Writers of Pro Football Prospectus 2008

Most Recent FO Features

ReplayIns14.jpg

» Scramble for the Ball: Getting it Right?

Instant replay review is one of the cornerstones of the modern NFL. The process and its myriad special rules have been internalized and constantly debated. Mike Kurtz wonders: is it worth it?

18 Dec 2007

Week 15 DVOA Ratings

by Aaron Schatz

The top teams in this year's DVOA ratings stayed the same week after week for a long time, but we're starting to see some movement. Not in the top spot, of course -- the Patriots are still the greatest team DVOA has ever measured, even though this week they barely put up a DVOA of 0%. Below the Patriots, however, things are shuffling around. This week's loss to Philadelphia drops Dallas into third place. The Cowboys' rating for this game ends up a lot lower than the score would indicate because a) they kept recovering their own fumbles and b) I have scored Brian Westbrook's final run as a touchdown for the purposes of Philadelphia's DVOA and Westbrook's own ratings.

A weird quirk in how the formulas work drops the Colts to fifth in weighted DVOA, behind Jacksonville, the same week they rise to second in total DVOA for the season. That's because the Colts' blowout of New Orleans has now dropped out of the weighted DVOA formula entirely, and it is their second-highest rated game of the season. The highest-rated game is the victory over Baltimore, so perhaps the Colts should ask to play on NBC every week.

Jacksonville's victory over Pittsburgh really epitomized the direction that each team is heading. If you want to see two strong trends, check out the week-to-week graphs for these two teams. By the way, just for fun, I added a thin line that represents last year's Jacksonville Jaguars through Week 15, just to show you the absurd difference in consistency between that team and this one.

More on the Pats vs. history and the Bengals vs. common sense after we run the tables...

* * * * *

Here are the Football Outsiders team efficiency ratings through 15 weeks of 2007, measured by our proprietary Defense-adjusted Value Over Average (DVOA) system that breaks down every single play and compares a team's performance to the league average based on situation in order to determine value over average. (Explained further here.) OFFENSE and DEFENSE DVOA are adjusted for opponent and consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS DVOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season.

To save people some time, we request that you please use the following format for all complaints:

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


TEAM TOTAL
DVOA
LAST
WEEK
WEIGHTED
VOA
RANK W-L OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
S.T.
DVOA
S.T.
RANK
1 NE 56.0% 1 50.6% 1 14-0 44.2% 1 -8.8% 5 2.9% 8
2 IND 35.4% 3 27.5% 5 12-2 27.2% 2 -13.2% 2 -5.0% 29
3 DAL 31.1% 2 29.1% 2 12-2 23.7% 3 -6.5% 9 0.9% 13
4 GB 26.9% 4 28.5% 3 12-2 20.4% 4 -1.1% 17 5.4% 5
5 JAC 23.2% 6 28.2% 4 10-4 18.6% 5 -2.0% 14 2.7% 10
6 TB 22.7% 7 22.9% 7 9-5 11.0% 9 -12.0% 3 -0.4% 16
7 SD 21.0% 8 25.5% 6 9-5 6.2% 12 -8.6% 7 6.2% 2
8 PIT 20.1% 5 10.2% 11 9-5 3.3% 15 -18.5% 1 -1.7% 17
9 SEA 14.8% 9 14.4% 9 9-5 3.9% 14 -8.7% 6 2.2% 12
10 MIN 12.2% 10 16.3% 8 8-6 4.8% 13 -4.6% 10 2.7% 9
11 PHI 7.7% 13 3.5% 13 6-8 8.6% 11 -3.4% 12 -4.4% 28
12 CLE 6.8% 11 13.6% 10 9-5 11.6% 8 11.0% 26 6.2% 3
13 TEN 2.6% 18 -4.3% 18 8-6 -5.9% 22 -11.1% 4 -2.5% 22
14 WAS 1.6% 15 1.2% 15 7-7 -2.7% 18 -3.9% 11 0.4% 14
15 CIN 0.4% 12 -0.4% 16 5-9 12.5% 6 12.2% 29 0.0% 15
16 DEN -1.2% 14 -4.0% 17 6-8 10.0% 10 8.0% 22 -3.1% 25
TEAM TOTAL
DVOA
LAST
WEEK
WEIGHTED
VOA
RANK W-L OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
S.T.
DVOA
S.T.
RANK
17 NO -2.5% 19 8.6% 12 7-7 12.4% 7 11.4% 28 -3.5% 26
18 NYG -3.4% 16 -4.5% 19 9-5 -2.9% 20 -1.4% 16 -1.9% 18
19 BUF -4.1% 17 2.2% 14 7-7 -5.1% 21 3.2% 19 4.1% 6
20 HOU -4.4% 23 -5.4% 21 7-7 2.9% 16 9.8% 24 2.5% 11
21 CHI -10.1% 21 -5.2% 20 5-9 -20.1% 31 -2.3% 13 7.7% 1
22 BAL -11.5% 20 -14.3% 24 4-10 -15.6% 26 -6.8% 8 -2.7% 23
23 ARI -11.6% 22 -12.9% 22 6-8 -2.8% 19 6.0% 20 -2.8% 24
24 DET -17.5% 24 -19.5% 25 6-8 -0.1% 17 11.1% 27 -6.3% 31
25 KC -19.8% 26 -23.4% 27 4-10 -19.3% 29 -1.9% 15 -2.4% 21
26 NYJ -20.2% 25 -14.0% 23 3-11 -11.2% 24 12.4% 30 3.4% 7
27 OAK -24.0% 28 -19.6% 26 4-10 -13.1% 25 6.6% 21 -4.3% 27
28 CAR -25.5% 29 -27.9% 28 6-8 -17.7% 28 1.6% 18 -6.2% 30
29 ATL -28.1% 27 -28.9% 29 3-11 -16.3% 27 9.8% 25 -1.9% 19
30 MIA -31.4% 30 -36.5% 31 1-13 -9.9% 23 19.3% 32 -2.2% 20
31 STL -34.8% 31 -30.4% 30 3-11 -20.0% 30 8.2% 23 -6.6% 32
32 SF -39.9% 32 -39.4% 32 4-10 -30.6% 32 14.9% 31 5.6% 4

