Writers of Pro Football Prospectus 2008

Most Recent FO Features

HarrisChr11.jpg

» Film Room: Chris Harris

Is Harris one of the league's top cover corners, or a product of the system in which he plays? Cian Fahey says the answer lies somewhere in the middle.

19 Dec 2006

Week 16 DVOA Ratings

by Aaron Schatz

Here are the latest DVOA ratings, with the commentary now available on FOXSports.com. This week's commentary talks a lot about special teams, since the Bears are on pace to finish with the second-highest special teams DVOA ever. You can check out a blog post about the top 10 special teams of the decade. Also in this week's commentary: the first appearance of the year for the DVOA week-to-week graphs.

(Wednesday Morning: Looks like FOX doesn't want to use these graphs, so I'll stick them up here. The point is to show the difference between a low variance, San Diego, and the absurd Jacksonville variance. Unfortunately, they don't fit side-by-side on FO and I don't have the time to re-do them.)

A few people asked in the Quick Reads thread for the DPAR ratings for Peyton Manning and Marvin Harrison last night. Well, prepare to be amazed.

Peyton Manning was worth 20.1 DPAR last night. That's the best game by any quarterback this year. (He was listed with 22.0 against Houston in Week 2's Quick Reads, but that was before opponent adjustments, and that game is lower now.) It is the seventh-best passing game by any quarterback in the last decade. I don't know where it stands in total value because I haven't done a sheet which adds up passing DPAR and rushing DPAR for 10 years of single games.

I'll try to do a blog post about the best quarterback games, or I'll do it in next week's Quick Reads. The best passing game of the decade is Randall Cunningham in Week 5 of 1998 against Green Bay: 20-for-32, 442 yards, four touchowns, 23.3 DPAR.

Harrison had 7.4 DPAR which would have ranked him first among wide receivers this week, but isn't one of the best games of the year. By the way, right now the top two wide receivers in DPAR are Colts and the next two are Bengals.

Pages are up now for offense, defense, and special teams. Individual stats are updated, adjusted line yards stats will be updated later tonight. You can also check out the Mike Harris Playoff Odds Report.

Don't forget to check out the Football Outsiders swag shoppe -- featuring ROBO-PUNTER shirts! They have our logo on the front, ROBO-PUNTER on the back with uniform number 1. Buy $50 worth of CafePress items before December 20, and get free shipping.

* * * * *

These are the Football Outsiders team efficiency ratings through Week 15 of 2006, 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 based on strength of opponent as well as to consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS DVOA is adjusted for type of stadium (warm, cold, dome, Denver/Mexico City) and week of season.

WEIGHTED DVOA is based on a formula which discounts games more than eight weeks ago in order to get a more accurate picture of how teams are playing now. This is the formula used for the rankings at FOXSports.com.

Remember that you can always use the keyword "DVOA" to access the latest DVOA commentary at FOXSports.com.

To save people some time, please use the zlionsfan template 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
W-L WEI.
DVOA
RANK OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
DVOA
S.T.
RANK
1 SD 33.9% 2 12-2 31.9% 1 25.2% 2 -4.9% 11 3.8% 5
2 CHI 32.0% 1 12-2 29.7% 2 0.1% 16 -23.7% 2 8.2% 1
3 BAL 28.0% 3 11-3 28.3% 3 1.0% 15 -23.9% 1 3.2% 7
4 JAC 25.6% 4 8-6 26.1% 5 2.2% 14 -22.5% 3 1.0% 14
5 NE 22.3% 9 10-4 27.4% 4 6.3% 8 -11.4% 5 4.6% 3
6 PHI 21.9% 5 8-6 16.4% 8 16.8% 3 -7.9% 10 -2.8% 23
7 IND 20.7% 10 11-3 22.6% 6 32.9% 1 8.1% 26 -4.1% 29
8 DAL 15.8% 6 9-5 16.8% 7 14.6% 6 -2.0% 14 -0.8% 19
9 NYG 14.5% 7 7-7 11.0% 9 10.8% 7 -3.0% 12 0.6% 15
10 NO 11.0% 11 9-5 10.3% 10 14.9% 5 4.5% 20 0.6% 16
11 CIN 11.0% 8 8-6 8.9% 11 16.1% 4 6.5% 25 1.4% 12
12 PIT 9.0% 13 7-7 8.2% 12 4.1% 10 -9.8% 8 -4.9% 30
13 KC 4.7% 14 7-7 7.1% 13 3.8% 11 0.5% 16 1.4% 11
14 CAR -0.2% 12 6-8 2.3% 14 -4.6% 20 -8.0% 9 -3.7% 26
15 BUF -0.9% 17 7-7 0.5% 17 -8.6% 23 -2.5% 13 5.1% 2
16 MIA -3.6% 15 6-8 1.3% 15 -11.7% 26 -10.0% 7 -1.9% 21
TEAM TOTAL
DVOA
LAST
WEEK
W-L WEI.
DVOA
RANK OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
DVOA
S.T.
RANK
17 DEN -3.8% 18 8-6 -4.6% 18 -2.8% 18 1.2% 17 0.2% 18
18 MIN -5.3% 16 6-8 -7.0% 20 -11.6% 25 -11.4% 4 -5.2% 32
19 NYJ -6.7% 19 8-6 -8.6% 21 2.2% 12 12.9% 28 3.9% 4
20 WAS -8.2% 22 5-9 -10.8% 23 2.2% 13 11.4% 27 1.0% 13
21 ATL -9.3% 20 7-7 -10.2% 22 -3.3% 19 2.6% 18 -3.4% 24
22 GB -10.0% 21 6-8 -6.7% 19 -6.8% 21 -0.6% 15 -3.9% 28
23 TEN -10.8% 23 7-7 0.6% 16 -10.0% 24 3.5% 19 2.7% 8
24 STL -13.0% 26 6-8 -19.8% 26 6.0% 9 15.5% 31 -3.6% 25
25 CLE -15.1% 25 4-10 -15.2% 24 -13.9% 28 4.9% 21 3.7% 6
26 SEA -17.7% 24 8-6 -20.7% 27 -13.5% 27 5.9% 24 1.7% 10
27 HOU -22.4% 27 4-10 -18.9% 25 -2.8% 17 16.9% 32 -2.8% 22
28 SF -23.1% 31 6-8 -24.8% 29 -7.3% 22 14.3% 30 -1.5% 20
29 TB -24.6% 29 3-11 -25.1% 30 -19.6% 31 5.5% 23 0.5% 17
30 ARI -24.7% 28 4-10 -20.7% 28 -14.2% 29 5.3% 22 -5.2% 31
31 DET -26.8% 32 2-12 -28.4% 32 -15.0% 30 14.2% 29 2.4% 9
32 OAK -28.8% 30 2-12 -25.8% 31 -35.3% 32 -10.2% 6 -3.7% 27

