yards. That's a stat. Everyone uses stats, it's just that Aaron binds himself to his instead of just using them as proof for whatever feeling he's getting on team X.
So, okay, let's do power rankings based on wins. Let's start with everyone's favorite trifecta: San Diego, Pittsburgh and New England.
So, NE beat PIT, so they should be higher. PIT beat SD, so they should be higher. But wait! SD beat NE. There's no ordering that satisfies those requirements!
whoops!
As for your argument about Denver, I could make an argument for PIT being higher because their stats include something like the 5th-worst QB performance in the history of DVOA. But that game was still played! And it counts! As does Denver's loss to Miami.
People need to learn to cheer for their team, but realize that just because they think their team is the best team ever doesn't mean it actually is.

96
by Karl (not verified) :: Wed, 10/19/2005 - 2:03pm

RE: #92

J Austin, I only wanted to address the first few sentences of your post:

Other than the misspelling of ridiculous, I just wanted to point out that the DVOA stats are in fact, far from "little". If you read the methodologies, you'll realize that each team's ranking is a value that encompasses every single play of the season and compares them with every single play of every team in the season. Simply, DVOA has watched more football than you have this season.

Also, you suggested that anyone who's watched the Patriots over the last four years, knows that games are won by teams that play better, rather than teams with better stats. Well I'll counter your claim by saying, anyone who's followed DVOA and Football Outsiders over the past four years knows that the Patriots HAVE been winning games because they play better. And it doesn't hurt that they've also had the best stats to boot:
35.6% DVOA in 2004 (#1 in NFL)
22.4% DVOA in 2003 (#2 in NFL)

97
by Aaron (not verified) :: Wed, 10/19/2005 - 2:11pm

Um, guys, I don't think that "J Austin" guy is coming back, nor do I think he's interested in actual response.

I got an e-mail today accusing me of being a coward. I'm not sure a coward would post his e-mail address next to his commentary and invite this sort of abuse. By the way, this e-mailer used a spoof e-mail address so that I could not send him a response. Maybe he was afraid of what I would say.

I think the biggest problem is that people don't understand the concept of "objectivity." That's why I get so many e-mails accusing me of bias, or being a Raider fan (shudders), or "not respecting the Broncos." One e-mail said "I understand objective ratings, but can't you make subjective changes if a team should really be in the top ten?" No, because then they aren't objective ratings.

Mea culpa: Estimated Wins numbers were incorrect. I forgot to pro-rate the teams that had a bye week. Now fixed above.

98
by Pat (not verified) :: Wed, 10/19/2005 - 3:04pm

Regarding DVOA:

I really wish we had a good handle on exactly how special teams contributes to winning in terms of matchups.

By this, I mean, when is punting important? When is kicking off important?

Rushing and passing are easy to "matchup" against teams, because, well, pass defense matches with pass offense, and you can then go and look up a team's run/pass ratio, and get an estimate for how well to match up those areas of a team.

But kickoffs and punting are way different. If you've got a defensive struggle, punting can make or break the game. But if you've got two teams putting up points left and right, kickoffs mean everything.

Take a look at NYG vs PHI, for instance. Both have roughly equivalent offenses, but PHI has a better defense. The Giants have radically better special teams, and that's the entire reason they're ranked the way that they are.

But the question then becomes how much are the special teams going to play into it? PHI's punting is actually above average, but with the Giants weak defense, are the Eagles going to be kicking off more than they punt?

99
by seamus (not verified) :: Wed, 10/19/2005 - 3:12pm

#39:

Great point about the various rankings. I think the main point is that those rankings seem to focus on *current* strength, so they always seem to be 30% focused on what happened last Sunday, and maybe 70% the rest of the season.

100
by Larry (not verified) :: Wed, 10/19/2005 - 3:36pm

The Broncos are also going to have to convert a third down once in a while. Denver ranks 30th in offensive DVOA on third downs, just ahead of those powerhouses in San Francisco and Houston. What makes their failure to convert third downs even more problematic is the fact that, because of their great running game, Denver faces an average of just 6.2 yards to go on third downs, the third-lowest number in the NFL.

This strikes me as exactly backwards (or at least inconclusive). Since DVOA compares against average performance in the situation, then a low DVOA when having favorable situations doesn't necessarily mean anything bad at all. It could indicate good 3rd down performance in terms of number of conversions. Not as good as it should be given the 1st and 2nd down performance, but potentially still good in terms of converting the 3rd downs. This all depends on exactly how much better than normal that 6.2 yds is, as well as the baseline performance in such situations.

101
by Pat (not verified) :: Wed, 10/19/2005 - 3:48pm

Larry:

That comment seems to me to indicate that the Broncos are severely underperforming. Their offense should be doing much better than they are, but they keep failing on 3rd and short.

102
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Wed, 10/19/2005 - 3:55pm

#92's comment at the end is pathetic. "just cheer for your team"--what a joke. What is the purpose of the whole website?
As for the Bronco's, they are getting shafted in this ranking system period. There maybe reasons but there is no way anyone can say that the 5-1 Bengals have beat the calibur of teams the 5-1 Broncos have.
Both teams are still suspect for different reasons--DVOA just points out the reasons Denver is and not the Bengals.
What happens if Denver wins the next 2 weeks in a row but fail to move up in any category? Please answer that question Aaron.

103
by Jerry P. (not verified) :: Wed, 10/19/2005 - 4:01pm

"Take a look at NYG vs PHI, for instance. Both have roughly equivalent offenses, but PHI has a better defense. The Giants have radically better special teams, and that’s the entire reason they’re ranked the way that they are."

Agreed, and then there's the point that the Giants special teams might not even be that good. Didn't they have a punt return and a kick return for a TD in week 1 against the Cardinals? Their special teams rating has been dropping since then. With so few plays that are special teams and it being only 6 weeks in special teams DVOA seems to rise and fall less drastically. I'm just waiting for Buffalo to regain the top spot in that category on the legs of Terrence McGee.

104
by Bj&ouml;rn (not verified) :: Wed, 10/19/2005 - 4:07pm

To answer #102:

If my Broncos win their next two games by a combined 60 points, they will move up in the DVOA rankings. However, if Denver wins their next 2 by a combined score of 6-4, while turning the ball over 20 times per game, they will move down.

Curiously enough, the best thing for Denver in terms of the DVOA rankings would be for Miami to run the table.

105
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Wed, 10/19/2005 - 4:17pm

comment 102 is not for #92 but instead #95--sorry

106
by Jerry P. (not verified) :: Wed, 10/19/2005 - 4:19pm

"There maybe reasons but there is no way anyone can say that the 5-1 Bengals have beat the calibur of teams the 5-1 Broncos have."

Hello and welcome to Football Outsiders. Allow me to help you. Please direct your attention to the second table in this article. Please take note of the Bengals are listed as having the 3rd easiest schedule so far this year. The Broncos are listed as having the 3rd hardest. So, in fact, no one is saying the Bengals have beat the caliber of teams the Broncos have. DVOA says the exact opposite. Feel free to give credit to DVOA for supporting you in this claim.

"Both teams are still suspect for different reasons–DVOA just points out the reasons Denver is and not the Bengals."

Half credit. It points out why both are suspect. Check the defense table under "Just the Stats". See where the Bengals are 28th in run defense. That's where they are suspect. But you're new here. It's not your fault you were not aware of the other, often times more useful DVOA tables (as well as the handy offensive and defensive line stats) under "Just the Stats". Personally, I think these tables should be included in the weekly DVOA threads but given the time constraints (the overall stats are done before the offensive and defensive breakdowns) I think we can let Aaron off.

"What happens if Denver wins the next 2 weeks in a row but fail to move up in any category?"

Aaron is so damn reasonable as a human being he'll attempt to find out why. But fair warning, he may ask for your personal participation. Changes have been made in the past (especially is the wild days of 2003) to the system based on input in these threads. Another article detailing the complete evolution of the DVOA metric might be in order for new visitors. But that's the Outsiders' call.

Take care and enjoy your stay at Football Outsiders.

107
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Wed, 10/19/2005 - 4:22pm

Re: #104

Pretty stupid to think that the Bronco's could win and move down due to close wins over good teams--irregardless of anything else.
Seems DVOA needs to implement wins and losses at a higher level

108
by mawbrew (not verified) :: Wed, 10/19/2005 - 4:24pm

Re: 101

OTOH, they're doing a great job of avoiding 3rd and long. That should serve them well (assuming it continues).