  • 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.
  • 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 NON-ADJ
TOT VOA
ESTIM.
WINS
RANK PAST
SCHED
RANK FUTURE
SCHED
RANK VAR. RANK
1 NE 56.0% 14-0 56.0% 12.8 1 2.2% 10 -34.9% 26 14.2% 16
2 IND 35.4% 12-2 33.6% 11.3 2 2.1% 11 -1.7% 17 16.8% 12
3 DAL 31.1% 12-2 34.0% 10.2 4 -0.9% 18 -23.9% 21 18.2% 10
4 GB 26.9% 12-2 32.0% 10.3 3 -3.4% 26 -27.6% 24 9.4% 26
5 JAC 23.2% 10-4 18.8% 10.1 5 3.6% 6 -28.4% 25 6.0% 32
6 TB 22.7% 9-5 25.6% 9.5 6 -5.2% 29 -65.4% 32 21.1% 6
7 SD 21.0% 9-5 16.0% 8.6 7 3.2% 8 -25.2% 23 23.4% 4
8 PIT 20.1% 9-5 20.2% 8.2 10 -0.8% 17 -46.3% 30 25.8% 2
9 SEA 14.8% 9-5 22.9% 8.4 9 -10.2% 32 -39.6% 29 11.1% 22
10 MIN 12.2% 8-6 16.7% 8.5 8 -3.8% 27 0.4% 16 19.5% 8
11 PHI 7.7% 6-8 -0.8% 8.1 11 5.9% 2 -6.6% 19 19.1% 9
12 CLE 6.8% 9-5 8.1% 7.9 12 -2.8% 25 -39.5% 28 6.3% 31
13 TEN 2.6% 8-6 -0.6% 7.3 14 1.1% 13 15.3% 11 18.0% 11
14 WAS 1.6% 7-7 -5.5% 6.9 17 3.3% 7 43.3% 6 8.9% 28
15 CIN 0.4% 5-9 7.2% 7.6 13 -2.2% 23 -24.7% 22 9.3% 27
16 DEN -1.2% 6-8 2.2% 6.6 19 0.4% 14 33.2% 8 36.5% 1
TEAM TOTAL
DVOA
W-L NON-ADJ
TOT VOA
ESTIM.
WINS
RANK PAST
SCHED
RANK FUTURE
SCHED
RANK VAR. RANK
17 NO -2.5% 7-7 -2.0% 7.2 15 -5.1% 28 -2.4% 18 13.7% 17
18 NYG -3.4% 9-5 -2.5% 6.8 18 -1.8% 22 51.9% 4 7.5% 30
19 BUF -4.1% 7-7 -9.4% 7.0 16 5.3% 3 4.2% 14 12.8% 20
20 HOU -4.4% 7-7 -6.4% 6.5 20 -1.2% 19 58.7% 1 13.4% 18
21 CHI -10.1% 5-9 -11.4% 5.9 21 2.9% 9 24.4% 10 7.6% 29
22 BAL -11.5% 4-10 -15.6% 5.5 23 0.3% 15 34.9% 7 15.5% 13
23 ARI -11.6% 6-8 -4.5% 5.3 24 -6.0% 30 -62.9% 31 22.9% 5
24 DET -17.5% 6-8 -16.5% 5.9 22 5.0% 5 7.1% 13 20.2% 7
25 KC -19.8% 4-10 -25.4% 4.6 25 5.2% 4 -37.6% 27 24.0% 3
26 NYJ -20.2% 3-11 -23.8% 4.5 27 6.2% 1 -17.1% 20 10.5% 24
27 OAK -24.0% 4-10 -21.9% 3.9 29 0.0% 16 44.2% 5 15.2% 15
28 CAR -25.5% 6-8 -23.6% 4.6 26 -1.7% 21 53.7% 3 15.3% 14
29 ATL -28.1% 3-11 -26.5% 4.2 28 -1.3% 20 3.2% 15 11.2% 21
30 MIA -31.4% 1-13 -35.5% 3.4 30 2.1% 12 56.3% 2 13.1% 19
31 STL -34.8% 3-11 -30.6% 3.0 32 -2.8% 24 8.5% 12 9.5% 25
32 SF -39.9% 4-10 -32.9% 3.2 31 -6.8% 31 29.4% 9 10.7% 23

Best and Worst DVOA Ever Watch

BEST TOTAL DVOA
AFTER WEEK 15
  BEST OFFENSIVE DVOA
AFTER WEEK 15
  WORST TOTAL DVOA
AFTER WEEK 15
  WORST OFFENSIVE DVOA
AFTER WEEK 15
2007 NE 56.0%   2004 IND 44.9%   2005 SF -61.4%   2005 SF -44.2%
1999 STL 46.6%   2007 NE 44.2%   1999 CLE -48.4%   1997 NO -39.6%
2005 IND 41.5%   2002 KC 37.8%   2000 ARI -47.4%   2004 CHI -37.7%
2004 IND 40.2%   2006 IND 32.9%   2000 CLE -46.0%   2002 HOU -37.4%
2001 STL 39.4%   2000 STL 32.1%   2004 SF -45.4%   2006 OAK -35.3%
1996 GB 37.5%   2004 KC 31.5%   2000 CIN -45.0%   1999 ARI -33.3%
2002 TB 36.7%   2000 IND 30.8%   2003 ARI -43.0%   2000 CLE -32.3%
2004 PIT 36.4%   2005 IND 30.6%   2002 ARI -41.6%   2007 SF -30.6%
2007 IND 35.4%   1998 DEN 30.3%   1998 PHI -39.6%   2004 MIA -30.5%
2004 NE 35.3%   2003 KC 30.0%   2007 SF -39.9%   2000 SD -29.7%

The Patriots are still the best team ever, and the gap between them and the other great teams of the past is still pretty big despite this week's uninspiring performance. This is because of a weird historical trend featuring upsets and/or great teams winning squeakers over bad teams in Week 15. Remember that last week, the "best DVOA ratings through Week 14" table had eight teams over 40%. This week we only have four. 14 teams since 1996 put up a DVOA of 34% or more through Week 14, and 12 of those teams dropped between Week 14 and Week 15, including this year's Patriots, Cowboys, and Colts. The only teams which went up between Week 14 and Week 15 were the 2004 and 2005 Colts. Of course, that sounds strange because this was the week the 2005 Colts actually lost for the first time, but it was a close loss to a very good San Diego team, so their DVOA stayed basically the same.

Brad Smith Option

One last note on the Pats-Jets game. This somehow got left out of Audibles, but Mike Tanier drew up a "TDZ" diagram of the Brad Smith Pistol Option play. As Mike points out, "In the bottom right corner, you can see Chad Pennington getting ready to come in, trying to throw passes into the wind."

Congratulations to San Francisco, which looks now to be pretty safely out of the "worst DVOA of all-time" neighborhood. Their win over the Bengals once again brings up the question: How the hell do the Bengals have a positive DVOA rating? They've dropped to 15th this week, but still... we're talking about a 5-9 team where most of the indicators of "luck" don't seem to apply. The Bengals have a below-average schedule. Opponents have been average on kickoffs and above-average on field goals. On offense, the Bengals have recovered less than half their fumbles. The one "luck element" that seems to be an issue is recovering fumbles on defense, where the Bengals have recovered 14 of 17 opponent fumbles.

Cincinnati is 29th in defense, which seems reasonable, and exactly 0.0% on special teams, so that's not the issue. The problem seems to be offense. They don't seem to be terribly efficient, so how can they be 13th in points and ninth in yards but sixth in offensive DVOA?

The answer seems to be that the Bengals are more efficient than we are giving them credit for, particularly on first down. The Bengals gain at least 45 percent of needed yards on 47 percent of first downs, which is third-best in the league behind New England and New Orleans.

The Bengals also have a weird pattern when they get close to the goal line. The Bengals have the best offensive DVOA in the league in goal-to-go situations. However, they are 23rd in the zone I call "RED2," which consists of non-goal-to-go red zone situations. That means the Bengals get stuck settling for a lot of short field goals. Looking at Jim Armstrong's drive stats through Week 14 (I don't have an update for this week yet), the Bengals are 19th in points per red zone appearance but second in field goals per drive. They score a field goal on 19 percent of drives. (The only team higher is, oddly enough, the Jets and FO whipping boy Mike Nugent, at 19.5 percent.) The average Cincinnati field-goal attempt comes from the 14-yard line, closer than every team except Indianapolis, New England, and Jacksonville.

The trick for Cincinnati is to turn a lot of those short field goals into goal-line situations. Last year, the Bengals were third overall in the red zone, so this is really a one-year trend. If it reverses next year, the 2008 Bengals could really surprise people... with lots of much-closer shootout losses handed to the other team by the still-crappy Cincinnati defense. Oh well.

On to the housekeeping... All the team stats pages and individual stats pages are now updated. Loser League results and the DVOA Premium database are both now updated. The playoff odds page is now updated as well.

For (short) comments on every team, remember to look for DVOA on AOL, every Wednesday. (This will be linked on the FO Goes Mainstream page.)

Posted by: Aaron Schatz on 18 Dec 2007

116 comments, Last at 21 Dec 2007, 6:07pm by B

Comments

1
by Otis Taylor 89 (not verified) :: Tue, 12/18/2007 - 5:30pm

Christmas has come early - Week 15 DVOA out before 6:00 pm ET!

2
by iowapatsfan (not verified) :: Tue, 12/18/2007 - 5:32pm

Why on the westbrook thing? It doesn't make sense to me, and seems like it leads to a slippery slope of things that probably would have happened.

3
by Bobman (not verified) :: Tue, 12/18/2007 - 5:36pm

The key takeaway for me is Indy's ST are now ranked in the 20s! (albeit #29)

Kneel before us... you... other three teams! bwa-ha-ha-ha!