  • NON-ADJ VOA shows what the rating looks like without adjustments for strength of schedule, luck recovering fumbles, or weather and altitude on special teams.
  • 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. With just two games left, there's no need to list FUTURE SCHEDULE anymore.
  • 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
TOTAL VOA
ESTIM.
WINS
RANK PAST
SCHED
RANK VARIANCE RANK
1 SD 33.9% 12-2 38.3% 11.0 1 -5.1% 27 8.6% 29
2 CHI 32.0% 12-2 39.8% 10.5 2 -8.9% 30 23.2% 2
3 BAL 28.0% 11-3 31.3% 9.9 4 -1.9% 21 16.9% 13
4 JAC 25.6% 8-6 23.8% 8.9 7 2.1% 11 34.4% 1
5 NE 22.3% 10-4 25.1% 9.2 6 -1.9% 22 21.0% 4
6 PHI 21.9% 8-6 16.8% 10.1 3 -0.4% 18 12.8% 20
7 IND 20.7% 11-3 16.9% 9.5 5 5.2% 6 18.2% 10
8 DAL 15.8% 9-5 16.8% 8.2 8 0.0% 16 17.5% 12
9 NYG 14.5% 7-7 10.3% 8.2 9 4.3% 9 10.7% 25
10 NO 11.0% 9-5 12.7% 8.1 10 -2.8% 23 13.5% 19
11 CIN 11.0% 8-6 7.4% 7.9 11 4.6% 8 11.5% 22
12 PIT 9.0% 7-7 7.3% 7.6 12 1.0% 14 19.1% 8
13 KC 4.7% 7-7 3.0% 7.4 13 -1.3% 19 21.5% 3
14 CAR -0.2% 6-8 -3.3% 6.6 16 0.8% 15 11.4% 23
15 BUF -0.9% 7-7 -5.2% 6.9 15 5.1% 7 16.5% 15
16 MIA -3.6% 6-8 0.4% 6.3 20 2.3% 10 14.4% 18
TEAM TOTAL
DVOA
W-L NON-ADJ
TOTAL VOA
ESTIM.
WINS
RANK PAST
SCHED
RANK VARIANCE RANK
17 DEN -3.8% 8-6 -3.0% 6.3 21 2.1% 12 18.3% 9
18 MIN -5.3% 6-8 -0.3% 7.3 14 -4.5% 26 6.3% 31
19 NYJ -6.7% 8-6 -6.2% 6.5 17 1.9% 13 14.9% 17
20 WAS -8.2% 5-9 -12.8% 6.3 18 5.3% 5 10.0% 26
21 ATL -9.3% 7-7 -1.3% 6.3 19 -1.7% 20 21.0% 5
22 GB -10.0% 6-8 -10.0% 5.9 23 -4.4% 25 16.9% 14
23 TEN -10.8% 7-7 -14.3% 6.2 22 10.2% 1 17.6% 11
24 STL -13.0% 6-8 -1.6% 5.8 24 -9.3% 31 12.0% 21
25 CLE -15.1% 4-10 -22.9% 4.9 26 6.9% 3 9.1% 28
26 SEA -17.7% 8-6 -10.5% 4.9 27 -10.4% 32 10.8% 24
27 HOU -22.4% 4-10 -24.3% 4.7 29 5.5% 4 15.2% 16
28 SF -23.1% 6-8 -22.2% 5.1 25 -3.8% 24 19.9% 6
29 TB -24.6% 3-11 -34.7% 4.2 30 8.0% 2 6.1% 32
30 ARI -24.7% 4-10 -15.6% 4.8 28 -8.3% 29 19.5% 7
31 DET -26.8% 2-12 -24.9% 4.0 31 -5.4% 28 6.4% 30
32 OAK -28.8% 2-12 -33.5% 3.6 32 -0.2% 17 9.9% 27

Posted by: Aaron Schatz on 19 Dec 2006

180 comments, Last at 24 Dec 2006, 3:58am by the other Chris

Comments

1
by Wanker79 (not verified) :: Tue, 12/19/2006 - 5:58pm

First!

;-P

2
by Andy (not verified) :: Tue, 12/19/2006 - 5:59pm

NYJ #19, below Buffalo, Miami and the Pats.

All is right with the universe.

3
by Sophandros (not verified) :: Tue, 12/19/2006 - 6:00pm

If I said this:

The Saints are clearly ranked too high because I want to have a reverse of the FOMBC. Voodoo is way better than this. Go Eagles, beat the Cowboys and the Falcons!

Do you think it will work?

4
by Wanker79 (not verified) :: Tue, 12/19/2006 - 6:01pm

And now something actually useful...the Harris Playoff Odds Report links to the Week 14 report.

5
by princeton73 (not verified) :: Tue, 12/19/2006 - 6:01pm

the Jax variance continues to be off the charts

6
by RecoveringPackerFan (not verified) :: Tue, 12/19/2006 - 6:02pm

I assume that the Mike Harris report has no adjustment for resting starters?

7
by Tom S (not verified) :: Tue, 12/19/2006 - 6:03pm

Wow, You guys are really early with the DVOA Ratings this week!

Way to go! Big Props to You!

LOL

8
by Pat (not verified) :: Tue, 12/19/2006 - 6:06pm

Ooh. Playoff odds report does not like New Orleans' chances of getting the #2 seed.

And thank God for Weighted DVOA. Can you imagine if Indy fans saw Indy below Philly? Of course, this makes sense - Indy's playing much better now, and Philly's playing slightly worse. But still. It'd be insane.

9
by M (not verified) :: Tue, 12/19/2006 - 6:09pm

#5 - It will be interesting to see if JAX variance comes down once home/away is brought into the DVOA model, because a large part of their variance can be explained by this. One thing really funny about "All He Does Is Win" Garrard is that I believe he has single-handedly lost both the Houston home game and this week's Tennessee game. BTW, even though Leftwich had his flaws, didn't he lead JAX on a relatively large # of 4th Qtr comebacks the past two years, and consistently have a higher DVOA on 3rd downs than other downs? Even though there is no such thing as clutch...