109
by Oz (not verified) :: Wed, 10/19/2005 - 4:26pm

Re: 97

Aaron, when prorating Estimated Wins did you perhaps skip over Green Bay?
It looks like they're the only 5-game team that has the same EW value as before.

110
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Wed, 10/19/2005 - 4:27pm

Interesting comments

You are right--I see both tables and what they say.

Final say is Denver is 13 and Cinn is 1.

Whatever the stats say within the table are interesting and factual but the final rating is what matters.
Agree or Disagree

111
by Aaron (not verified) :: Wed, 10/19/2005 - 4:41pm

Actually, the final rating isn't what matters. Winning the Super Bowl is what matters. Oh, and yeah, thanks OZ, for some reason the program listed GB's bye as 7 instead of 6. But that means that their ST rating is incorrect, so actually they have 1.8 estimated wins and ST drops from 28th to 29th.

112
by mawbrew (not verified) :: Wed, 10/19/2005 - 4:42pm

Re: 100

Assuming the third down DVOA sited is, in fact, adjusted for distance (I'm not sure that's the case), I tend to agree with you. It would seem less problematic at shorter distances because the overall liklihood of success is greater in those situations.

113
by doktarr (not verified) :: Wed, 10/19/2005 - 4:44pm

re: bobman #57:

Vanderjagt has been below arerage this year. That's not a statement of my personal opinion; it's a statement of the FG/XP DPAR. Now, I'll grant that the sample size (only two difficult kicks attempted, one missed) is very small. But it's easy to see why DPAR thinks he has played poorly, because, in very limited time, he has.

Is this a statistical fluke? It's entirely possible. All I was trying to do was explain WHY the Colts have had bad numbers.

Your supposition that the Colts have bad return numbers because of the fair catches is dubious. The main reason, I suspect, is turnovers. Is this because of injuries? Maybe. Will they improve as the year goes on? Maybe.

No point in mentioning Hunter, since the Colts' coverage game is fine.

Again, all I was trying to do is explain WHY the overall rankings have the Colts behind the Bengals. Based on this year, it is reasonable. If you think the Colts' special teams struggles are just an early season fluke (and I wouldn't really argue with you), then that's fine; pencil the Colts in at #1. I'm just looking at what the algorithm looks at, and I think it's interpretation of the data is reasonable.

114
by Bj&ouml;rn (not verified) :: Wed, 10/19/2005 - 4:47pm

Let me break this down for you so you understand, Bronco Fan.

A win is just a number. A 5-1 team is no more likely to win a game than a 3-3 team, based just on those numbers. If Denver plays terribly and wins in spite of itself (like against Washington), they do NOT deserve to move up!

If you don't understand this, then just go back to the CBS Sportsline message board from whence you came.

115
by Aaron (not verified) :: Wed, 10/19/2005 - 4:59pm

It is adjusted for distance -- i.e., the benchmark for success is different on third-and-5 than it is on third-and-10 -- and I'll go through the Denver third downs in the mailbag.

As we speak I'm working on some research with estimated wins to see if it might make sense to base the FOX ratings on those rather than straight DVOA. I'm also playing around with "weighted estimated wins" and other such things.

116
by B (not verified) :: Wed, 10/19/2005 - 5:08pm

I think estimated wins would be the best metric for the Fox power rankings.

117
by josh (not verified) :: Wed, 10/19/2005 - 5:10pm

If you want to use these numbers to make predictions, shouldn't you regress toward the mean. I would think that NY Giants special teams DVOA especially needs to be regressed.

118
by Brian (not verified) :: Wed, 10/19/2005 - 5:18pm

Wow, the Jaguars' schedule for the rest of the year is unbelievably easy. Of course, they had to play the hardest schedule in the league the first 6 weeks to get it. Indy is most likely going to have to win a LOT of games just to win the division! If they should lose to JAC in their second meeting they could go 14-2 and not win the division. Other than the showdown with the Colts, Jacksonville's 2nd most difficult game the rest of the season is the Browns.

119
by B (not verified) :: Wed, 10/19/2005 - 5:24pm

And to think, I haven't been taking Jax seriously. Maybe they are for real.

120
by josh (not verified) :: Wed, 10/19/2005 - 5:25pm

I have a suggestion, not about DVOA itslef, but about someting that I think would be intersting to include on the chart. Along with expected wins you could run future expected wins based on the remaining schedule DVOA. We could then, in future threads, refer back to actual predictions of the system. This should correlate better win actual wins than DVOA anyway as it takes into account the remaining schedule. Is that possible?

121
by Pat (not verified) :: Wed, 10/19/2005 - 5:29pm

josh:

I don't think there is currently a metric for "head to head estimated win percentage". See arguments above regarding special teams.

I think:

DVOA correlates to point spreads.
Estimated wins correlates to total wins at end of season.
No current metric correlates to win probability.

At least, I don't think any metric is designed to.

122
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Wed, 10/19/2005 - 5:36pm

While you try to clear everything up for me--Jax is good but look, Denver, while losing to Miami, beat Jax at Jax in every category.
But you are right, a win is just another number by DVOA power rankings--but as Aaron mentioned, winning the superbowl is what really matters. I couldn't say anyone is going to beat Indy so the rest of this seems worthless--unless you are a gambling man.

123
by B (not verified) :: Wed, 10/19/2005 - 5:49pm

Indy is the class of the AFC, but they do have some weaknesses, espicially on defense. They struggle to stop the run, and they struggle against passes to running backs and tight ends. Of course, if thier offense scores on every possession, the defensive weaknesses are irrelevant.

124
by James Gibson (not verified) :: Wed, 10/19/2005 - 5:50pm

Let me caution Broncos fans (and I am one) - don't forget a recent history of starting with a winning record only to fold as the season goes on. They still haven't won a playoff game since Elway retired. And the especially painful season to me was not the two times they got mauled by Indianapolis in the playoffs, it was the year before when they started 6-2 and finished 9-7, missing the playoffs entirely.

Now, I didn't start following DVOA very closely until last year, but one of the impressive things last year to me was San Diego's high DVOA early in the season with a 3-3 record and then the Chargers finishing 11-5.

125
by B (not verified) :: Wed, 10/19/2005 - 5:58pm

One benefit of the Patriots slide into medirocity is the "dvoa is stoopid because it disrespects my team" are no Broncos fans and not Patriot fans. Maybe later we can have a Manning vs Plummer debate? Perhaps we should give the Broncos fans one they can actually win. How about Edge vs Bell?

126
by B (not verified) :: Wed, 10/19/2005 - 5:59pm

Only two dropped passes? I think that qualifies as a good day for Dez White.

127
by B (not verified) :: Wed, 10/19/2005 - 6:02pm

Oops, wrong thread. I feel like Carl.

128
by MJK (not verified) :: Wed, 10/19/2005 - 6:03pm

Bronco Fan,

With regard to your contention that the "final rating is what matters", I would counter that this isn't the case. Fans (myself included) enjoy putting one number on a team to rate how good it is, and we get happy when someone puts a good number on the team we root for. But if you're interested in using numbers to understand the game, one number doesn't do enough. The reason is that football is a multi-faceted game, and different teams strengths will play into different teams' weaknesses. See some of the above comments in this thread regarding how important special teams is depending on how often the two teams score.

It is quite possible that a team ranked higher by total DVOA (or by more primitive win/loss percentage, or any other objective ranking method) could still be at a disadvantage playing a team ranked lower, based on its individual strengths and weaknesses. For example, Team A might be ranked higher than Team B due to the fact that A has an unbelievable WR corps that usually more than makes up for having an abysmal running game, giving A a slightly above average offense against an average defense fielded by B. But if B's defense is average because it's phenomenal against WR's, but porous against the run, then B's strength cancel A's strength and A is unable to capitalize on B's weakness. Meanwhile, if B's TE's are above average, and A's TE defense is below average, then B will score on A even if, overall, B's offense and A's defense are average. Sorry for the long winded example, but that's the best way I can think of putting it.