The Pennington diagram above is priceless.

This is where I start hating and ignoring DVOA and pretty much all stats. For leading teams, the last two games are pretty much meaningless--even with a perfect season on the line. Sigh. Is it playoffs yet?

4
by Lee Gibbons (not verified) :: Tue, 12/18/2007 - 5:40pm

Re 2: I agree. If you are going to do this then you should probably take away Westbrook's late touchdown against the Redskins as they let him score in order to get the ball back.

It makes the ratings more subjective and once you start changing one thing, you have to change others or else the ratings lose meaning.

5
by Bobman (not verified) :: Tue, 12/18/2007 - 5:40pm

Oh, and thanks for the Jags/Steelers charts. I love those things.

Ya know, the 07 Jags kind of remind me of somebody... can't think who... oh yeah, the 05 Steelers, a bit. Just a bit.

6
by Lee Gibbons (not verified) :: Tue, 12/18/2007 - 5:41pm

Re 4: I meant the 2nd Redskins-Eagles game this year for reference.

7
by rageon (not verified) :: Tue, 12/18/2007 - 5:43pm

In re: #2 -- true, it would be nice if everything could be done without a subjective element, but it's hard to not do so when doing so clearly improves the accuracy of the system. For all intents and purposes, Westbrook could have scored a TD had he chosen to, and should probably be rewarded as such. It's such a unique situation that it probably doesn't put the system on too much of a slippery slope. Really, how many times can one honestly look back and remember something similar happening. (And no, running backwards on Techmo bowl in order to inflate your stats doesn't count.)

I think the way "garbage time" plays are handled is much more likely to lead to a slippery slope that clearly flukey plays like the Westbrook "Smartest Play Ever!!", though I certainly believe they need to be accounted for as well.

8
by kevinNYC (not verified) :: Tue, 12/18/2007 - 5:44pm

Ugh... the Giants run defense took a beating the past 2 weeks and has brought the rest of the team down. Oddly enough, the poor offense hasn't moved at all the past month.

Is it me of has the Philly defense moved up dramatically the past few weeks?

9
by AndyE (not verified) :: Tue, 12/18/2007 - 5:47pm

iowa2: probably because dvoa doesn't (yet) have a metric for (made the play that guaranteed victory). Remember last year's game non-ending interception by SD against the Pats? A kneeldown wins the game. By definition, that play should have the highest possible value in a measuring system.

10
by jaredtaskin (not verified) :: Tue, 12/18/2007 - 5:48pm

3. leaving it as is suggests that the cowboys stopped him, which also isn't what happened.

11
by Will Allen (not verified) :: Tue, 12/18/2007 - 5:48pm

Boy, maybe the Bills will pack it in now that they have been eliminated, but I sure don't like the Giants as favorites in Buffalo this Sunday.

12
by panthersnbraves (not verified) :: Tue, 12/18/2007 - 5:49pm

Realizing that they have the 3rd hardest remaining schedule, how exactly are the Panthers still in the playoff hunt?

Does it require a bizzare series of tie games and Elvis coming back?

13
by Tom (not verified) :: Tue, 12/18/2007 - 5:49pm

It seems like DVOA needs an adjustment for plays near the end of the game where a first down guarantees the end of the game.

14
by mawbrew (not verified) :: Tue, 12/18/2007 - 5:51pm

Re: 7

I can understand the decision to treat the play as a Westbrook TD, but don't you have to count it as a TD against the Dallas defense ratings then too?

I can't see why it would count for the Eagles but not against the Cowboys.

15
by AndyE (not verified) :: Tue, 12/18/2007 - 5:51pm

Future schedule currently shows the sum of opponent DVOA, not the average.

16
by White Rose Duelist (not verified) :: Tue, 12/18/2007 - 5:52pm

In regards to the Westbook situation - there's no way to know what happens against Washington if they played it straight. He may have scored anyway or stopped in the backfield, or anything in between. In this case, there is no doubt that he would have scored a TD if he had not chosen to stop. I agree that there is a potential for problems, but as long as the adjustments are kept to certainties, things should work out fine.

17
by Independent George (not verified) :: Tue, 12/18/2007 - 5:53pm

#2, 4 - I don't see it as being any different than not scoring an incomplete on a spike, or a carry on a kneel-down. On the other hand, I thought that DVOA didn't care much about that last foot anyway? What's the actual difference in scoring?

Also, as ridiculous as the Brad Smith/option series was, I do give Mangenius credit for trying something different; isn't one of the recurring criticisms around here that coaches are too risk averse to try something wacky? They're playing a historically great team, on the road - what do they have to lose?

18
by zip (not verified) :: Tue, 12/18/2007 - 5:53pm

#12

Just because they aren't mathematically eliminated doesn't mean they are really "in the hunt"

19
by GBS (not verified) :: Tue, 12/18/2007 - 5:54pm

I have a question about the playoff odds report. Does it take the proper tiebreakers into account, or are two teams that finish with the same record and fail to win a division assigned equal chances of making the playoffs?

20
by Will Allen (not verified) :: Tue, 12/18/2007 - 5:55pm

Bobman, those of us who are not fans of teams which have won at least 12 games five years in a row find complaints about stat keeping after divisions and byes have been clinched, and asking, "Is it playoffs yet" to be rather Yankeefannish, if ya' get my drift.......

21
by Ken (not verified) :: Tue, 12/18/2007 - 5:55pm

when do you think weather gets built into DVOA? obviously this hurts the teams playing in monsoons and blizzaards

22
by Holy smokes! (not verified) :: Tue, 12/18/2007 - 5:59pm

Great diagram..
but did anyone notice Pennington threw a better game than Brady? --in the same elements ..?

23
by GBS (not verified) :: Tue, 12/18/2007 - 6:00pm

Will, I'm afraid most of us Colts' fans relate more to Atlanta Braves' fans than Yankees' fans.

24
by Will Allen (not verified) :: Tue, 12/18/2007 - 6:01pm

GBS, I'm pretty sure tiebreakers are factored. The Giants had a 97% chance to be a wild card entry last Tuesday. It'll be interesting to see what that drops to today.

25
by pete (not verified) :: Tue, 12/18/2007 - 6:04pm

re 22: If chad pennington got to play behind a ridiculous offensive line and in the same system for 8 years, he'd be Tom Brady. Conversely, if Brady played behind a revolving door of an offensive line and in what, 4 different systems, he'd be Chad Pennington. Can't say i'm too surprised

26
by Charles the Philly Homer (not verified) :: Tue, 12/18/2007 - 6:05pm

8.

Subjectively, I'm not surprised. The defense has been making strides all year, and the secondary injury circus of earlier weeks appears to be over at long last.

27
by James, London (not verified) :: Tue, 12/18/2007 - 6:07pm

"Yankeefannish"?

Will, them's fightin' words.

28
by Otis Taylor 89 (not verified) :: Tue, 12/18/2007 - 6:08pm

Two things:
-Strength of schedule plays a huge role in what type of season you are going to have if you are in the middle area of teams. The Eagles #2 ranked schedule gave them long odds on making the playoffs with a QB coming back from a major injury - if they had a SEA or TB schedule they would probably be the top WC. PIT schedule went from #29 or 30 to 10 in two weeks;
-Staying on the schedule line, I would be worried if I were GB,SEA,CLE,TB and MIN as they are playoff teams that have not played a playoff regular season schedule. On the flip, JAC and SD have played playoff type schedules.

29
by Mark (not verified) :: Tue, 12/18/2007 - 6:10pm

NE is one of the greatest teams of all time as everyone already knows, but so are the Colts. To think that they are listed in the greatest DVOA ever, and yet they've been without their all-pro WR most of the year, have had to use a makeshift offensive line, lost their third and fourth receiving options for a couple of games...on defense they lost one of their starting DT and their all-pro DE to IR, have missed a dozen or so starts at the LB position, are now playing without their other starting DT and DE, as well as a safety...