10
by Rocco (not verified) :: Tue, 12/19/2006 - 6:10pm

The Steelers have a 5.7% chance of making the playoffs? I like those odds. Can we get MDS to declare that they have no chance of winning? It worked out last time. :) And they're up to #30 in special teams DVOA. I guess blocking a punt, running one back for a TD, and avoiding disasters will have that effect. Those are lofty heights.

Any chance we can have the NFC West declared a disaster area, and give its playoff spot to the #7 AFC team?

11
by ernie cohen (not verified) :: Tue, 12/19/2006 - 6:11pm

Something seems amiss in the playoff odds report. SEA is 96% to make the playoffs, but only 1.7% to win the two playoff games to make it to the conference championship game? I think I''d like to buy in on some of this 50-1 action. Or some of the 8-1 action against them winning the championship game if they get there...

12
by Andrew (not verified) :: Tue, 12/19/2006 - 6:14pm

Okay, have you added in the McNabb to Buckhalter TD from Redskins @ Eagles yet?

13
by RecoveringPackerFan (not verified) :: Tue, 12/19/2006 - 6:15pm

11: Seattle's magic number in the division is one. Their DVOA is -20%. I'll lay you that action.

14
by Kyle W (not verified) :: Tue, 12/19/2006 - 6:18pm

I would just like to congradulate whatever team gets to play the NFC West winners (If any of the teams ever decide they actually want to win that division) in the playoffs, Will the winner of that division be the worst playoff team ever?

Also what is keeping STL (just) above SEA? I can't think of any particular thing either team has done well in a long time now.

15
by the other Chris (not verified) :: Tue, 12/19/2006 - 6:19pm

Mike Harris' playoff forecasts are clearly ranked too high because they don't code the tiebreakers completely. nfl-forecast.com is way better than this. My kode rullzz h@x0r !!!!111eleven111

Seriously, I was preparing a graphical comparison between my playoff forecasts and Mike Harris' when I noticed that some of his forecasts are improbable, if not impossible. For example Mike shows some scenarios where St. Louis finishes in the 4th or 5th seed, when I'm pretty sure it is mathematically impossible for them to qualify in any seed but the 6th. This points to either an incomplete coding of the tiebreakers or errors in that code. These can lead to significant errors in the playoff predictions. For example, Mike projects the Chiefs with a 13% chance of making the playoffs. I have them rated below 2% and having looked at the scenarios in detail, I assure you it is much closer to 2% than 13%. Pittsburgh who is ranked nearly the same as KC in DVOA, has a head-to-head tiebreaker advantage over KC, and yet Mike Harris' projection for them (5.7%) is less than his for KC.

I was hoping to use a comparison of the performance of the two playoff forecast systems as a comparison of DVOA to my wins only based system. I've been impressed with the DVOA predictive capabilities as I've commented in previous weeks. However, this exercise as it stands now isn't an apples-to-apples comparison.

If Mike Harris or anyone on the FO staff wants to collaborate to put the DVOA games prediction algorithm in my software, I'm open to the idea. It would save duplication of work either on Mike's part in coding all the tiebreakers or on my part building a game prediction algorithm based on DVOA. Having tested my tiebreaker algorithms over several seasons, I'm pretty sure of its accuracy and the interactive interface is something most math-geek football fans find useful, especially at this time of the year. Just a thought.

16
by kevinNYC (not verified) :: Tue, 12/19/2006 - 6:19pm

I wonder if those playoff odds account for the NFL's asinine tiebreaker policy? The idea that the Giants could not go to the playoffs despite having an equal record and head to head tiebreaker over Atlanta is a bit scary.

17
by Rocco (not verified) :: Tue, 12/19/2006 - 6:20pm

#11-

According to DVOA (and most observers), any team Seattle faces in the playoff is likely going to be a better team. I could see them winning one game at home, but going on the road is likely to be fatal. It makes sense to me. I don't think the projections take kindly to a team that gets swept by the 49ers.

18
by the other Chris (not verified) :: Tue, 12/19/2006 - 6:23pm

16. The Giants do not control their own destiny. They mentioned this on MNF last night and I get the same result with my code. I don't remember what the exact scenario was, but if the Giants have a head to head tiebreaker, it probably involves a 3-way tie.

19
by Rocco (not verified) :: Tue, 12/19/2006 - 6:24pm

#15-

If I had to guess at the discrepancy between KC and Pittsburgh, it's because Pittsburgh has games left against Baltimore and @Cincy, whereas KC has @Oakland and Jacksonville. The projections probably give the Steelers such a low chance of winning this week (not that I blame them) that it makes it near-impossible for the Steelers to get to the playoffs. Also, the Steelers would lose a tiebreaker with the Jags and Broncos, while the Chiefs would win tiebreakers with both of them if they won out (I think).

20
by Jay (not verified) :: Tue, 12/19/2006 - 6:24pm

Looking at Mike's playoff odds report (which is really cool, obviously) it seems a little bit off to me - Chicago has an 83% chance to make it to the conference championship game? Unless I'm a moron, doesn't that mean the system predicts them to beat someone from the roughly {NO,Dallas,Philly,Seattle,NYG} 5/6 times? I'd take the under on that one.

21
by Pat (not verified) :: Tue, 12/19/2006 - 6:25pm

#15: It is impossible for St. Louis to finish other than the #6 seed (Philly holds the tiebreaker by common games). Looks like the head-to-head and common game tiebreakers aren't implemented right.

22
by Crushinator (not verified) :: Tue, 12/19/2006 - 6:25pm

Man, that's not good. 2 of the top 4 DVOA teams are also the top 2 in Variance. The Bears being 2nd in Variance could mean a lot come playoffs.

and the #1 Chargers are also one of the most consistent teams in the NFL. They're going to be brutal in playoff team.

23
by RoniFrassole (not verified) :: Tue, 12/19/2006 - 6:26pm

How do the Pats move up so many spots from last week?? It looks like its all due to their defense, but they played the Texans, so I gotta think there was a major opponent adjustment there.

24
by Pat (not verified) :: Tue, 12/19/2006 - 6:27pm

I wonder if those playoff odds account for the NFL’s asinine tiebreaker policy? The idea that the Giants could not go to the playoffs despite having an equal record and head to head tiebreaker over Atlanta is a bit scary.

Darn that whole "division" thing. :)

25
by dbt (not verified) :: Tue, 12/19/2006 - 6:28pm

Clickable charts updated except non-weighted VOA is hosed, fixing now.

26
by the other Chris (not verified) :: Tue, 12/19/2006 - 6:31pm

Rocco, you mentioned about the only two teams that KC does have tiebreaker advantages over if they win out.

Cincinnatti and Pittsburgh both have head to head wins.