The DVOA rating on Fox attempts to combine all the different little DVOA pieces into one general rating. To be honest, I am not wonderfully well versed either with Denver or with how DVOA is combined. But I would argue that what the power rankings are NOT saying is that, because Denver is ranked about 12th or so, means that Denver will finish the season as the 12th best team measured by W/L. Which maybe is what Aaron meant by saying that "winning the superbowl is what matters".

To Aaron or any other better versed Outsider/fan: is there a good description somewhere of how total DVOA is put together? Is it some weighted combination of rush DVOA+Pass to WR DVOA+Pass to TE DVOA - RunD DVOA - PassD DVOA...etc? Or am I misunderstanding? If it is a combination, how are the weights calculated? Could this be a tweak to improve correlation with performance?

129
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Wed, 10/19/2005 - 6:04pm

I agree on the Bronco's recent history-Plummer is going to have to win games for them sooner or later.
Typically they rely on Shanahan to strategize them through the beginning of the year. The other teams see about all he has after week 7 or 8 and then the players have to come through.
The problem is that there is no Elway anymore.
Interesting about SD last year, I didn't consider that.
Manning loves the Broncos ever since the Broncos killed them in the regular season a few years ago. Since he just dominates them and will this year if they play

130
by Catfish (not verified) :: Wed, 10/19/2005 - 6:05pm

Born a Broncs fan-

You are right, all that really matters is wins and losses. Nothing can change the fact that the Broncos are 5-1. What DVOA shows is how well the team has played every play of the season, and is thus a better indicator of overall quality. This makes DVOA actually better at predicting how well a team will do for the rest of the season, or next season than wins and losses.

Here's an extreme example (nothing to do with the Broncos, merely to illustrate a point). Say Team A is 6-0, but all its wins came on last second field goals. Team B is 4-2, but their 2 losses were by last second field goals and their wins came by 4 TD's each. Assuming both played an average schedule, which do you think is the better team? Right now, Team A is ahead in the standings; but based on Team B's performance, you would expect them to win more games over the rest of the season. Aaron's rankings would reflect this and put Team B ahead of Team A.

131
by MJK (not verified) :: Wed, 10/19/2005 - 6:11pm

And if it's any consolation for Broncos fans, based purely on Wins and Losses and strength of schedule, Denver ends up ranked around 3rd or 4th or so (depending on the ranking algorithm you use).

What DVOA says is that if the Broncos want to continue being ranked so highly, they had better play better than they have on average, especially better than they did in the Miami game, because right now they're overperforming their average level of play.

132
by james (not verified) :: Wed, 10/19/2005 - 6:12pm

B,

Is the defense vs. running backs, tight ends, wrs available on the site?

133
by Brian (not verified) :: Wed, 10/19/2005 - 6:16pm

Indy does have several weaknesses.

Poor tackling is still a problem, although much improved over the last several years. This is the reason for their poor run defense rating, they usually get to the runner quickly but sometimes miss the tackle.

Inexperienced secondary is also still a weakness, although once again it is MUCH improved. They are giving the line more time to get to the quarterback, but Monday night showed that if the opposing O-line can hold them off without using a max protect package they can get some big plays.

The offense misses Pollard. Clark is an outstanding #2 TE, but as the #1 he has had some struggles (just like last year when Pollard was hurt). The TEs have dropped quite a few 3rd downs this year to kill drives. The rest of the offense has been a little inconsistent at times with dropped passes and bad throws, but those seem to be correcting themselves. The much less dominant TE packages are not.

That said, all the teams seem to have glaring weaknesses this year. Go Colts! (says the Indy resident)

134
by Jimmy Two Times (not verified) :: Wed, 10/19/2005 - 6:18pm

Question:

I know the FO staff has been kind enough to start posting defensive DVOA by receiver type under the "Team Defense" section of Just the Stats. The thing I find myself wondering about most often, though, is situational DVOA; not only DVOA by down (which is the most important thing), but DVOA by half, on the road, in domes, etc.

Are any of these numbers available? I'm a little surprised that at least DVOA by down isn't up here anywhere; am I just missing it?

135
by B (not verified) :: Wed, 10/19/2005 - 6:19pm

James: Yes, it is:
http://www.footballoutsiders.com/stats/teamdef.php
Scroll about halfway down.

136
by J Austin (not verified) :: Wed, 10/19/2005 - 6:22pm

I apologize to any who were offended by #92 because of grammer. I didn't realize that this website was based on who had the best english. I was under the impression that we were arguing opinions about football. So sorry if I looked like an idiot. Granted I did not watch this website the previous 4 years so I did not see the patriots rankings but I do however watch football like I have ocd every Saturday and Sunday. Pardon me for not caring about anyone but my Broncos. I do not know the overall stats of the patriots however I have seen numbers of games where they were dominated everywhere but on the scoreboard. Some of my information may be wrong but I am glad to see that the website is looking into weighing in qaulity wins when I mentioned Denvers record against .500 teams. After all it doesn't matter how you beat them as long as you beat them. As Aaron said it's not the final rankings that matter but who wins the super bowl. So if my Broncos finish the season #1 in these power rankings but lose there first playoff game again I will be once again mad and dissapointed.

Live Eat and Die by the Broncos

137
by R.J. (not verified) :: Wed, 10/19/2005 - 6:29pm

My elementary understanding of DVOA is that it values sustained drives that lead to TDs over big play "drives" that result in TDs. Is that right? It makes sense that it would, because an offense that demonstrates it can convert 3rd downs and sustain a long drive seems more likely to continue to succeed than one that relies more on big plays.
Also, wouldn't a defense that allows sustained drives but prevents "big plays" be generally rated lower under DVOA as one that forces a lot of 3 and outs but also gives up a couple of big plays?
Just from watching it seems obvious that the Denver Broncos on both offense and defense have been succeeding in a way DVOA would not particularly value (if my understanding is correct). A majority of their TD "drives" have been of the "one big play" variety and the defense has given up a lot of long drives but almost no "big plays" in their wins. Perhaps it's mostly a fluke or perhaps the Bronco can sustain it, given their specific ability to make big plays on offense (especially in the running game) and their specific ability to take away big runs and big pass plays (while giving up a lot of short pass plays).
Does any of the foregoing make sense?

138
by Larry (not verified) :: Wed, 10/19/2005 - 6:30pm

Pat (#101):
I think Aaron made my point while I was having lunch, but to clarify, I can see only the following interpretations of low 3rd down DVOA combined with on-average shorter 3rd downs.

1) Given that 1st and 2nd down have been good (and that's not guaranteed, since they might not be getting as many 1st downs on those plays) we'd expect 3rd downs to be better. Even granting the first premise, this isn't really a very good conclusion, since the expectations are so much different (45%/60% of needed yards vs. 100%). And besides, one of the FO anti-CW bits is 3rd down performance may vary considerably. Although, I can't recall having seen any indication on how well 3rd down DVOA correlates to 1st and 2nd down in the same season as opposed to the following one.

2) The net effect for the Broncos is just fine. They get to 3rd and short and convert them, but not as much you'd expect given that these are shorter than average 3rd downs. But, it's still better than the average team since the average team has 3rd and 9 instead of 3rd and 6.2 (note: made up numbers alert).

2b) The Broncos have a major problem, they get to 3rd and short and don't convert them at all.

We can't distinguish 2 vs. 2b from DVOA rank only. Actual VOA numbers, perhaps, and not even then, we'd need the baselines. But, definitely not DVOA rank.

139
by B (not verified) :: Wed, 10/19/2005 - 6:30pm

" I do not know the overall stats of the patriots however I have seen numbers of games where they were dominated everywhere but on the scoreboard."
Care to back this up with actual games? Tuck rule games don't count.

140
by Catfish (not verified) :: Wed, 10/19/2005 - 6:32pm

Wow, 8 posts since I started typing

141
by MJK (not verified) :: Wed, 10/19/2005 - 6:32pm

Hypothetical question for Aaron and Co:

How difficult would it be to modify the "Just the Stats" pages so that teams were sortable by different headings? I.e. you could click on the "Defense vs. TE" heading and it would sort the list by who was best (according to DVOA) against Tight Ends? ESPN or Sportsline (I don't remember which) does that in their statistics (at least for MLB), and it's a lot of fun to sort lists by different qualities of the team and look for trends.