AND THEY ARE STILL ONE OF THE GREATEST DVOA EVER. Imagine if they get healthier how much of a battle the AFCCG will be next month.

30
by Pat (not verified) :: Tue, 12/18/2007 - 6:11pm

Why on the westbrook thing? It doesn’t make sense to me, and seems like it leads to a slippery slope of things that probably would have happened.

It's not a "probably would have happened." It's "what did happen."

That play guaranteed victory for the Eagles. It was worth more than a TD, not less, but that's difficult to assess. At the least, however, it can be scored the same as a TD.

Re 2: I agree. If you are going to do this then you should probably take away Westbrook’s late touchdown against the Redskins as they let him score in order to get the ball back.

No - the reason you change Westbrook's play here to a TD is that a run to the 1 is, in general, worth less than a TD, but the play that happened is worth more than a TD. You change it to a TD so you're at least closer to the proper value.

In the Washington game, that play was still worth a TD. If Westbrook had fallen down at the 1, that play should've been valued at least as a TD too.

Basically, the idea is to treat any offensive play which results in a first down which then results in a sequence of kneel-downs which ends the game as at least worth a TD. It's not really playing "what might have been." It's saying "what was this play worth in terms of value?"

The Westbrook play versus the Redskins two weeks ago doesn't change, because it still was a TD - even though the Redskins got the ball back afterwards, their chances of winning were diminished by that play as much as any other TD would be. Just not as much as if he would've fallen down at the 1.

31
by Johonny (not verified) :: Tue, 12/18/2007 - 6:13pm

Chad Pennington played well last year and has played ok this season. I think he's in the classic position of being on a bad team where the organization is fixated on their players weaknesses instead of their strengths. Aaron's 2006 and 2007 numbers argue that Chad deserves a chance to start somewhere in 2008.

32
by Devin McCullen (not verified) :: Tue, 12/18/2007 - 6:32pm

It shouldn't surprise anybody who read Audibles that I agree with Aaron on the Westbrook thing. I don't think it's too far removed from not counting INTs on Hail Marys or omitting kneeldowns and spikes. After all, isn't Westbrook's ability to get into the end zone an indicator of his talent, whether he actually got there or not?

But I would tend to agree that you also shouldn't count it when the defense lets the offense score. Which then raises the question of what to do if that actually works, and the trailing team ties the game up. That I don't have an answer for.

33
by Devin McCullen (not verified) :: Tue, 12/18/2007 - 6:40pm

Actually, after further review (i.e. after reading Pat's post), I'll just withdraw my comments on how to handle the defense giving the offense a free touchdown. Maybe DVOA should be adjusted in some way, but I don't have any particular thought about how you'd do it.

34
by Pat (not verified) :: Tue, 12/18/2007 - 6:43pm

But I would tend to agree that you also shouldn’t count it when the defense lets the offense score. Which then raises the question of what to do if that actually works, and the trailing team ties the game up. That I don’t have an answer for.

Why would you need to do anything? Their DVOA would reflect a team that allowed a touchdown, then drove down the field to score a touchdown.

Keep in mind - defenses don't let teams score "just because" - they let them score because they've already failed - i.e., they gave up a first down when they really needed to prevent it. (In theory, the defense could also let someone score when they haven't failed at all, but that would require the offense being idiotic, because in all those situations, the offense could simply kneel out the clock.)

So a defense letting someone score is a failure for the defense, and a bad failure. Maybe it's not quite as bad a failure as a TD is during 'normal play,' but it's got to be close.

A good analogy is a DB tackling a receiver before he catches the ball when he's beat deep with no safety help over the top. Yes, it's a smart play to save a touchdown, but it's still a failure.

35
by admin :: Tue, 12/18/2007 - 6:44pm

I love that the main subject of the discussion so far is one play that was probably worth, I don't know, 0.1% DVOA to two teams, one of which isn't going to the playoffs.

Playoff odds now updated.

For those wondering: Yes, I need to work on a weather adjustment. I've talked about it a lot recently, just not in this article. I was going to do some introductory research as part of this week's ESPN MNF article until I realized the game is actually in San Diego...

36
by burbman (not verified) :: Tue, 12/18/2007 - 6:46pm

How did Shaun Alexander earn a 16 in LL? Did he get the point he earned in play, plus the penalty?

37
by Independent George (not verified) :: Tue, 12/18/2007 - 6:49pm

#25 - Well, Brady has also been battling shoulder problems for his entire career.

38
by David (not verified) :: Tue, 12/18/2007 - 6:57pm

36: Yes. The penalty is in addition to the player's fantasy score for the week, not instead of it.

Oddly enough, that actually worked out in my favor with Kellen Clemens this week, since even though he took the penalty he came out to 13 points instead of 15 because of a turnover.

39
by Will Allen (not verified) :: Tue, 12/18/2007 - 6:58pm

Intuitively, the Vikings' and the Giants' chances of making the playoffs seem too high, and the Saints' and Redskins' chances seem too low.

40
by Fat Tony (not verified) :: Tue, 12/18/2007 - 6:59pm

Pennington’s conventional and FO stats are better than Brady's this week in large part because of game specific circumstances. Playing most of the game with a two score lead in bad weather, Belichick chose to minimize risk and run clock as much as possible rather than seeking to maximize scoring opportunities. On the other hand, the Jets had to take more risks and, to his credit, Pennington played well enough to keep the game close. With that said, he played roughly ninety percent of NY's snaps versus Brady’s one hundred percent and each offense managed zero TDS and two FG attempts apiece (Nugent 1-2, Gostkowski 2-2). The weather was the story of this game, not Brady v. Pennington. If the NYJ had jumped to a ten point lead, I think it’s highly likely each side would have swapped strategies and Pennington and Brady, roughly speaking, would have swapped stat lines.

As far as the two of them swapping careers, that’s a very interesting if ultimately unanswerable question, but I’m inclined to agree with the premise that a healthy Pennington could have replaced Brady and lead NE to at least one SB win, which is not intended in any way as a knock on TB.

41
by Taollan (not verified) :: Tue, 12/18/2007 - 7:04pm

I love Tampa Bay, but can someone explain to me how dominating a bad Atlanta team boosts them from a #6 defense to a #3 defense in a single week? Don't get me wrong, I think TB has a #3 defense in reality.. but I don't see how a single game against a bad opponent shifts them so much this late in the season.

42
by Jim (not verified) :: Tue, 12/18/2007 - 7:04pm

Aaron -

In the main article, you wrote "The Cowboys’ rating for this game ends up a lot lower than the score would indicate because [of your scoring of Westbrook's non-TD and one other factor]." In post 35, you say that it made essentially no difference. Which is it?

43
by ammek (not verified) :: Tue, 12/18/2007 - 7:06pm

Interesting that the FO preseason picks of Green Bay, Tampa Bay and Jacksonville (and the Pariots, of course) have materialized, but not necessarily for the reasons expected. The Green Bay defense was supposed to be the motor of improvement, but has now slipped into the bottom half of the league - making it the worst of any only top ten team. And I think most people would have assumed the Jags would rank 5th in defense and 14th in offense, and not the other way round.

Also, I argued against the preseason ranking of GB special teams as worst in the league on the basis that the personnel had changed in every significant position bar one (punter, and that's a second-year starter) and that Ted Thompson's strategy of accumulating lower-round picks was specifically designed to address and rectify the special teams problem. Well, the Packers now rank fifth on special teams, and I wonder if for next year's book it wouldn't be worth adding a few words along these lines if any similar scenario occurs (I don't recommend changing the numbers on such a subjective point).

By the way, where did the Packers/Rams special teams dvoas rank in terms of best/worst-of-the-season?

44
by throughthelookingglass (not verified) :: Tue, 12/18/2007 - 7:08pm

Minus the San Diego and KC games (when everyone was hurt), Indy has a DVOA of roughly 49%. Over their last 6 "relatively healthy games", their DVOA is about 39%.

Jacksonville has a DVOA of roughly 42% in the last 6 weeks.

How about the Packers the last 6 weeks? Around 39%.