And KC has the lowest conference record possible, having all of their losses in the AFC.

Even if I set KC's odds of winning their last two games to 100%, their odds are only 4.5% according to my system. I see no way they could be as high as 13%.

27
by Vince (not verified) :: Tue, 12/19/2006 - 6:34pm

The Playoff Odds Report gives Atlanta a non-zero chance to win their division. They've been mathematically eliminated. Otherwise, good stuff!

28
by David (not verified) :: Tue, 12/19/2006 - 6:37pm

The system also gives New Orleans a .1% chance to finish in a wild card slot, when they have locked up their division. Something's off.

29
by NewsToTom (not verified) :: Tue, 12/19/2006 - 6:37pm

PIT has exploded all the way to 30th in ST after a comparatively sparkling performance last week. The last place team in ST is not even Arizona, but instead Minnesota, who was a mere 27th two weeks ago. Go, Vikings!

Biggest weekly risers and fallers (DVOA):
+4: NE (9-5)
-3: CIN (8-11)
Biggest weekly risers and fallers (WDVOA):
+3: IND (9-6), PHI (11-8), PIT (15-12), SF (32-29)
-4: MIN (16-20)
Biggest WDOVA-DVOA disparities:
+7: TEN (23-16W)
-3: WAS (20-23W)

Tennessee's schedule strength came crashing down last week, but with games remaining against #5 NE and #15 BUF should remain strong.

30
by admin :: Tue, 12/19/2006 - 6:39pm

Funny thing. So many readers requested a Playoff Odds Report. For a long time, I didn't want to do one until we could get it perfect. But so many readers kept asking and asking for it. Finally, I decided that if the readers really wanted it, we might as well do one that was imperfect. So we finally introduced one, admitting right off the bat that it was imperfect. And what is everybody doing? Complaining that it is imperfect. Wow, there is just no pleasing people.

31
by Rocco (not verified) :: Tue, 12/19/2006 - 6:44pm

#26-

Like I said, I'm guessing. The playoff projections make my head hurt. :) You're probably right that the tie-breakers are off.

32
by Dr. Lofty Ambitions (not verified) :: Tue, 12/19/2006 - 6:45pm

Does anyone have a link to -- knock! knock! -- hold on, someone's at the back door.

back door opens
Oh, it's the New Orleans Saints! Sure, you're welcome.

Anyway, got a link to a good breakdown of the scenarios in which the Steelers make the playoffs? A good breakdown is defined as one in which the author knows how playoff tiebreakers actually work, including the cruical point that teams are first ranked within their division, and then only the highest-ranked as-yet-unseeded team in each division advances to the wildcard tiebreakers. A particular example of this crucial point in such a breakdown this year would be the hypothetical situation in which the three 9-7 teams tied for the last wildcard spot are the Steelers, Chiefs, and Broncos, in which case the Chiefs would win the tiebreaker over the Broncos on division record, but the Steelers would take the wildcard over the Chiefs on head-to-head.

(The Pittsburgh papers have horrendously misleading articles about tiebreaker scenarios. The NFL's own articles only mention teams that can clinch this week.)

I'd write such a good breakdown myself, if I weren't so busy with a toddler and a newborn. Hmm, freelance FO article in future years...

33
by Wanker79 (not verified) :: Tue, 12/19/2006 - 6:45pm

Am I the only one that the "Mike Harris" link sends to last week's report?!?

34
by RecoveringPackerFan (not verified) :: Tue, 12/19/2006 - 6:46pm

Sorry, Aaron. Mine wasn't intended as a complaint, just a search for an assurance that I could provide some level of support to my homerness giving the Pack a better chance of getting in than 1%.

35
by Kevin (not verified) :: Tue, 12/19/2006 - 6:46pm

16, 18:

If the Giants win out, but finish tied with both Atlanta AND Philadelphia, they lose out to both Atlanta and Philadelphia in the tiebreaker scenario.

There is some convoluted way that the Giants can clinch a playoff berth this week, though. I might be wrong here, but I believe it involves a Giants win, Philly win, Atlanta loss, a Minnesota loss, and a scenario where Seattle clinches the NFC West (either Seattle win or SF loss).

36
by Not saying (not verified) :: Tue, 12/19/2006 - 6:51pm

I notice the Pats are still the only team in the top ten in all three categories. Of course, the Chargers are just one spot away (11th in defense). Plus they have a much lower variance (29th) than the Pats (4th). SD is probably the actually most balanced team in the league, week in and out.

37
by the other Chris (not verified) :: Tue, 12/19/2006 - 6:51pm

33. wanker79

Mike Harris numbers his report by the week finished, whereas FO numbers there DVOA rankings by the current NFL week, which begins on Tuesday. So for Mike Harris, week 15 is the current week.

38
by Vince (not verified) :: Tue, 12/19/2006 - 6:51pm

Funny thing. So many readers requested a Playoff Odds Report. For a long time, I didn’t want to do one until we could get it perfect. But so many readers kept asking and asking for it. Finally, I decided that if the readers really wanted it, we might as well do one that was imperfect. So we finally introduced one, admitting right off the bat that it was imperfect. And what is everybody doing? Complaining that it is imperfect. Wow, there is just no pleasing people.

Of course you can please us. Just produce a Playoff Odds Report that is 100 percent accurate, all the time, starting in week 1, and get it done by this afternoon. Then we'll all be happy!

39
by Kevin (not verified) :: Tue, 12/19/2006 - 6:56pm

32:

The Steelers would only win a tiebreaker against Kansas City, having beat them head-to-head.

They lose any tiebreaker involving Cincinnati, even if they win in the last week of the season (conference record).

They lose heads-up tiebreakers with Denver and Jacksonville due to head-to-head losses.

They lose out on conference record tiebreakers with Denver and Buffalo.

They lose out on strength of victory tiebreakers with Jacksonville, the NY Jets and Tennessee.

So, basically, Pittsburgh's only shot at the playoffs is to finish 9-7, with the Denver/Cincy loser, Jacksonville, the Jets, Buffalo and Tennessee all finishing 8-8 or worse.

40
by Kevin (not verified) :: Tue, 12/19/2006 - 6:56pm

32:

The Steelers would only win a tiebreaker against Kansas City, having beat them head-to-head.

They lose any tiebreaker involving Cincinnati, even if they win in the last week of the season (conference record).

They lose heads-up tiebreakers with Denver and Jacksonville due to head-to-head losses.

They lose out on conference record tiebreakers with Denver and Buffalo.

They lose out on strength of victory tiebreakers with Jacksonville, the NY Jets and Tennessee.