142
by B (not verified) :: Wed, 10/19/2005 - 6:43pm

Re #137: I think you need to consider the drives that don't end in points. If the defense forces a lot of three and out, typicially they will give thier offense good field position, making it easier for the offense to score. On the other hand, a team that employs a "bend but don't break" defense will make it harder for thier offense to score points. Now, if that offense employs a strategy where they either get 3 and out or a big play, if they have a couple bad breaks (IE 3 & out on successive drives), that gives the other team a short field. This puts a lot of strain on the defense, which makes it harder for you to win games.

143
by Aaron (not verified) :: Wed, 10/19/2005 - 6:44pm

Re 134 and 141: There were plans to do both of these things during the summer, and we ran out of time. During the season, time constraints become even more of an issue. I don't have the time to put tables together manually; nor do I have the time to manage a large programming project, even though I have programmers willing to do their part. We'll see what we can do, but I can't make any promises.

144
by Tim Kirk, York(UK) (not verified) :: Wed, 10/19/2005 - 7:05pm

re: 138 (and 101, and others...)

Listening to Bronco's game Sunday one stat that was mentioned was that the Broncos were having a terrible season so far on third and even medium long... I think they are probably converting OK on third and short, but when they do throw a first down incompletion if the second down doesn't do much they are going 3&out a lot more than a winning team should. Anyone playing the Broncos surely knows that if you double cover Rod Smith and can get any pressure on Plummer then you've a good chance of stopping a third down and more than about 6 yards... Denver need Lelie to start catching again, and some production from TE and 3rd WRs which has been only patchy for a long time.

They are also however putting together drives which yomp down the field without nary a third down in sight... But until they do that for most of a game and put 40+ points past someone it isn't going to look much more than having some big play potential and making it work just enough.

The problem is that they are generally playing better inth e first half than the second (and have played well enough in some first halves to run away with games that have closed up on them late). If they'd been doing this the other way around people would be thinking they have been lucky making last minute comebacks (like Washington earlier this year, and Jacksonville last season). Until they actually play for a whole game at the level they look capable of I'm not going to believe they will do much in the post season and I've been a fan for over 20 years.

DVOA does a really good job of pointing out that kind of issue, and given that Denver deserved to lose to Washington & Miami, and failed to put away New England properly, there are definite potential weaknesses. I don't know if they are real weaknesses, they _could_ always be brilliant judged games saving the maximumfor next week each time (but somehow I doubt it). Having a very friendly home-field for mitigates a tough looking schedule somewhat, and this weekend at the Giants is going to be an interesting test. (Also Miami & Washington are both teams playing much better than last year, so as the season progresses those performance may well rate slightly better [less worse...] too...).

There does seem to be something about this season though - with a number of teams doing much better or worse than expected, and some divisions really falling apart (NFC North) - or strengthening up (NFC East) that is going to really test any analysis I think, and make the weighting effects as the season progresses very interesting...

145
by Fnor (not verified) :: Wed, 10/19/2005 - 7:06pm

Aaron:
The whole process is automated, yes? You mentioned having a spreadsheet of each play that's imported from the play by play into something similar to the charts you have us making. So if it's automated, through excel or whatever you use, couldn't you just have the system spit out the numbers you want to see? If it is all just a big excel thing, then, maybe into another workbook which can then be imported into an sql table. That way you wouldn't have to really oversee anything, you'd just be providing variable values from your sheet into a form-table that was put together for you.
If that makes any sense.

146
by Jimmy Two Times (not verified) :: Wed, 10/19/2005 - 7:18pm

Thanks for the response, Aaron, and thanks also for all the stuff you do provide.

Given how much the book talked about DVOA by down (great stuff, BTW), I think I'm probably not the only one who'd be interested in seeing it. The comment that piqued my interest today was the one you made on FOX about how bad Denver had been on 3rd down. It seems to me that, given that we already know that 3rd down DVOA tends to regress towards 1st and 2nd down DVOA, compressing some of the runaway 3rd down DVOAs might give better projections.

Then again, I know that DVOA isn't supposed to be a predictive stat, so I'll shut up now.

147
by Todd S. (not verified) :: Wed, 10/19/2005 - 7:18pm

#128

This may not be exactly what you're looking for, but the general weight is 3/7 offense, 3/7 defense, 1/7 special teams.

148
by J Austin (not verified) :: Wed, 10/19/2005 - 7:45pm

It's me again. I agree with everyone that the broncos have had more than there fair share of 3rd down troubles this year. But I was just wondering if anyone caught the Jacksonville Jaguars game where the Broncos went 80 yards on 17 plays in 9 minutes killing almost the whole 2nd quarter. 8 running plays and 9 pass plays. Followed that up with another 65 yard drive that killed the last 5 minutes of the first half to take a dominant 14-0 halftime lead. I know I saw it live and in color and it was beautiful. So granted they have struggled but also showed against a good jaguar defense (that held the high flying colts to 10 points) that they are capable of putting together good long drives when needed. whith a few 3rd down conversions. Here's hoping they snap that streak and win the next couple so you all will become the believers in them that I have. Thank you

149
by putnamp (not verified) :: Wed, 10/19/2005 - 8:14pm

I've been wondering for some time if the "bend-dont-break" defense is just a myth used to describe defenses that give up a lot of yards but don't get scored on "as often", or just happen to be really good in the red zone. Seems like one of those rationalizations that announcers/writers use to compliment something they like even when it doesn't deserve it.

150
by B (not verified) :: Wed, 10/19/2005 - 8:21pm

The "bend but don't break" defense is kind of a myth, caused by a team that has a good offense, good special teams and a mediocre defense. The good offense and special teams constitently puts the opossing tewam in lousy field position. Then the defense gives up a lot of yards, but manages to stop before the red zone. See 2003-2004 Jets for an example.

151
by R.J. (not verified) :: Wed, 10/19/2005 - 8:25pm

#142
Of course it's true that relying on a "big play" offense and a prevent the "big play" defense makes it harder to win, unless it doesn't because the team is very good at both of those things.
To use a baseball analogy, imagine a team that only averages 6 hits per game but also averages 4 HRs per game. On the pitching side that team gives up on average 12 hits per game but also averages giving up less than one HR per game. (Unlikely, sure, but there are in fact hitters who don't hit for a high average but hit a lot of HRs and there are pitchers who give up a lot of hits but don't give up many HRs.) A team made up of players like that would win a lot of games even though many of the team stats (especialy BA and BA-against) wouldn't look very good.
I don't know if the Broncos really are that kind of team, but from watching their games it seems like they have an unusual ability to make big plays (especially with the run and play fakes off the run) and to prevent big plays (speed to the ball, sure tackling, good downfield coverage and a good pressure to prevent deep throws).
But also it seems they they have had a lot of 3 and outs on offense (poor 3rd down conversion, poor "drop-back" pass game) and they have not been good defending short, quick passes (given up lots of yards and sustained drives).
Sustainable formula for winning games? We'll see.
Explanation for relatively low DVOA ranking? I'd like to know.

152
by B (not verified) :: Wed, 10/19/2005 - 9:35pm

Sticking with the Denver theme, look at what happened to them in the last two games. They got out to an early lead, but then let the other team crawl back into the game by failing to convert critical third downs on offense. Fortunately they were able to make plays in the end to stall out the potential comebacks.

153
by Trogdor (not verified) :: Wed, 10/19/2005 - 9:36pm

I hate the Broncos.

But I do have some consolation for their fans, regarding them being ranked behind Cincy, despite having the same record against a tougher schedule. Keep in mind that opponent adjustments are still not yet at full strength, and won't be for 2 more weeks. So what? Well, this means that if they each played Generic Opponent (completely average in every facet) for the next 2 weeks, and had exactly the same results on every single play (that is, neither team played better than the other), the Broncos would gain ground on Cincy in the rankings, because their tougher schedule would give them more credit, while Cincy's cupcake schedule would give them less credit. (The same would happen if each team had a bye next week - Denver would rise and Cincy would fall).

Um.... I just noticed something. Denver has a positive DVOA schedule, yet their DVOA is lower than their VOA. Is that right? Shouldn't their DVOA be higher? Could this whole controversy be caused by a misplaced negative sign?