The last 6 weeks, New England's DVOA has been roughly 38%.

I'm not saying that NE isn't the best team, because they are, but the playoffs will be very interesting. And I think these are the teams that NE will face in the playoffs.

45
by CA (not verified) :: Tue, 12/18/2007 - 7:09pm

Question for Aaron or someone else in the know (I know Aaron often skips the comments, so I may have to email this question to him directly): I should probably already know the answer to this question, but does DVOA directly reward good clock management? Does DVOA reward the offense / punish the defense for a play that runs the clock when the offense's team is leading late in a game or reward the defense / punish the offense for a play that stops the clock when the defense's team is trailing late in a game? Let's say the offense's team is ahead by a TD with a few minutes left in the 4th quarter. Is a running play on 2nd and 6 that gains 4 yards and keeps the clock running worth more than a running play on 2nd and 6 that gains 4 yards but stops the clock because the back stepped out of bounds? It is in real life, obviously, but does DVOA capture that directly?

46
by Mark (not verified) :: Tue, 12/18/2007 - 7:17pm

I think it's funny how the one ranking lists Jax ahead of Indy when the Colts have handily beaten them both times this year. Shows you what stats are good for.

47
by Bobman (not verified) :: Tue, 12/18/2007 - 7:23pm

#20 Will, yikes, you are right and I am guilty on both counts. Though, with the Yankees recently, I have been much happier with the regular season than the post season. (whine)

Plus, my NL team, the Pirates, should be considered a pathetically redeeming quality. 3 good years out of the last 25.

Sorry to have sounded so snobby. But it really is the preseason all over again right now (and you know THAT feeling) for fans of at least 3 teams. I'll assume NE fans are charged up for 16-0; I sort of am too. Watching the final 8 seeds fill up, like last night's game, is good (congrats, BTW). Actual games are always a good thing.

But in terms of statistical looks at how things change at the top over the next two weeks, it's pretty meaningless. My baseline assumption is that the Colts metrics will all slip because they'll be playing backups half the time, so not only do box scores mean "nothing" to me in terms of assessing my team's quality, they have the potential to irritate me since they make my team look less competent in the eyes of the TV/Internet talking heads, who blather on and on and who convince the know-nothing fans on the street, then some guy stops me on the elevator at work to quote Pete Prisco or Clark Judge about how Indy's fallen off the last two weeks or Michael Wilbon about how the Jags are the second best team in football.... Must stop. Innocent victims....

Clearly, this has a lot more to do with my capacity to remain sane than the NFL's ability to remain commpellng in Weeks 16 and 17. For the next two weeks, I'm a staunch Pats, Browns, and Vikes fan--even to the point of poring over their FO stats and those of their opponents. It'll give me something constructive to do (when shirking my actual paying job duties, of course). Go Teams!

48
by B (not verified) :: Tue, 12/18/2007 - 7:24pm

42: The "one other factor" is the Cowboys three fumble recoveries on offense, which affected the DVOA more than the Westbrook play.
45: As far as I know, both plays would be equal by DVOA.

49
by footballprofessor (not verified) :: Tue, 12/18/2007 - 7:26pm

Playoff Odds are clearly bogus because Conference Dominance is at play. WTM is way better than this. NFC teams should not be listed as having more than a .1% chance of winning the Super Bowl because the AFC pwns the NFC! I like beans!

50
by aih (not verified) :: Tue, 12/18/2007 - 7:28pm

Am I the only one who finds it surprising that the odds of the Pats going undefeated are still below 70%? I realize that for most of the season, with all those games left, the media was overestimating their chances. But now! Is there really a 25-30% chance that they will lose to the Giants?

51
by Sam (not verified) :: Tue, 12/18/2007 - 7:36pm

42/Jim:
The play may have been significant to the game's DVOA, but not the teams' overall DVOA for the season - possible that you and Aaron are considering two different number here.

52
by MC2 (not verified) :: Tue, 12/18/2007 - 7:37pm

The diagram of Pennington made me LOL.

53
by TGT (not verified) :: Tue, 12/18/2007 - 7:38pm

I have a question about the luck elements around Cincinnati's DVOA. Recovering less than half their offensive fumbles and the opponents being above average on field goals seems to match the Bengal's record being worse than the DVOA. Since there's no mention of how far they are from the norms, I see no reason to assume they are near average. I'm actually drawn to believe the contrary.

54
by PatsSF (not verified) :: Tue, 12/18/2007 - 7:51pm

#29 -- Even more amazing that three of the best DVOA performances all come from 2004. Hell of a year. (BTW, this is me handing out credit to whoever deserves it if this has been noted before).

55
by MJK (not verified) :: Tue, 12/18/2007 - 8:01pm

It seems like DVOA needs an adjustment for plays near the end of the game where a first down guarantees the end of the game

I've often thought that DVOA needs some way of valuing plays more highly that bleed clock in an obvious clock-bleeding situation (or that prevent clock bleeding in a little time situation).

For example, when clinging to a small lead late in the game, with the other team low on timeouts, a 1 yard run on 1st and 10 is probably more valuable than a 5 yard pass where the reciever steps out of bounds, even though DVOA wouldn't see it that way. And a 0 yard run is MUCH more valuable than an incomplete pass, even though DVOA values them equally.

Similarly, when a team is desparately driving in a two minute drill and is out of timeouts, an incompletion on 1st down is actually better than a seven yard pass that gets tackled in bounds, and a 10-yard sideline route that stops the clock is preferable to a 15-yard pass over the middle.

56
by MJK (not verified) :: Tue, 12/18/2007 - 8:03pm

Agree on Pennington's and Brady's statline. Once one team got the lead in those situations, its QB was going to play it safe and focus on running out the clock, whereas the other QB had to try to move the ball. Combine it with the fact that the Jets defense was playing more aggressively (again because the Pats had the lead) and I think that explains it.

I've always liked Pennington as a QB, but Brady definitely has the stronger arm, so I'm not sure Pennington could do what Brady has done in New England. However, I do think that Pennington throws better on the run than Brady--the Jets used to love the naked bootleg...

57
by DolFan 316 (not verified) :: Tue, 12/18/2007 - 8:20pm

So in other words, if the Fins hadn't beaten the Ravens then 0-16 would've been a certainty since the Bengals are much better than their record. Glad they won then.

58
by Charger Jeff (not verified) :: Tue, 12/18/2007 - 8:28pm

May I just point out, as an unabashed homer, how guldurned satisfying it is to have the Bolts pass D ranked in the top 5?? Ahhhhh ...

Oh and according to the Pass D vs. #2 WRs, Antonio Cromartie must be the greatest CB ever.

59
by admin :: Tue, 12/18/2007 - 8:42pm

Individual stats and ALY pages now updated.

By the way, I wrote about this a couple weeks ago, but actually FOUR of the top DVOA ratings of all-time were in 2004. The 2004 Eagles are 11th through Week 15; their rating gradually dropped at the end as they sat their starters before the playoffs.

60
by Raiderjoe (not verified) :: Tue, 12/18/2007 - 9:17pm

Raiders should be higher than Jets and Chiefs. Raiders beat Chiefs/ and have more wins than Jets. Your computers broken again

61
by Charger Jeff (not verified) :: Tue, 12/18/2007 - 9:21pm

Ignorance is NEAT!

62
by jimm (not verified) :: Tue, 12/18/2007 - 9:29pm

Sometimes the individual ratings on this site don't seem to add up. Take Minnesota for instance:

T. Jackson - 37th in DPAR

No wide receiver in the top 45

TE 30th

Off Line Adj yards - 15th, pass blocking 29th

RB - Peterson 4th

But the Vikings rank 10th in DVOA for Offence and 5th in the NFL in yards per play.

The numbers just don't add up. How can your offensive line be worse than average? Your QB and WR,TE absolutely terrible, and your RB very good, but only 4th?

63
by Bobman (not verified) :: Tue, 12/18/2007 - 9:29pm

I think you mean neet!