So, basically, Pittsburgh's only shot at the playoffs is to finish 9-7, with the Denver/Cincy loser, Jacksonville, the Jets, Buffalo and Tennessee all finishing 8-8 or worse.

41
by ernie cohen (not verified) :: Tue, 12/19/2006 - 6:56pm

re: all the responses to 11:

What people seem to forget is that

(1) real team performance varies significantly over a period of a month - SEA might very well be a decent team in 3 weeks (it's not like they don't have any talent);

(2) game outcome is highly randomized by stupid things like turnovers - very few playoff matchups are 8-1 odds, period;

(3) SEA will have homefield in the first round, and likely a "virtual rest" week the week before, unlike their opponent. True, the opponent will be the loser from the east, but it's still nothing like 8-1.

42
by Wanker79 (not verified) :: Tue, 12/19/2006 - 6:57pm

Re: 37

When I click on the link it sends me to the Week 14 report.

43
by GlennW (not verified) :: Tue, 12/19/2006 - 7:13pm

> "Anyway, got a link to a good breakdown of the scenarios in which the Steelers make the playoffs?"

Here's the breakdown I've worked up on the Steelers. I believe it to be accurate, with only the very last line (TEN requirement) subject to an assumption: that TEN would win a three-way tiebreaker between PIT-NYJ-TEN on strength-of-victory, due to the fact that TEN has a comfortable lead that would be nearly impossible to overcome in that as-yet-TBD category.

PIT advancement (at 9-7) requires elimination of FIVE of SIX teams with the
following requirements:

CIN DEN > CIN

DEN (CIN > DEN and (SF > DEN or (KC > OAK and KC > JAC))) or
(DEN > CIN and SF > DEN and KC > OAK and KC > JAC)

JAC (NE > JAC and KC > JAC) or
(NE > JAC or KC > JAC) and (TEN > BUF and TEN > NE)

NYJ (MIA > NYJ and OAK > NYJ) or
(MIA > NYJ or OAK > NYJ) and (BUF > TEN or NE > TEN)

BUF TEN > BUF or BAL > BUF

TEN (BUF > TEN or NE > TEN) or
(MIA > NYJ and OAK > NYJ) or (NYJ > MIA and NYJ > OAK)

44
by chris clark (not verified) :: Tue, 12/19/2006 - 7:14pm

30: Aaron, I think the "complaints" about the playoff reports are really just nit-picks, rather than complete, "this is worse than nothing" kvetches. However, the ones from Chris Cox, "the other Chris" where he is offering you to work with his code, should be considered. In fact, if FO doesn't get to that by the end-of-the-season, I will probably do it to his code. I'm just busy on work, as in livelihood, until the start of the year and doubt that I will get it done before then.

45
by the other Chris (not verified) :: Tue, 12/19/2006 - 7:14pm
45
by Mike H (not verified) :: Tue, 12/19/2006 - 7:14pm

All,

Yup, the tiebreaker system in the Playoff Odds Report is incomplete. I view everyone's comments not as complaints, but as helpful constructive criticism :). Hopefully we'll have everything ironed out by next week. Sorry for the confusion and occasionally inaccurate percentages.

47
by AlexDL (not verified) :: Tue, 12/19/2006 - 7:15pm

Is the variance seperated by offense, defense and special teams anywhere?

According to your variance Numbers, the Patriots at Jaguars game looks to be the game to stay away from when it comes to laying bets. The #32 Jags (who have been better and more consistant at home) vs. the #20 Pats (who have been playing better and more consistantly on the road)are set to square off in an early playoff game.

I would have never thought that Jacksonville would be the #3 Defensive DVOA ranked team, not only are they #3(-22.5), but they are just 1.4% from #1(-23.9) Baltimore and then there is a huge jump to the Patriots who are #5(-11.4), that's an 11.1% jump. If we look at the team that is 11%+/- from the Patriots we have KC at #16 with a defensive DVOA of (0.6%). These numbers are saying that Jacksonville is dominating on defense, on par with Chicago and Baltimore, and league average on offense and special teams.
Then we have New England that is (according to DVOA, not rankings) better than average, maybe good, in all three fields, but not great in any one. Combine this with the above variance issues and we have ourselves one hell of an unpredictable football game.

Anyone else notice any games with unusual matchups this coming weekend?

48
by kevinNYC (not verified) :: Tue, 12/19/2006 - 7:16pm

#35... That's only true if the Eagles beat the Cowboys and the Giants lose the Skins. The Giants losing a tiebreaker to Philly is totally based on division record... if it goes past that the Giants will win common opponents easy.

49
by RecoveringPackerFan (not verified) :: Tue, 12/19/2006 - 7:19pm

41: How does the possibility of sudden improvement matter? It seems that Seattle could fall apart just as easily. I agree that 8-1 is high, but the game would almost certainly be on the road against a vastly superior team. I wouldn't actually lay the action, but I think that odds are well within the realm of possibility.

50
by GlennW (not verified) :: Tue, 12/19/2006 - 7:19pm

> "They lose out on strength of victory tiebreakers with Jacksonville, the NY Jets and Tennessee."

You can double-check, but I think the Steelers win individual tie-breakers with NYJ and TEN on record against common opponents, with strength-of-victory only coming into play on multiple-team tie-breakers involving those teams. That helps the Steelers a little bit...

51
by the other Chris (not verified) :: Tue, 12/19/2006 - 7:21pm

Mike H:

I'e be interested in seeing a redacted version of weeks 9 to date with the correct tiebreakers in, if you are up to it.

52
by admin :: Tue, 12/19/2006 - 7:24pm

A lot of people asked for last night's Peyton Manning and Marvin Harrison DPAR. That's now added above. Prepare to be amazed.

Variance for offense, defense, and special teams is on the individual pages for those units, although special teams variance is pretty much useless.

53
by CROdell (not verified) :: Tue, 12/19/2006 - 7:26pm

Apparently, Miami has not yet been mathematically eliminated. Under what insane scenario do they make the playoffs? As far as I can tell, it involves Cincy taking the #5 seed, a whole gaggle of teams losing over the next two weeks, with the end result that several teams (DEN, NYJ, etc.) sit at 8-8, with 5-7 conference records. And then my head starts to hurt, and I need to lie down.

54
by RobinFiveWords (not verified) :: Tue, 12/19/2006 - 7:28pm

20:

Over the past sixteen seasons, #1 seeds have won 81% of their playoff openers and #2 seeds have won 78% of their playoff openers. The combination of home field and a bye week really does make for a huge advantage.