Re: 137

I don't think it's entirely accurate to say that long drives are favored over big plays. What the system does is limit the bonus from an extremely long play over just a really long one. For example, what's the difference between a 50-yard TD run and an 85-yarder? About 99.99999% of the difference is just having more field to work with. The back who scores from 50 most likely would also have scored from 85 or 109 (in Canada) or 426 (if the field was that long). So there's not a whole lot of difference between running 60 and running 80 in DVOA (I think there is some small difference, but not 33.3%).

154
by Aaron (not verified) :: Wed, 10/19/2005 - 10:05pm

Fnor, it's like any other project -- if you want it to look right and work right, someone has to manage it. That's my time issue.

The third down thing is much more of a year-to-year issue than it is game-to-game during a single season. When I did the midseason projection system last year, I learned that third-down performance doesn't necessarily return to "the team's natural mean" during the season -- and if it does, the effect is much smaller than during the off-season. Much of the issue is personnel changes made each off-season because teams know they have third down problems.

DVOA is supposed to be predictive, just as a long-run thing, not a "this next game" thing.

155
by R.J. (not verified) :: Wed, 10/19/2005 - 10:47pm

#153
By "big plays" I did not mean 85 vs 50 yd plays, I meant plays from scrimmage like more than 20-yds on running plays and more than 50-yds on pass plays. My impression was that Denver had "a lot" of plays like that and had given up very few. Looking at games logs, in their 5 wins the Broncos had 10 such plays and gave up only 1. I'm not sure if that means all that much except that again, it sure "seems" like Denver's strengths on offense are particularly the ability to bust a long run from scrimmage or complete a long pass after a play fake and on defense to not allow big runs (speedy, sure tackling LBs) and to stop long passes (good pressure, good deep coverage). Is it possible that DVOA does not fully "appreciate" this kind of "big play" team.
(On a perhaps related "big play" note, when looking at the logs I noticed that in the 5 wins Denver "forced" 5 turnovers that lead directly to points and gave up none.)

156
by Catfish (not verified) :: Wed, 10/19/2005 - 10:53pm

Um…. I just noticed something. Denver has a positive DVOA schedule, yet their DVOA is lower than their VOA. Is that right? Shouldn’t their DVOA be higher? Could this whole controversy be caused by a misplaced negative sign?

From my understanding, not necessarily. If I understand correctly, the opponent adjustments aren't made on total VOA based on opponent DVOA, but on a play-by-play basis. Teams get adjusted up when they gain 5 yards on a run on 2nd and 7 at midfield in the 3rd quarter, but their opponent allows an average run of only 3 yards in the same situation. Because of the complexity involved, I wouldn't doubt the existence of a situation where a team plays a harder than average schedule, but has a lower DVOA than VOA. Or it could be a simple mistake as you suggest.

157
by Trogdor (not verified) :: Wed, 10/19/2005 - 11:00pm

Yeah, I thought of that. It just seems really unlikely to me that their opponents could be so strongly positive, yet the adjustment would be strongly negative. Situations would have to line up almost perfectly for it to happen that strongly. Their adjustment of -4.9% is the same as Philly's, even though their schedule is 3rd and Philly's is 26th. That just strikes me as really odd.

158
by Pat (not verified) :: Wed, 10/19/2005 - 11:53pm

MJK:

To Aaron or any other better versed Outsider/fan: is there a good description somewhere of how total DVOA is put together? Is it some weighted combination of rush DVOA+Pass to WR DVOA+Pass to TE DVOA - RunD DVOA - PassD DVOA…etc? Or am I misunderstanding? If it is a combination, how are the weights calculated? Could this be a tweak to improve correlation with performance?

Total DVOA is off+def+ST.

Offensive DVOA is the average of all offensive plays. So if a team was 50% pass, 50% rush, it's just (rush+pass)/2. Philly, for instance, was roughly 3/4 pass, 1/4 rush: so total offensive is 3/4 pass + 1/4 rush. (If they were 50/50, they'd be 8.7% - they're actually 15.9%) Ditto for defensive. (Most defensive ratings are basically rush+pass/2, since that's what things average out to.)

Special teams is magic. :)

159
by thad (not verified) :: Wed, 10/19/2005 - 11:55pm

So far this year Denver has had on offense
9 rushes of 20 yards or more
10 passes of 20 yards or more
2 passes of 40 yards or more
on defense
2 rushes of 10 yards or more
17 passes of 20 yards or more
3 passes of 40 yards or more
according to NFL.com

160
by R.J. (not verified) :: Thu, 10/20/2005 - 12:25am

#159
I was only looking at the games Denver won, thinking that possibly DVOA was missing something particular about them in their good games. In those games Bell had a 68-yd run and scored on 55-yd & 34-yd runs, Anderson scored on a 44-yd run and Plummer completed downfield passes of 72 and 55 yds. They also had runs of 39, 24, 22 & 20. The D gave up one run longer than 20 (21-yds). The D also gave up a single "downfield" pass of more than 40 yds. (A couple of "short" passes turned into long plays but they weren't long throws).
This kind of performance seems unusual but I don't really know if it is. Do other teams have that many long runs from scrimmage? Do other teams give up so few long plays?
Finally, it seems like DVOA would rate higher a team that on offense consistently gained 4-6 yds per play over a team that gained 1-3 yds on 90% of their plays but had huge gains on the other 10% and on defense consistently gave up 3-4 yds per play but almost never gave up a big play. Am I wrong about that?

161
by zip (not verified) :: Thu, 10/20/2005 - 12:30am

Re: #154

Given the geekiness of football stats, I bet this site has quite a few html/mysql/php programmers lurking who'd be honored to help. Or maybe I'm an abberration.

Obviously it's never that simple, but I thought I'd point it out.

162
by zip (not verified) :: Thu, 10/20/2005 - 12:34am

#160

I think you're right, I'm pretty sure I remember that "home-run hitter" type backs (Barry Sanders would be the most notorious example) tend to be considered less valuable by DVOA and that has been mentioned in many articles. I would have to agree with that.. if two backs have the same expected value (in yards) per rush but one has a much larger variance (the home-run hitter), I'd rather have the consistent guy.

163
by Nate (not verified) :: Thu, 10/20/2005 - 12:50am

To all those who want sortable charts:
Although it would be nice, it's trivially easy to just cut and paste the tables into a spreadsheet of your own.

164
by yeah right (not verified) :: Thu, 10/20/2005 - 4:47am

Let me tell you something about Barry Sanders to all the football stat "experts". EVERY single coach would trade their present running back for Sanders. In a second! Except perhaps Chargers. And the excitement and value someone like Sanders brings to an organization is impossible to measure. Its players like Sanders that contribute to the making of a "storied" franchise. Its their stories that create the folklore for generations to hear and see on espn classic.

165
by J Austin (not verified) :: Thu, 10/20/2005 - 7:32am

#164
Not dissing Barry Sanders or nothing because he's one of the greatest of all time. But which storied franchise are you talking about. Surely not the Detroit Lions. A team who can't even win games with the all star lineup that the Colts have. Granted Harrington will be replaced this weekend but the Lions are not legendary because of Barry Sanders. Only Barry Sanders is legendary because of his play. And I'm sick of all this excel spreadsheet formula hoopla. Next week let's talk football and not who can come up with a better computer program. Thank you

166
by Aaron (not verified) :: Thu, 10/20/2005 - 8:02am

Nobody who writes for this site has ever said that Barry Sanders was not a great running back. For some reason he keeps coming up as a running back that our stats don't consider as good even though the only year that we have stats for him is his last year, which happened to be his worst year in general. I'm not sure why this "Football Outsiders thinks Barry Sanders wasn't valuable" meme still exists but it makes us look like idiots for something that we have never said.

167
by Karl (not verified) :: Thu, 10/20/2005 - 10:06am

I just noticed the FoxSports Aikman Efficiency Ratings. Why would they publish these if they now know they have access to the more reliable FO Efficiency Ratings? Because it's Aikman?

[The link is in my name]

168
by Aaron (not verified) :: Thu, 10/20/2005 - 10:39am

Yes, because it is Aikman... They're going to put me in touch with the guy who helps to compute those, looking to share ideas.

Notice that the Aikman Rankings have Denver lower than the DVOA rankings on defense, and the exact same on offense. That Troy Aikman is clearly a computer geek in his mom's basement who never played the game.