64
by jimm (not verified) :: Tue, 12/18/2007 - 9:31pm

- to finish the pt above - the individual ratings on this site suggest that everyone other than Peterson on the Viking offence is below average or downright awful, yet they rank 10th in DVOA and 5th in the league in yards per play.

Sorry that just doesn't add up.

65
by Waverly (not verified) :: Tue, 12/18/2007 - 9:42pm

OK, I've asked this before, and it's even more puzzling this week:

How can Tampa Bay's future schedule DVOA be -65.4% when there aren't any teams with a DVOA less than -40% ?

(FUTURE SCHEDULE lists average DVOA of opponents still left to play this season....)

66
by Seth Burn (not verified) :: Tue, 12/18/2007 - 9:47pm

Can someone please give me an explanation for how weighted DVOA is calculated? I understand how DVOA works but I don't understand how the weighting system works. Thank you.

67
by Will Allen (not verified) :: Tue, 12/18/2007 - 10:18pm

jimm, you really have to put the individual and line stats in context. For instance, the 29th rank for the Vikings in pass blocking is really saying the Vikings' offensive line, while blocking for Tavaris Jackson, Brooks Bollinger, Kelly Holcomb, and their merry band of receivers, ranks behind other offensive lines who have much better qbs and receivers.

Strangely enough, the Cowboys pass blocking rank started climbing quickly last year after Bledsoe was benched. Total team efficiency is much more easily interpreted with DVOA. Next comes offensive and defensive efficiency, although even there some context needs to be waded through. When it gets to individual or line rankings, one really has to work hard to interpret what is being measured.

68
by Raiderjoe (not verified) :: Tue, 12/18/2007 - 10:20pm

re61

You porbably talking to me since you make post after mine. You are ignorant one saying A Cromartie best cornerback ever. Maybe you too young to have heard about Mike Haynes and Willie Brown.

69
by Michael (not verified) :: Tue, 12/18/2007 - 10:21pm

" If chad pennington got to play behind a ridiculous offensive line and in the same system for 8 years, he’d be Tom Brady. Conversely, if Brady played behind a revolving door of an offensive line and in what, 4 different systems, he’d be Chad Pennington.""

Am I the only one who read this?

70
by Lyford (not verified) :: Tue, 12/18/2007 - 10:41pm

"Remember last year’s game non-ending interception by SD against the Pats? A kneeldown wins the game."

No matter how many times people say that, it just isn't true. Mcree made the interception, and then fumbled at the 41 with 6:16 left. Less than two minutes later, San Diego had the ball back, albeit in a game that was tied rather than with an 8-point lead. But they went 3 and out, punted, and the Patriots immediately marched down the field to the San Diego 15. At that point, they ran clock, burned San Diego's timeouts, and kicked the game-winning field goal.

But if Mcree hadn't fumbled, there's no guarantee whatsoever that the Patriots don't force a 3 and out, and march down the field to the tying score. An 8 point lead with 6 minutes left and the ball is better than an 8 point lead with 6 minutes left and without the ball, but it isn't a guaranteed win. There were over 6 minutes left, the Patriots had two timeouts left, and the Chargers had done almost nothing offensively in the second half. Somehow the myth has sprung up that the fumbled interception could have sealed the deal, and it just is not true.

71
by mm (not verified) :: Tue, 12/18/2007 - 10:42pm

66-dvoa for the season simply adds up how a teams done since week 1 with no adjustments

Weighted dvoa adds the most recent weeks at full value but reduces the value of previous weeks. What happened in a game 6 weeks ago tells you less about how a team is playing right now than the game they played last week, so the game 6 weeks ago doesn't count as much.

Weighted dvoa is better for telling you how teams are playing at this point in the season, and who should be favored if two teams play next week; total dvoa gives you a better idea of how teams have done over the whole season.

72
by Fat Tony (not verified) :: Tue, 12/18/2007 - 10:47pm

Off topic but pro-football reference has finished a major upgrade (now with mega hyperlinks like pro-baseball reference and basketball-reference) and it is now amazingly awesome. Not sure of the relationship between FO and that site but it's so good now it might deserve an XP. (link in name)

73
by Chris M (not verified) :: Tue, 12/18/2007 - 10:51pm

If the drop in the Colts system is a "weird quirk," well then maybe that shouldn't be in the system.

Not that I disagree with the concept of weighted DVOA - I actually think you should defend it more strongly. The Colts honestly have slipped a bit in the past month (primarily due to injuries, of course). That win over the Saints doesn't tell us a ton about the team as currently constituted.

74
by mm (not verified) :: Tue, 12/18/2007 - 10:55pm

"That win over the Saints doesn’t tell us a ton about the team as currently constituted."

which is why it isn't counted in this week's weighted dvoa

75
by Chris M (not verified) :: Tue, 12/18/2007 - 11:01pm

74: Agreed. But Aaron was calling it a "weird quirk" as if it was some glitch in the system, or some error. But I actually think weighted DVOA is basically correct regarding the Colts.

76
by Raiderjoe (not verified) :: Tue, 12/18/2007 - 11:33pm

Still mad about rankings.
I don't need quirky computer to tell me Chiefs better than Raiders. I know what my eyes tell me and I know the Raiders are better than the Chiefs.
eyes > computer

77
by mm (not verified) :: Tue, 12/18/2007 - 11:57pm

jimm- I'd like to expand on what Will said about evaluating all the numbers in context.

About 4 years ago the Saints had Aaron Brooks at Quarterback and an offensive line that was poor at run blocking and mediocre in pass blocking. Whenever Deuce McAllister was in the game defenses would play the run first; quite frequently at least one defender would get into the backfield before McAllister could reach the line of scrimmage. But McAllister would drive forward or sidestep defenders, turning what would be 2 yard losses into 2 yard gains. He turned many plays that should have been losses or no gain into small gains.

This lead the Saints to have a poor running DVOA, but gave them decent adjusted line numbers. The poor running DVOA was totally justified; they weren't getting good yardage out of the running game. The adjusted line numbers; however, didn't give the correct impression at all. The line numbers give the line 'credit' for having few runs for loss or no gain. However, the reason the Saints had few losses was because of McAllister, not because of the line.

The staff here wasn't watching a lot of Saints games (this was before they started charting every game), so they saw the Saints had a decent number for the Offensive line and a poor running DVOA, and they called McAllister the most overrated back in the game. Repeatedly. I tried to point out where the hole was in their analysis but they didn't really pay attention.

Last year the Saints got an overhauled offensive line and a QB that forced defenses to respect the pass. All of a sudden when McAllister got a handoff he could run to the line of scrimmage (or beyond!) before having to worry about defenders. He put up good numbers and the staff were puzzled about how he 'turned it around'.

The truth is he ran pretty much the same, but the circumstances were quite different. The problem wasn't the numbers, but they needed someone who was watching the games to interpret them correctly. The Offensive line ranking, in particular, is useful in some instances but is nowhere near perfect.

I haven't watched any Vikings game closely. If I were to guess, though, I'd say your young QB and young RB are making some mistakes in the backfield that are being attributed to the numbers for your Offensive line, bringing down their ranking. Of course, their running number is also hurt because they're facing defenses that are bring 8 men to the line of scrimmage all the time. This increases the probability that the defense will get a few more stops for little or no gain, while increasing the chance that a few short runs will turn into big gains (because there's no safety if you get downfield). Note your offensive line doesn't get credit for long runs in the Adjusted line number.

78
by Ben (not verified) :: Tue, 12/18/2007 - 11:58pm

re:76 If you look at the weighted VOA (which is a measure of how the teams are playing recently) you'll see the Raiders ahead of the Chiefs...

79
by Pat (not verified) :: Wed, 12/19/2007 - 12:03am

I don’t need quirky computer to tell me Chiefs better than Raiders. I know what my eyes tell me and I know the Raiders are better than the Chiefs.

Didn't your eyes also tell you that the Raiders weren't going to suck this year?