Unless you're the Bears.

55
by Moridin (not verified) :: Tue, 12/19/2006 - 7:30pm

Heh, just 2 more games for the Vikings to attempt to finish with -40% in their Run D. A couple more percentage points to go :D

56
by Dave (not verified) :: Tue, 12/19/2006 - 7:31pm

San Diego is clearly ranked too low because first place just doesn't sufficeintly recognize LT's greatness. Scoring records are way better than this. Raiders suck, Chargers rule, good night now...

57
by David (not verified) :: Tue, 12/19/2006 - 7:38pm

35 is correct, though, that if the Eagles win, Seahawks clinch, Falcons lose and Vikings lose, the Giants clinch a wild-card berth. The amusing thing is that in that situation the Eagles also clinch some form of playoff spot, and the Cowboys already have one locked down. That locks up all six NFC berths (although not seeding or the NFC East crown) this weekend, which would be a humorously anticlimactic ending to this whole NFC logjam.

Oh, and on the playoff odds report: nobody expects the odds themselves to be on-target - not that we could even check if they were. It's when the model starts reporting impossible outcomes that people complain.

58
by turbohappy (not verified) :: Tue, 12/19/2006 - 7:39pm

Thanks Aaron! Peyton was on fire last night...only a couple throws that weren't perfect.

59
by CaffeineMan (not verified) :: Tue, 12/19/2006 - 7:39pm

Woo hoo, data. Thanks Outsiders. And thanks for the charts, dbt. Now all I need is that guy to update the DVOA graphs, and I can make my office pool picks. :D

On the Pats moving up, it's the defense (a lot) and special teams (a little). I think the offense dropped a little numerically, but not in ranking.

I'm surprised the Pats offense didn't drop more, given how bad Houston's defense is. The offense's loss is Gostkowski's gain, however, as the kid crawls towards league average in FG/XP (-1.3 versus -3.3 last week). I'd love to see the split in the KICK stat for how much is the kicker and how much is the coverage team, as the Pats are now 4th in that category.

As to the defense, Houston's offense is not nearly as bad as their defense (-2.8%, 17th). And it was a pretty impressive game if you look at Houston's drives:

PUNT, 4 and out, 9 yds
INT, 1 play
INT, 5 plays, 36 yds
PUNT, 3 plays, -1 yds
PUNT, 3 plays, 0 yds
PUNT, 4 plays, 14 yds
PUNT, 5 plays, 26 yds
TD (run), 10 plays 70 yds (lots of runs)
PUNT, 6 plays, 11 yds
INT, 1 play
INT, 2 plays, 11 yds
PUNT, 6 plays, 18 yds
END OF GAME, 1 play, 2 yds

The Texans had only 1 series all game long (the one where I was missing Wilfork and Seau) where they had more than 2 series of downs on offense. Still, the Pats rose a little more than I would have guessed. Other teams fell slightly as well, I think.

60
by Not saying (not verified) :: Tue, 12/19/2006 - 7:45pm

Re: 47 Is the variance seperated by offense, defense and special teams anywhere?

When the stats are updated on the pages for each kind (see: Just the Stats, Team Offense, etc.), you can see the variance for each part. Right now they're only updated through Week 15.

Then we have New England that is (according to DVOA, not rankings) better than average, maybe good, in all three fields, but not great in any one.

The special teams DVOA seems pretty great. I mean, it's not Chicago, but it's second tier.

61
by Travis (not verified) :: Tue, 12/19/2006 - 7:47pm

Apparently, Miami has not yet been mathematically eliminated. Under what insane scenario do they make the playoffs?

There isn't any. They can't even make it out of the AFC East at 8-8 - both the Jets and the Bills beat them in tiebreakers.

62
by GlennW (not verified) :: Tue, 12/19/2006 - 7:47pm

Miami has been mathematically eliminated. They can't pass NYJ in their own division, and both CIN and DEN have already secured at least a 6-6 conference record.

63
by Gerry (not verified) :: Tue, 12/19/2006 - 7:49pm

"And what is everybody doing? Complaining that it is imperfect. "

Everybody? I wasn't. I was sitting here complaining about the stupidity of my Giants.

And thinking about having a beer.

64
by the other Chris (not verified) :: Tue, 12/19/2006 - 7:52pm

57. David - acually it will be possible to check the accuracy of the playoff odds reports after several years results have been tabulated. You can check if teams who have are predicted as 75% chance of making the playoffs actually qualify at a rate close to 3 out of 4 times. If such teams actually qualify significantly more or significantly less that 3 out of 4 times, it's an indicator that the predictions need tweaking.

65
by Stillers Rule (not verified) :: Tue, 12/19/2006 - 7:56pm

Pittsburgh is clearly ranked too low because Santonio Holmes has learned to telekinetically cause a penalty when he's about to fumble a punt. Stephen Colbert's gut is way better than this. The Steelers are the champs and until their not, there number 1! WOOO!

66
by David (not verified) :: Tue, 12/19/2006 - 8:12pm

True in the long run, but in its first month of existence, with predictions specific to the tenth of a percent? That's not really verifiable by any method other than formulating a different statistical model and arguing that yours is better.

67
by admin :: Tue, 12/19/2006 - 8:22pm

Re: 63. Mmmm, beer.

Adjusted Line Yards are now updated. Along with the new ALY numbers goes this blog post (click link on my name) asking what the hell is going on with the Indianapolis Colts being ranked number one defensively against runs left end.

68
by Will Allen (not verified) :: Tue, 12/19/2006 - 8:26pm

Other Chris, could you e-mail Brad Childress, and ask that he inform the passing-challenged Vikings that they still have a better than 20% chance of making the playoffs? I assume that only holds if the win out; I can't imagine there is any scenario that could get them in at 7-9, but a 1 out of 5 chance ought to be enough to keep the guys from folding their tents just yet.

I was surprised to see the Vikings maintain their top five defensive ranking after yielding about 250 yards passing in the first half last week. I guess nobody immediately below them had a great week defensively, either. They had better improve Thursday at Lambeau, however, cuz' I don't think The Era of Tavaris Jackson Domination is ready to start just yet. Also nice to see them fall to dead last in special teams performance.

Ah, Aaron makes mention of the '98 Randy Moss breakout game, versus the Packers, back when teams trembled at the the thought of crowding the line of scrimmage against the Vikings. Now, they put everybody but the beer vendors within two yards of the neutral zone. Still, I'm not quite ready to be nostalgic for those days. Give me another year or two of Troy Williamson, Billy McMullen, and the like, however, and I might get there.