169
by Karl (not verified) :: Thu, 10/20/2005 - 11:01am

Just don't let them turn the letters DVOA into TROY.

170
by B (not verified) :: Thu, 10/20/2005 - 11:42am

For a team that's like Denver, perhaps we should look at the 2003 Panthers. Like Denver, that was a team who's offense got big plays (mostly through the passing game, although DeShaun Foster had a bunch of big runs) but would struggle picking up 3rd downs. (Look at Steven Davis's success rate that year 46%, right around where Bell and Anderson are this year). It's not a perfect fit cause Carolina was a little better on defense and a little worse on offense, but that too was a team that kept winning despite DVOA's low ranking. Then during the playoffs, CArolina had the three best games of thier season, playing way above expectations, and they were a field-goal away from winning the superbowl. Could that happen again? Maybe, but to follow the pattern, they'd have to lose to the best team in the NFL (according to DVOA). That would most likely be the Colts. Look on the bright side, that probablly won't happen till the AFC championship game, so that means Denver will win at least one playoff game without Terrell Davis or Elway.

171
by zip (not verified) :: Thu, 10/20/2005 - 12:02pm

#165

And I’m sick of all this excel spreadsheet formula hoopla. Next week let’s talk football and not who can come up with a better computer program

You're at the wrong site, my friend.

#162

(Barry Sanders would be the most notorious example)

I knew I shouldn't have said that. What if I said Deuce McAllister? Deshaun Foster? Oh, screw it.

I apologize for suggesting that DVOA might have a lower opinion of a back "like Sanders" than the general public would. As Aaron mentioned there aren't actual numbers on Sanders so it's just conjecture...

But shouldn't you be vandalizing Tommy Maddox's house or something right now?

172
by B (not verified) :: Thu, 10/20/2005 - 12:28pm

Deuce McAllister should never be compared to Barry Sanders. If you want to look at how a back like Sanders would be rated by DVOA, look at Tomlinson. That's not even a fair comparison because Tomlinson is more like Sweetness that Sanders.

173
by TroyF (not verified) :: Thu, 10/20/2005 - 12:41pm

Being a fan of Denver, I really don't care where they are ranked here. It doesn't matter. I do think the rankings "look" a bit loony. . . but we'll see what happens. Maybe they will fall apart again and maybe they are a paper tiger.

One thing though. . . while many of their second half have been below average (and there can be a variety of explanations for that up to and including they suck), the most important game they played this year was the second half against SD. The offense performed horribly that half. But the defense allowed something like 41 yards and 3 points to the #2 offense in DVOA.

Personally? I think this Denver team is different from the last few years. They aren't turning the ball over (17 quarters in a row) and they are running the football well. Moving to defense, KC, SD, JAX and NE all scored a TD less than their season averages. And Denver didn't have Champ Bailey for most of those games and was playing 3 rookies extensively in the secondary.

But you can't discount the game against Miami. And you can't discount the second halves of Washington or NE or even KC. Not when using statistics as the base. We can argue about how the statistic is put together or even say the stat is horrific, but you can't just use some stats you like and discount all the ones that don't fit.

And with the parity and closeness of the teams in the NFL, you have to KNOW there will be some big variances with any statistical model. It'll be interesting to see how Denver/Miami/Cincy fare over the next five or six weeks and see what happens with the overall DVOA. For that matter, it'll be interesting to see the Colts as their schedule gets a bit stiffer.

Good stuff Aaron.

174
by B (not verified) :: Thu, 10/20/2005 - 1:02pm

Regarding Denver and the second half, what i think is going on is Shannahan's offensive strategy tends to be build a big lead in the first half, and protect that lead by running the ball in the second half. This has worked great in the past, but recently (Perhaps since the loss of Portis), the running game has lost it's consticency. They still get the big runs, espicially from Bell, but thier overall success rate has dropped below 50%. This means they're not sustaning long drives as often as they used to, and it shows up more in the second half.

175
by J Austin (not verified) :: Thu, 10/20/2005 - 1:36pm

I don't think the problems in the second half are due to the loss of Clinton Portis. Mike Anderson and Tatum Bell are both terrific backs. I think the problem is the offensive line is not what it used to be. As well as teams figuring out Shanahan's strategy and a way to stop it. They stack the lines and dare Jake Plummer and the receivers to burn them on 3rd down and so far this year they have not been able to come through on that. Mainly due to teams double covering Rod Smith due to lack of a good tight end or 2nd or 3rd receiver. Ashley Lelie needs to begin to step up so that Plummer can find Rod Smith on 3rd down the way he did to end the game against he Patriots. Until they step up the running game will continue to struggle in the second half forcing the defense to spend too much time on the field. And when Champ Bailey returns to full health that will do a number for our pass defense.

176
by Aaron (not verified) :: Thu, 10/20/2005 - 1:39pm

Here, let's try this analogy. Barry Sanders is to Deuce McAllister as Emmitt Smith is to LaMont Jordan. In other words, similar style as far as consistency vs. boom-and-bust, but totally different planets when it comes to total value.

177
by R.J. (not verified) :: Thu, 10/20/2005 - 3:18pm

As one of the Denver followers who have kept this discussion going, let me say that I don't have a problem with the rankings. I may be a fan of the Broncos but I'm a bigger fan of good football analysis, which is why I come to this site. Some of the most interesting and insightful discussion of football goes on here. Kudos.
As far as the DVOA rankings, it does strike me as odd that a team with this many wins against the 3rd most difficult schedule would be ranked so "middle of the packish". So I would like to understand what Denver is doing that allows them to win games without making much of an impact on a "play-to-play" analysis and comparison that the DVOA rankings does. I'm wondering aloud here whether their specific strengths are overlooked somehow. Or if maybe their clear weaknesses are "overcounted" by DVOA analysis. Or if they are flukishly overperforming their expected win totals. Either way it will be interesting to see if they continue to win this way or if whatever DVOA sees catches up to them.

178
by Fnor (not verified) :: Thu, 10/20/2005 - 4:00pm

#177:
Amen. That is what I'm talking about, that people decided was heresy. Root for you team AND good analysis. Analysis isn't about respect. It's about numbers and helping us understand the game.

#175:
Bailey is a fun case. See also Weighing in on Champ Bailey. Expect a boost, but not an astronomical one, I'd say.

179
by J Austin (not verified) :: Thu, 10/20/2005 - 4:31pm

Good article Fnor. I missed that article because I just started looking at this website yesterday. Even as a biased fan I do not expect the Broncos to jump to the top of the pass defense charts just by adding in Champ Bailey. Obviously he will have bad games as well as the other corner still having to perform. Which often times they do not for the Broncos. As seen in the second half against the Patriots team. Seems they have this problem on both sides of the ball with recievers as well. I do however expect the defense to improve somewhat with his presence and allows Darrant Williams and Foxworth to split time on the other side. Hopefully taking out Lenny Walls who is a joke at corner. I've seen him live and half the time he loses the reciever from the snap. I think with the presence of Champ Bailey they can rush the Quarterback more in the second half without worry of getting burned on every play. I don't expect an astronomical difference but hopefully enough to change some stats and help us close out dominant performances. Eventually the power rankings will go up or we will lose and earn what we have now. Also the charts on corners are a little off. Granted it looks bad on Champ Bailey to give up catches on 59% of balls thrown his way. But as noted you can not see the corner after the snap unless the ball is thrown his way. How many times was a play broken up because the go to receiver was covered by Bailey and was not able to get open. It may not help yet. But % of snaps played vs. # of completions given up should also way into what kind of cornerback he is. Or at least # of pass plays not neccissarily over plays.

180
by Bj&ouml;rn (not verified) :: Thu, 10/20/2005 - 7:23pm

Can we all go back to talking about the Patriots, please? All this attention for my team is kindof... creepy.