80
by Raiderjoe (not verified) :: Wed, 12/19/2007 - 12:07am

so Raiders have a f ew setbacks in this season. all for the better, good learning experience for 2008 when Raiders become great. Jamarcus Russell playing in games now to get feet wet for 2008. Raiders going to get good draft pick again, maybe take top WR in draft, trade down and get more picks

81
by Athelas (not verified) :: Wed, 12/19/2007 - 12:18am

Michael #69-
When you don't feed the trolls, they wither away.
Just one more thing I love about FO.

82
by Chad Gerson (not verified) :: Wed, 12/19/2007 - 12:21am

Raiderjoe, I make you a gift of this quotation, so you may put it more elegantly next time:

"Who are you going to believe, me or your own eyes?"
--Grouch Marx

83
by Karl Cuba (not verified) :: Wed, 12/19/2007 - 12:34am

81: I agree with your idea about not feeding the trolls but the Brady/Pennington comment wasn't that absurd. I thought a lot of Pennington before he hurt his shoulder and while pete's comment is a little extreme, I can see where he's coming from. The denizens of this website are too unwilling to separate players and their surroundings.

84
by Andrew (not verified) :: Wed, 12/19/2007 - 12:49am

jimm #62:

Because a lot of Peterson's yards have come on huge touchdown runs which don't get credited to the line.

85
by QB (not verified) :: Wed, 12/19/2007 - 1:08am

55 - While running out the clock, a 0 yard run may be more valuable than an incomplete pass, BUT it is not an indicator that the offense is any better than an offense who throws an incompletion. The offense certainly isn't trying to run for 0 yards, and the failure is just as bad as any other 0 yard run. Unless we want to start measuring and evaluating each team's playcalling, there's no reason to give extra credit for a play like that.

86
by mmm... sacrilicious (not verified) :: Wed, 12/19/2007 - 1:25am

#55,85: Besides the moral issue of whether DVOA should measure clock management, doing that may represent a physical impossibility. First, it may be tough to incorporate the time a play takes from the play-by-play data into a statistical formula. Second, there may not be a hard and fast rule for when it's good to bleed time and when it isn't.

Saying that it's always good to run clock when you're ahead in the final minutes and it's always good to go out of bounds if you're behind is overly simplistic. Imagine that a team is down by one point at the 2 minute warning, but has a first-and-goal at the 10 yard line. They would probably want to stay in bounds and bleed the clock, whereas their opponent would want to force them out and call timeouts. I can't imagine crafting an algorithm that would deal with all these possible scenarios and increase the accuracy of DVOA.

87
by jimm (not verified) :: Wed, 12/19/2007 - 1:26am

Will, mm, Andrew - I think you're kind of missing my point. The sum of the parts don't add up.

According to these ratings the Vikings have a below average offensive line, horrible wr's and te's, a terrible QB and a good RB who ranks 8th in DVOA, . If that was indeed true there is no way the Vikings could possibly be the 10th best offence in the league.

With respect to the rating of the OLine as 15th best in run blocking - I find that so absurd that I don't personally assign any value to the run blocking rating at all. The Vikings average 5.4 a rush while every team commits to stopping the run. Peterson has been great (5.9) but both Taylor and Moore average 5.3 as well.

88
by RZR (not verified) :: Wed, 12/19/2007 - 1:53am

Re: DVOA accounting for clock-killing

IIRC from all the stats explanations, DVOA does in one way adjust for game situation that could be considered awarding clock killing. Late in a game, a team sufficiently ahead (8+ points I think) has lower thresholds for "success" as defined by DVOA, and a behind team has those thresholds increased, to reflect that a team ahead is more concerned with killing time than gaining [lots of] yards and a behind team needs more yards quickly to make up.

Perhaps more should or could be done, but it's not entirely true it is not counted at all.

89
by Alternator (not verified) :: Wed, 12/19/2007 - 1:57am

jimm, you're comparing the below-average DPAR rankings with an above-average DVOA. You wouldn't compare points scored to yards gained, so why compare these two stats that measure completely different things?

90
by Charger Jeff (not verified) :: Wed, 12/19/2007 - 3:53am

Raiderjoe, I know you take a ton of crap on here for your ridiculous rantings, and I know you're PORBABLY (to quote yourself) drunk, but I don't recall Hayes and Brown ever being rated at -100-% DVOA against WR2s ...

Besides, didn't the Chiefs and Raiders split this year? That makes them even if you throw out DVOA.

How does your superior team lose at home to the lowly Chiefs anyway?

91
by Chris M (not verified) :: Wed, 12/19/2007 - 8:37am

87, 89:

To expand on that - note that while Peterson is 8th in DVOA, Taylor is also 17th. That's pretty impressive.

As for the YPC - Adjusted Line Yards cuts off runs after 10 yards. It seems (and this is subjective, I've only watched him three or four times) that Peterson breaks a lot of long ones, which, after a few yards or so are based his cutting ability moreso than the line quality. Moreover, he gets stopped at the line a good bit (maybe this was just against Dallas and Chicago?), which will hurt ALY.

92
by Not saying (not verified) :: Wed, 12/19/2007 - 8:54am

Re: 41
I love Tampa Bay, but can someone explain to me how dominating a bad Atlanta team boosts them from a #6 defense to a #3 defense in a single week?

It's partly that the teams ahead of them (BAL and SEA) fell in the same week. There also wasn't that much difference in raw numbers (rather than ranking). The same is true this week: it's only a matter of a couple of percentage points.

Re: 42
In the main article, you wrote “The Cowboys’ rating for this game ends up a lot lower than the score would indicate because [of your scoring of Westbrook’s non-TD and one other factor].” In post 35, you say that it made essentially no difference. Which is it?

I think the point is that the TD or not makes very little difference in the DVOA, but it makes a lot of difference in the way people perceive the game through the score. 10-6 looks a lot different than 17-6.

93
by jimm (not verified) :: Wed, 12/19/2007 - 9:46am

89. Alternator - if you use DVOA it makes even less sense...the line is still ranked the same - and Peterson drops to 8th in DVOA as opposed to 4th in DPAR. One WR - Rice moves up to 23 but the rest (QB, WR, TE) have equally bad DVOA ratings.

94
by Karl Cuba (not verified) :: Wed, 12/19/2007 - 11:22am

90: If you're going to argue with the force of nature that is Raiderjoe, then at least make sense. How could Hayes and Brown be rated according to DVOA when it didn't exist back then?

95
by Brooklyn Bengal (not verified) :: Wed, 12/19/2007 - 11:58am

Thanks so much for the explanation of how the atrocious Bengals can be ranked so high by DVOA. My eyes and brain still have a problem accepting that Cincinnati's offense is efficient, as every game I've seen has Carson overthrowing receivers, receivers breaking the wrong way on their routes, the run game sucking horribly.

Still, here's to hoping the Bengals surprise everybody next year...

96
by Charger Jeff (not verified) :: Wed, 12/19/2007 - 12:15pm

94: Karl I was being sarcastic. The whole point about Cromartie being the best CB ever was in the same vein.

97
by Charger Jeff (not verified) :: Wed, 12/19/2007 - 12:16pm

However, that's one helluva DVOA number against WR2s for the Bolts. Pretty impressive turnaround for what had been a moribund pass defense.

98
by M (not verified) :: Wed, 12/19/2007 - 12:18pm

jimm, why don't you do a detailed analysis by DPAR which adds up all of the running and passing components by individual player? It should approximate the total - note that DVOA will not "add up" because it is a rate stat rather than a counting stat.

If you do a detailed analysis of the components you are talking about, then perhaps you may find there is a flaw (which Aaron would be happy to fix if that is shown to be the case). It is also possible this year's Viking team is so unique offensively that you are observing a weird numerical interaction which intitially makes no sense, but actually fits together once all (including very small) underlying components are incorporated.

99
by JasonK (not verified) :: Wed, 12/19/2007 - 12:32pm

I am absolutely shocked to find my Loser League team in 7th place in the entire league, in spite of my double-penalty QB duo of Campbell and Griese.