69
by Alex DL (not verified) :: Tue, 12/19/2006 - 8:35pm

#52 thanks...although wordpress is telling me to F-off...another day perhaps wordpress, another day.
Why is the ST variance useless?

70
by Pat (not verified) :: Tue, 12/19/2006 - 8:35pm

If such teams actually qualify significantly more or significantly less that 3 out of 4 times, it’s an indicator that the predictions need tweaking.

The number of years that you'd need to do that is pretty large, though. You can't cheat and use the same team multiple times after multiple weeks, either - that'd give you false confidence in the predictions (i.e. if team A has an 80% chance at week 15, an 85% chance at week 16, and a 90% chance at week 17, you can't use the same team to check the 80%, 85%, and 90% brackets. The results are obviously dependent).

Depends on the accuracy you wanted, though, I guess.

I think that in some sense that's already done for the DVOA playoff odds: they're using the percentage of teams that win games when separated by X% in DVOA, I think.

71
by the other Chris (not verified) :: Tue, 12/19/2006 - 8:39pm

Will Allen belive it or not, there is still at least one scenario where a 7-9 Minnesota team can make the playoffs. It requires a victory at Green Bay, but beyond that I haven't tracked that scenario down. An 8-8 Vikings team has a better than 50-50 shot, assuming my individual game odds are anywhere near correct. They definitely still have a reason to play.

72
by CROdell (not verified) :: Tue, 12/19/2006 - 8:41pm

#61,62:

Thanks. Though I'm kinda disappointed; I was hoping there was some wonderfully byzantine scenario involving strength of victory/schedule tiebreakers, Nick Saban making human sacrifices Aztec-style, etc. No, it's just FoxSports.com not updating their playoff picture page correctly.

73
by paytonrules (not verified) :: Tue, 12/19/2006 - 8:44pm

Bears have fallen quite a bit on defense in the last two weeks. Being ranked second may not seem like much, but if memory serves it's a 7% drop. Ouch.

This is interesting to me though. FO has pointed out that defenses typically fall off from the super high that the Bears put up a year ago, and Aaron or MDS (I forget which), mentioned that two years of top-level performance is nearly unprecedented, meaning the Bears were sure to go down next year.

Well the Bears started going down already, but are likely to end as a top 5 defense (not the BEST EVER ) which is just like what FO originally said about excellent defenses. They fall - but don't collapse (except for Buffalo).

Does this mean that maybe the Bears aren't as likely for a collapse next season - because they aren't actually bucking any odds? Could they stick top 5 year after year?

Who am I kidding.

74
by Crushinator (not verified) :: Tue, 12/19/2006 - 8:50pm

Awesome. I checked out the Harris Report. My dream SB of Titans/Niners has statistically NO chance of occurring.

75
by the other Chris (not verified) :: Tue, 12/19/2006 - 8:58pm

Pat: I agree with everything you said. I tried the percentage of home teams that win games when separated by X% in DVOA, for 2005. Interestingly, it did not give a monotonically increasing function, even when I used a running average of over 20 games. There were some pretty strong dips, where the probability of winning went down by as much as 10% as DVOA difference went up 10%. I see the same thing in my power ratings, (actual outcomes moving in the opposite direction of predicted outcomes) so the take away message is that individual games are very difficult to predict in the NFL.

76
by Travis (not verified) :: Tue, 12/19/2006 - 8:59pm

Will Allen belive it or not, there is still at least one scenario where a 7-9 Minnesota team can make the playoffs. It requires a victory at Green Bay, but beyond that I haven’t tracked that scenario down.

1. Minnesota beats Green Bay, but loses to St. Louis.

2. St. Louis loses to Washington.

3. Green Bay loses to Chicago.

4. Carolina beats Atlanta, but loses to New Orleans.

5. Atlanta loses to Philadelphia.

6. The Giants lose to New Orleans and Washington.

7. San Francisco doesn't win twice.

NYG, WAS, MIN, ATL, CAR, and STL (and possibly SF) finish at 7-9. Break division ties first; NYG beats WAS on division record, and CAR beats ATL on head-to-head (if SF finishes 7-9, they beat STL on division record). MIN has a 7-5 conference record, better than NYG, CAR, or STL (or SF), so they get the second wild card berth.

77
by Travis (not verified) :: Tue, 12/19/2006 - 9:01pm

Correction: in the scenario above, CAR beats ATL on division record, not head-to-head.

78
by NewsToTom (not verified) :: Tue, 12/19/2006 - 9:08pm

Re #74
Hey, don't be doggin' the 49ers. They could still end up hosting TWO playoff games. Sure, the ocmbined chances of (i) winning the division, (ii) making the NFC CG, and (iii) facing the other #6 seed in the NFC CG are almost inifinitesimal, but TEN's playoff chances were at .02% when they were 2-7 and are now much better. C'mon, you gotta have faith!

79
by Scott C. (not verified) :: Tue, 12/19/2006 - 9:40pm

Wanker79

Maybe your browser is caching it.

Hold down control and click the refresh button.

80
by Arren (not verified) :: Tue, 12/19/2006 - 9:52pm

Long time, no comment.

Congratulations to all of you Outsiders on FO's continuing ascendancy!

That said, perusing the comments lately, I'm taken aback by the moronic "First!" posts..... what is this, slashdot?

Carry on with your great work, Aaron and co.! Naysayers be damned!

81
by MJK (not verified) :: Tue, 12/19/2006 - 10:03pm

Wow,

The #4 and #5 teams (either way, using either total or Weighted) DVOA square off next week. They are rated #1 and #4 in variance.

I'm afraid to even try to predict what the score might be...

82
by Pat (not verified) :: Tue, 12/19/2006 - 10:03pm

Pat: I agree with everything you said. I tried the percentage of home teams that win games when separated by X% in DVOA, for 2005.

Statistics will kill you. Murder you, in fact.

Percentages aren't exact with finite statistics: you're limited by quantization effects and simple binomial noise. It's worse because you're killed by both the numerator and the denominator as well. A 10% drop isn't significant until you've got probably 30 or so events in that bin. And that might be a little low.

83
by Sam! (not verified) :: Tue, 12/19/2006 - 10:10pm

47:
Going into the season, I fully expected the Jaguars Defense to be in the top-3 DVOA. They just had too much talent not to be there, with the addition of Brian Williams at right corner.

What I did not think was that they could be a top-3 defense missing Mike Peterson at MLB, Donovin Darius at SS and without Reggie Hayward at DE, and without a healthy Marcus Stroud for much of the season at DT. Mike Smith (Jax DC) deserves a look at head coach, but I hope he keeps getting passed over!