181
by Jim (not verified) :: Thu, 10/20/2005 - 8:10pm

I would like to say thank you Aaron for your hard work and keep up the good fight.I cant believe so many posters have all these wonderful ideas on how you should tweak something to make the broncos higher,if you want to be a homer call your local talk radio.I would never lay my cash down only on Dvoa #'s but I do like to use as a tool to formulate or to back up my instincts.What some people have to relize is nothing is ever going to be 100%.Just like when a team decides to put 7 men in the box and Denver cant chop block, Plummer is not going to put the team on his back and win!Beat some good teams on the road (not the Giants)and then you homers can flap your gums,Get off Aarons back and dream about Elway and davis when jake blows your playoffs. And I dont care about grammer and spelling,Just cashing my tickets on monday FLAMEOUT

182
by B (not verified) :: Thu, 10/20/2005 - 9:46pm

Bjorn: I don't want to talk about the Pats, it's too depressing, but I will talk about the Giants, if that makes you happier. I don't think they're nearly as good as DVOA indicates. Well, thier special teams is for real, but the offense is highly overrated. Eli is a good QB, but he is really incostient. The only really impressive game I've seen the offense have was when they beat up on the Rams, but that was because the Rams defense never figured out that Burress covered by a CB a foot shorter than he is will always win jump balls. Combine this with Brandon Jacob's upright running style and the fact that nobody on the team can hold onto the ball except Tiki, and i don't see this as a top offense. Combine this with a defnese that's pretty mediocre, and i see a team that'll get creamed by Denver on sunday. That's why I'm picking the Giants to win.

183
by R.J. (not verified) :: Thu, 10/20/2005 - 10:12pm

#181
It seems to me that winning at Jacksonville would count to most reasonable people as beating a good team on the road. The Bengals couldn't do it (and neither they nor the Colts have yet to beat a "good team" anywhere). The double irony of course is that of all the football sites (and all the threads on this site), you decide to take a position here that a team needs to "win" games to prove itself. Priceless.

184
by Bj&ouml;rn (not verified) :: Thu, 10/20/2005 - 11:26pm

B: Here's my two cents on the NY-Denver game. If I'm Denver, here's what I do. Put 2 men (Lenny Walls and Darrent Williams) on Burress. Walls is 6 foot 4, so Plax won't win the jump balls. That leaves Champ Bailey to cover Amani Toomer. Seems to me like the Giants will have to go three wide if they want to pass, because there is no way Toomer is beating Bailey, and Plax doesn't have much experience with the double team (probably). I would trust DJ Williams, Al Wilson, and Ian Gold to be able to stop Barber.

As an aside, this game should be a clue about how the Broncos will play against Randy Moss and the Raiders. I have been asking myself since Moss arrived in Oakland if Denver will go man to man with Bailey, man to man with Walls (oh god no) or some kind of system like I think Denver will use on Burress. Only problem against the Raiders is that Jerry Porter is good enough to beat Bailey.

185
by B (not verified) :: Thu, 10/20/2005 - 11:38pm

Bjorn: Hey, I thought you said you didn't want to talk about the Broncos. Although it is good to know they have a tall corner, that's one key to stopping the Giants. The other is thier kickoff & punt coverage.

186
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Fri, 10/21/2005 - 12:45am

Boy--the Bronco's are popular here. They will probably get killed this week and that will justify the rating.
Aaron, this maybe a stupid question but I am still learning what the DVOA is all about. Does the system take into account the situation at all--for example, a 3rd and long from the 22 yard line at the end of a half and behind 4 instead behind 3?
So much of success in football is situational--seems to me this is maybe the biggest factor for 2nd half and in particular 4th quarter performance.

As to the Bronco's winning on the road--read man--JAX is a favorite for the rest of the way according to DVOA and Denver dominated them on the road. That is one of the reasons football fans are arguing the Bronco's rating--they have one of the toughest schedules and have the second best record so 12 or 13 doesn't match up. We just want to completely understand why--not tweak it for our team.

187
by Aaron (not verified) :: Fri, 10/21/2005 - 1:22am

Bronco Fan, I gotta tell you, of the 150 or so emails that I've gotten and the 300 or so Denver message board posts that I've read, I'd break it down like this:

10 people who follow the site and are worried about Denver's low rating.

10 people who are seriously interested in how the ratings work and offering suggestions for why the ratings might be mistaken about Denver. (Usually this has to do with the weather in Miami, which is a reasonable objection.)

50 people who hate FOX for being Republican even though FOX Sports and FOX News are in no way connected.

50 people who ask me questions that have already been explained in the commentaries.

300 people who just want to call me an idiot who knows nothing about football.

30 people who want to call me a faggot or threaten my family.

188
by zip (not verified) :: Fri, 10/21/2005 - 2:08am

#187:

"Every Message Counts: Aaron's Email"

189
by Kibbles (not verified) :: Fri, 10/21/2005 - 3:32am

Aaron, any word yet on why Denver has the 3rd hardest schedule to date according to DVOA, yet their VOA is higher than their DVOA (meaning they're getting adjusted DOWNWARD when you account for opponent, as if they've played an easy slate)? It seems very... strange, to say the least.

To wit: Of the 16 teams that have faced the toughest schedules to date, according to DVOA, 12 have a higher DVOA than VOA. Denver, Dallas, Tennessee, and NYJ are the exceptions. Dallas's schedules was ranked 13th, Tenn's 14th, and NYJ's 10th, and Denver's THIRD. Why are they getting adjusted downwards, then? I can understand Dallas and Tenn, actually, it's just NYJ and especially Denver that confuse me.

Adding fuel to the fire... every single team with a schedule ranked between 16-32 has a DVOA lower than their VOA. So basically, teams with schedules ranked 3, 10, 13, 14, 16-32 all had their VOA adjusted downwards to get their DVOA. The 13, 14, and 16-32 make perfect sense. The 10 I can sort of kind of halfway see. The 3 just stands out like a sore thumb. Is there any chance the numbers are off, or could you explain this one in a mailbag?

190
by Karl (not verified) :: Fri, 10/21/2005 - 8:44am

Jeeze Aaron, that's nuts! Keep your head up, don't get an ulcer, and set up some type of email filtering to relieve some of the flow of negativity. It must be draining.

191
by Trogdor (not verified) :: Fri, 10/21/2005 - 9:04am

I've been doing a little more thinking on that, and I've thought of ways that it's possible. I'll use a completely exaggerated example to illustrate.

Let's say the Broncos have played the 2003 Chiefs this year. Those Chiefs finished the year ranked #1 in DVOA (though only #6 in weighted DVOA). So by playing them, their schedule strength should go up, because "past schedule" appears to be just the average DVOA (or VOA) of teams previously played.

Now this is where it gets fun. Remember that KC team had a craptacular run defense (28th). So Denver decides to run the ball on nearly every play. They manage to hold the ball for about 50 minutes that game, running about 85 times and passing maybe 10. You see what's going to happen? The vast majority of plays in this game are being run against a nearly league-worst unit. So every successful run gets downgraded because it only came against KC. Also, by holding the ball so long, they don't defend many plays against KC's top-ranked offense. So when all the plays are totaled and weighted, there will be a ton for Denver that are downgraded, and very few that are upgraded. The net result is, even though their "past schedule" will go up, their DVOA for this game will be lower than VOA.

I hope that explains how this phenomenon is possible. The question now is, how is it happening to Denver? Do all/most of the teams they've played have some big unit/situational weakness that Denver has exploited and abused? If so, isn't that good coaching, to identify and exploit weaknesses? If that's the case, I wouldn't get too upset if I was a Denver fan (well, except I'd have to kick my own ass for becoming a Denver fan), because the lowering of ranking here would just be a side effect of considering accomplishments in context, which will always under-rank weakness exploiters. In the long run, would you rather have a team that consistenly identifies an opponent's weakness and exploits it, but is ranked lower in DVOA because of it, or one that gets a higher DVOA by attacking an opponent's biggest strength? If that's what's been going on in Denver, of course. I can't say for sure because I can't stand the Broncos and don't watch them very often. Screw them.

192
by Brad (not verified) :: Fri, 10/21/2005 - 1:50pm

I haven't read the whole thread, so someone else may have discussed this, but I think it's very interesting that the top two teams have such significant downward adjustments based on their past opponents. Right now, I'm not sure whether this adjustment will become even stronger (i.e., will the "easy" opponents in the past remain as easy) or will those other teams show themselves to be stronger than they look now (The teams will have quite a few opponents in common since the AFC South is playing the AFC North this year). I suppose I don't have anything of value to add to the discussion, but I wanted to point this out.