100
by Hank (not verified) :: Wed, 12/19/2007 - 12:38pm

It's unfortunate more people are not talking about the jags, a team that could very well go 12-4 and have as godo a chance as any team in the playoffs even though they are on the road.
The one thing most people say is they are built to compete with the colts- but this is just a trite misunderstanding.
How they are built is a good team, with solid defense, a great o-line, power runners and qb who excels at making good decisions. Of course a team ike that will play the colts tough- they would play any team tough.
Garrard's int-attempt ratio is amazing. If Brady had only 2 ints at this point, national holidays would be declared.
Fred Taylor is avg 5.1 ypc and mjd is close behind with 4.6.
And there two bad losses to the saints and colts were when injuries took garrard out of the game.
They should be in the super bowl discussion and are playing as well as any team, including new england, at this point in the season.

101
by jimm (not verified) :: Wed, 12/19/2007 - 1:22pm

98 M. - I think the point you make about the uniqueness of the Vikings offence may be pertinent, but they don't seem particularly unique to me - they've attempted 360 passes, which is about 100 less than average but there are 5 or so teams within 50 attempts of the Vikings.

The rating of the players is of course very difficult due to the interconnected nature of the game. I would argue for instance that Randy Moss when he played on the Vikings helped the running attack more than any other player because of the adjustments in defensive strategies he forced on other teams. But how on earth do you measure that.

Perhaps the Vikings WR's and TE's are actually far more valuable in the whole when blocking is taken into account. As Will points out Kleinsasser is destroying lbs and de's.

102
by mm (not verified) :: Wed, 12/19/2007 - 1:37pm

jimm-

I just checked the team stats page. The Vikings offense as a group is 3rd in rushing DVOA, 20th in passing DVOA, and 10th overall. If you consider how defenses defend them, it's probably fair to say they have the best running offense and a passing offense that's worse than 20th.

While some teams have a better running dvoa when one particular back is running, the Vikings only have a small drop off, resulting in a number that's better than all but 2 other teams. (and again, the Patriots and the Eagles play teams that actually have to defend the pass).

103
by jimm (not verified) :: Wed, 12/19/2007 - 1:54pm

M - I would agree with your assessment. I made a comment on this site early in the season if the league would simply outlaw the forward pass the Vikings would be Super Bowl bound.

104
by silentdibs (not verified) :: Wed, 12/19/2007 - 2:05pm

I can't be the first person to think "Rorshach from the Watchmen" whenever I read raiderjoe's comments.

I was going to say something about the Vikes' DVOA vs DPAR discrepancy, but others beat me to it. So this is all I got.

105
by Todd S. (not verified) :: Wed, 12/19/2007 - 2:37pm

#76 Raiderjoe, please read the introduction to the rankings. You'll find your complaint should look closer to this:

"Raiders are clearly ranked too low because they beat the Chiefs once out of two tries. My eyes are way better than this. Da RAiderz r going to the Super Bowl next year!"

106
by RickD (not verified) :: Wed, 12/19/2007 - 4:13pm

re: 22
I would think that anybody who watched the game would have noticed that Pennington was throwing better than Brady. It was pretty obvious. Yes, there was wind, but Brady was making bad decisions and poor throws.

re: 25
Put Pennington behind the Pats' O-line and he still has a weak-ass arm that cannot go deep. It's really too bad, since he had the makings of an excellent QB before his arm strength disappeared.

107
by DGL (not verified) :: Wed, 12/19/2007 - 5:38pm

#106: Put Pennington behind the Patriots' (current) O-line and he never would have had the two shoulder injuries that destroyed his arm strength.

108
by Nicky P (not verified) :: Wed, 12/19/2007 - 6:01pm

What I'd like to know about all this Brady-Pennington nonsense - is Brady mentioned because of convenience, because they happened to play each other this week and Pennington "outplayed" Brady?

For those that believe Pennington wins SBs with NE's OLine and system consistency, does he win SBs with Indy's OLine and system consistency?

Was the original statement trying to build up Pennington, knock down Brady, or both?

I don't understand.

Tom Brady can win 5 more SBs and people will still discredit him one way or the other. For the first 3 SB wins, he wasn't that great because they only won those SBs by 3. Now he's only great because he has the weapons. Which is it? The guy can't seem to catch a break with some people.

What will it take for the naysayers to admit that he's great? A few SBs and some SB MVPs? Record-setting statistics and a regular season MVP? Oh, wait...

109
by bravehoptoad (not verified) :: Wed, 12/19/2007 - 6:42pm

And what will it take for the nay naysayers to admit that most people admit that Brady's great? The foundation of a new religion? Being awarded some Superbowl MVPs and a regular season MVP? Oh, wait...

110
by Nicky P (not verified) :: Wed, 12/19/2007 - 7:06pm

109-

Nope, just keeping the names "Brady" and "Pennington" out of the same sentence will do.

If Pennington didn't lose the arm strength....

If Pennington had Brady's OLine...

If the Queen had balls she'd be King.

111
by enderwiggins (not verified) :: Thu, 12/20/2007 - 1:38am

wow the cowboys have 11 men in the probowl, just imagine how good brady would be if he was on AMERICAS TEAM.

112
by Bobman (not verified) :: Thu, 12/20/2007 - 2:57am

#100 Hank, I'll jump on board. They remind me a lot of the 05 Steelers, a #6 seed who didn't beat the QB-less Bengals all that impressively, dominated but still beat the #1 seed Colts with some luck, beat Denver more soundly, IIRC, and then more or less played the Seahawks to a draw but won the SB nonetheless. It was the Steelers' year and they won, no taking it away from them.

The Jags can do the same. They'd need some luck, but so does everybody. Not looking at seeding trees right now, so not sure if/when they'd face Indy, but they know each other well and a healthy Colt team would be a little tougher than the team the Jags took down to the wire a few weeks ago.

Still though, most Jags talk focuses on the "snubbing" of Fred T for Hawaii. Probably because most playoff talk focuses on the inevitability of the Pats winning it all. Sounds a bit like 2005 to me. Neither of the top 2 seeds made it last year; could be the same again this year (SD is looking tough and Rivers might remember how to play, who knows?).

113
by Scott (not verified) :: Thu, 12/20/2007 - 3:36am

Hoping for a Jags/Pats divisional game in bad weather. That could be interesting. You look at the playoff game two years ago there, and it was 28-3, but last year it was only 24-21 NE after Garrard fumbled late. He's a lot more efficient this year and their run game is as good as it's ever been. They can win that game, but I think they'll need a weather factor to help slow down NE's passing game.

114
by MJK (not verified) :: Thu, 12/20/2007 - 6:16pm

There isn't an overwhlemingly high probability of the Jags playing the Pats in the divisional game. The Pats will play the lowest seeded team after the wildcard round. That could be the Jags if they win and the 6 seed loses, but it's just as likely that the 6 seed wins, or that the Jags lose.

Personally, my money is on Cleveland to take the 6 seed, beat Pittsburg, and then lose to the Pats, while the Jags have to tangle with San Diego and then (if they make it) Indy--both non-weather teams with good offenses and strong running games. It's a tough road for the Jags unless they get lucky enough to face Cleveland or Pittsburg in the WC round, and even then they'll probably need some weather in Foxboro for the divisional round.

115
by Rich Conley (not verified) :: Thu, 12/20/2007 - 7:26pm

"Garrard’s int-attempt ratio is amazing. If Brady had only 2 ints at this point, national holidays would be declared."

Brady has only 6, but has 45 TDs and 500+ attempts. He also only has been sacked 17 times.

Garrard only has 2, but hes been sacked 20 times in only 300 attempts. He's only got 16 TDs.

When teams play the Pats, they play to stop Brady. When teams play the Jaguars, they play to stop Jones-Drew/Taylor.

116
by B (not verified) :: Fri, 12/21/2007 - 6:07pm

114: I think the Jags have a better chance of beating SD than Cleveland does of beating Pittsburgh. Of course, Pittsburgh still hasn't clinched the NFC North, so this is rather premature.