193
by Bj&ouml;rn (not verified) :: Fri, 10/21/2005 - 3:44pm

RE: #185

Perhaps I should rephrase. I will always enjoy reading balanced, intelligent and informative comments on Denver, but I am already sick of people who think that "the broncos r better than the'yre rating. so shut up." That gets old real fast.

Also, to anyone who says that they are still learning about DVOA, try finding the page on this website that explains DVOA in full.

Hint: That page is linked in the first few lines of the story. Durrrr.

194
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Fri, 10/21/2005 - 4:07pm

Aaron,

Let me start by apologizing. Didn't mean to seem critical of the system but trying to make sense of it.

Of course, I love the Broncos so questioning the system comes from being a Bronco fan.

However, I hope this ranking system isn't only for the few who regularly use it but rather for all who love to analyze the sport. So if someone new comes along with interest in your way of breaking the nunbers down, are they welcome to do so?

In my last posting, I asked a question about situational impact on the numbers. I see in the description of how the system analyzes each team but I am not sure I understand it completely. If you don't want to answer my question, then I will try to figure it out myself--just thought I would ask.

As a compliment, this system is by far the most detailed and complete system I have seen and therefore probably the best at predicting the best teams.

I love stats and I just found this website this week-so yes I am green to your system.

195
by Brian (not verified) :: Fri, 10/21/2005 - 4:33pm

#183 you said "It seems to me that winning at Jacksonville would count to most reasonable people as beating a good team on the road. The Bengals couldn’t do it (and neither they nor the Colts have yet to beat a “good team� anywhere)."

The Colts beat Jacksonville this year, although it was at home not on the road. Plus, people seem to forget that the Ravens were a much better team before the Colts put their starting QB out for an undeterminate amount of time. The Colts are now 3/6 in knocking out the starting QB of the other team for a significant amount of time in the game (although Leftwich came back to finish the game). Not sure what that stat means, but it's interesting.

196
by Aaron (not verified) :: Fri, 10/21/2005 - 4:35pm

Hey Bronco Fan, this site is absolutely open to everyone and I'm always striving to make my stats more accessible, which is why the FOX articles have all the explanation and jokes. I was just responding to the last line, "We just want to completely understand why, not tweak it for our team." That may be true of you and a few other folks who have e-mailed me, but it certainly is not true of the great majority. By the way, your questions will be answered in a mailbag later this afternoon.

197
by B (not verified) :: Fri, 10/21/2005 - 5:11pm

Re #195: Do you really think that having Kyle Boller as starting QB would make that much of a difference for the Ravens? I know Wright is pretty terrible, but Boller isn't any better.

198
by MCS (not verified) :: Fri, 10/21/2005 - 5:36pm

RE 195, 197:

The 3/6 thing is interesting though. Strange things are afoot at the Circle-K

199
by B (not verified) :: Fri, 10/21/2005 - 5:43pm

See MDS's commentary on Indy defensive line coach John Teerlinck:
http://www.footballoutsiders.com/ramblings.php?p=2971&cat=11

Of course, this doesn't explain the Bulger injury, cause he was hit by a linebacker. I'm not saying they did anything dirty, but Indy defenders do target the QB a lot.

200
by tunesmith (not verified) :: Fri, 10/21/2005 - 5:45pm

#191 - Trogdor, awesome comment. Pretty impressive you can see it from a Bronco fan's POV when you hate them so much. ;-) That's a really good point about exploiting matchups, too. I've been wondering about that. Hopefully Aaron will tell us if that's actually happening to any degree. If true, it's almost like beating a consistently mediocre team would count for more than beating a really good team by exploiting a glaring weakness.

201
by TMK (not verified) :: Fri, 10/21/2005 - 5:56pm

B, it may be easy to swipe Boller, but it's wrong. Up to and including the Indy game, Boller was 10 TD's and 6 INT's in his last 8 1/2 games. His team was also playing Indy very evenly up to that point, the Indy lead coming after a bizarre stretch of Raven playcalling in which they abandoned the run -- see MDS' excellent Ogden-Freeney piece from earlier this year.

Wright has been 5 TD's and 7 picks since he took over. I think it's pretty clear that Boller has been doing better, since Wright hasn't had a good game almost since his comeback against Seattle in 2003.

And somebody check on MDS. He might faint when he finds out I complimented something he wrote about the Ravens. ;)

202
by Jim (not verified) :: Fri, 10/21/2005 - 5:58pm

re 183
hey R.J. I said SOME, even bad teams can beat a good team in the NFL.,and as smeone beat me to it I believe the colts beat Jax so your point makes no sense to me other than the fact all you can see is orange.Even plummer can have an ok game once in a while but yor broncos are 25th in yards passing and 3rd in rushing,tou have played 2 games on the road (losing one to miami) and 4 at home wher most will agree you get more than just a home field advantage there.Wait till you play a team that is good against the run and can rush the passer a little and see what happens,the giants will probably do that sunday but I would not call them a real good team.Relax bronco fans maybe you are better than some people think after a few more weeks we will all know

203
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Sat, 10/22/2005 - 2:02am

Aaron,
Yes another comment regarding the Broncos. Sorry but trying to understand still.
The loss vs Miami seems to be a big reason they are not rated higher but lets recap some things about that game.
The Broncos were inside the 1 yd line 2 times and one time Plummer fumbled and it was run back for a TD (in the last minute of the game). The Secondary was depleted by the second half due to injuries and they were still in the ball game late in the 3rd. I know the score was 34-10 but most of it was late against a rookie secondary.
Why does this hurt the rating so much but the domination of JAX (their only other road game so far) not help just as much or more?
Also, do the stats not take into account: #1 how close the Broncos were to scoring on two different occasions (basically a 21 point turn around) and #2, the drastic injuries to the secondary? Is it possible to incorporate it?

Another question--How does a player DPAR influence a teams DVOA? or doesn't it at all?

204
by Kibbles (not verified) :: Sat, 10/22/2005 - 4:05am

Re #191: No, I understand theoretically how a disparity like that happens. What I DON'T understand is how it happens to the team with the 3rd toughest schedule in the entire NFL.

205
by admin :: Sat, 10/22/2005 - 11:29am

Numerous questions from this thread now addressed in FO Mailbag.

206
by R.J. (not verified) :: Sat, 10/22/2005 - 2:18pm

Re: #202
If beating "SOME good teams on the road" were necessary to have a high DVOA ranking, how are Cincy and Indy ranked where they are? (It’s a rhetorical question). Obviously it isn't that.
So what to make of Denver's current DVOA ranking? I see three possible explanations:
1. DVOA is a good measure of every team so the Broncos aren't as good as their W-L record.
2. DVOA is a good measure of most teams but Denver is doing something specific and perhaps unique that DVOA is missing.
3. DVOA is not a good measure of most teams and its ranking of Denver is an example of that.
In good faith I'm ruling out 3. I've already implied it’s probably 1 (post #177) but if it’s 2, I’d like to know why. Figuring that out seems precisely what a discussion thread about the current DVOA ranking is for. In any case, just beating “some good teams" doesn't seem relevant to the discussion.

207
by Walt Pohl (not verified) :: Sun, 10/23/2005 - 8:30pm

The Football Outsiders Comment Board Curse strikes! Denver fans doom their team to heartbreak.

The ironic thing is that I expect that Denver's DVOA ranking goes up this week...

208
by 2 cents (not verified) :: Mon, 10/24/2005 - 5:30am

Whats wrong with defenses going after the QB? Thats a big part of the defensive game. From the stated figures, the colts D line has been terrific under that coach! If a QB cant take a hit, he should go play golf. If a QB gets hits too often, get better OL players, or a better OL coach or maybe the QB is simply too stupid. Dont sissy up the game by suggesting non-physical football.

I... LOVE... QBs EATING DIRT! AND THE TWINS!!

209
by Freddie (not verified) :: Tue, 10/25/2005 - 5:01pm

ATL at 12
Denver at 13
At some point you just have to say the system is flawed. According to this system Miami should be better than ATL.
They play in 2 weeks. Anybody want to put there money where there mouth is???

Ridiculous.

210
by Bonnie Smith (not verified) :: Sat, 12/17/2005 - 9:43pm

Please tell me which NFL team went 16-0 and which year this was accomplished