84
by Jeremiah (not verified) :: Tue, 12/19/2006 - 10:50pm

The best passing game of the decade is Randall Cunningham in Week 5 of 1998 against Green Bay: 20-for-32, 442 yards, four touchowns, 23.3 DPAR.

How much was Manning's 22/26, 377 yards, 5 touchdowns against Denver in the 2003 playoffs worth?

85
by the other Chris (not verified) :: Tue, 12/19/2006 - 11:14pm

82. Pat-Yeah that's what is happening. It was actually down 15% (rather than 10%) but that is still just 3 victories fewer than expected out of 20 games. Easily explained by small sample size.

86
by Paul (not verified) :: Tue, 12/19/2006 - 11:29pm

#1 offense plays #32 defense...why do I expect the Texans to hold the Colts to under 20?

87
by admin :: Tue, 12/19/2006 - 11:42pm

84: Whoops. I should have said "regular season." I haven't broken down playoff stats for individuals, just teams. 81: The New England comment will read basically just like that, with an added shout-out to an old co-worker of mine in Jacksonville.

And, as if on cue, the commentary appears: http://msn.foxsports.com/nfl/story/6288640. I just e-mailed FOX to tell them that the Jacksonville comment with the two graphs only works if you actually run the graphs, so hopefully those show up sooner rather than later.

88
by calig23 (not verified) :: Tue, 12/19/2006 - 11:42pm

Re:#50

This is correct. The Steelers come out ahead against the Jets and Titans in the event that they finished tied 9-7.

89
by b-man (not verified) :: Tue, 12/19/2006 - 11:43pm
90
by Will Allen (not verified) :: Wed, 12/20/2006 - 12:10am

Yeah, Chris, my gut feeling at mid-season was that if the Vikings could go 8-8, with four losses in conference, they'd have a better than even chance to make the playoffs, so I'm not shocked. I just hope Tavaris Jackson begins his inevitable march to Canton Thursday night, thus giving them some spark. I have a bad feeling that The Gunslinger is going to have a big night, however. If they don't play with desperation, regardless of outcome, it does not bode well for the Reign of Childress.

91
by Catfish (not verified) :: Wed, 12/20/2006 - 12:12am

Hmmm, I think I detect a theme in this weeks commentary.

92
by b-man (not verified) :: Wed, 12/20/2006 - 12:17am

Did NE have one of the best ST games of the year last week(around 38%)? Walter had 36 Net/punt on 4 punts with 3 inside the 20. Gostk was 4 for 4 on field goals. Got a TO off a punt. Returned a KO for a TD. Punts 4 for 11/per.

93
by Ima Pseudonym (not verified) :: Wed, 12/20/2006 - 12:35am

92: I hope so, because apparently Walter sold his sole to the devil for that game. He is on IR now. The Pats are going through punters like Spinal Tap goes through drummers... or like the Pats go through defensive backs.

94
by throughthelookingglass (not verified) :: Wed, 12/20/2006 - 12:41am

93
or like the Bucs go through quarterbacks...or like the Browns go through centers..or like the Bengals go through the courts...i could do this for hours.

95
by Riceloft (not verified) :: Wed, 12/20/2006 - 1:13am

Fear not, Cadillac Williams plays the Browns this week. He'll magically appear in the 3rd quarter when it looks like the Browns are sitting pretty with a 10 point lead.

96
by Trogdor (not verified) :: Wed, 12/20/2006 - 1:27am

Atlanta's been playing a running back out of the shotgun for several years.

Did any of the NFL teams start running those plays before Arkansas busted out their Wildcat offense? That's just a great set of plays to watch.

97
by Andrew (A.B.) (not verified) :: Wed, 12/20/2006 - 1:31am

"...there is a limit to the number of weeks I can make a Matt Millen joke."

Is there a limit to the number of times you can make a "last home game ever" joke?

98
by Kalyan (not verified) :: Wed, 12/20/2006 - 2:34am

I am mailing this to Aaron seperately as well:

I have no complains in seeing NE Pats at #5. But i am not sure that they deserve it. I cannot point a finger to the exact nature of their ills (though the DVOA ranking states that they have the #8 offense!), but somehow the offense seems to lack the 'IT' factor. I am not sure if the offense can stand up to good defenses. I hope i will be proven wrong, but my worries still remain.

99
by Boots Day (not verified) :: Wed, 12/20/2006 - 2:40am

The important thing is, it looks like Chris Gardocki is having fun again.

100
by Brian (not verified) :: Wed, 12/20/2006 - 3:01am

I must say these rankings are really what everyone needs to look at. These really go in-depth and they make sense in my opinion and really show who are the best teams. I have tried all year to get people to realize teams like cincinnati and kansas city are overrated. And that the Eagles and Jaguars really are good. All those statistics espn shove in your face really dont take many variables into account like these ratings do. But I tell you, am i the only person who really is looking forward to the AFC playoffs. A team like Jacksonville wow! I mean this team is so intriguing to me because these ratings show as everyone knows that the Jaguars are the more inconsistent upper-echelon team in the league. But if you throw them in the big dance they really can not only beat but flat out dominant teams. If they play Indy in the wildcard, they could easily win that. One of the main reasons the Bengals screwed up on monday against the colts was because they didnt play to their strength. They just tried to expose the Colts weakness by running it most of the time. The difference with the Jags is that not only is rush defense the Colts weakness but thats actually the Jaguars strength. That combo between Fred Taylor and Maurice Jones-Drew is flat out dangerous. That defense has obviously proven by now that it is right up there with Baltimores and San Diego's. The greatest thing the Jags have going for them is that nobody outside of these ratings knows that the Jags defense is that good. I mean they are playing without 3 of their best players. One of them is a pro bowler and one is an arguable pro bowler. I mean the job that coaching staff has done is nothing short of outstanding and it will be a shame if they don't make the playoffs because they really have the potential to shake things up.

101
by Ferg (not verified) :: Wed, 12/20/2006 - 3:22am

98: Well, look at their pass DVOA the past three years: 39.3%, 31.4%, 15.4%. So if it seems that the offense declined, that's because it did, and it was a pretty big decline. But since it was awesome before, it could decline quite a bit and still be pretty good.

102
by Andrew (not verified) :: Wed, 12/20/2006 - 3:24am

Isn't it amazing, just like clockwork, that all 11 of the teams that started 0-2 are for all intents and purposes, dead in the water once again, and most likely not to make the playoffs.

103
by hector (not verified) :: Wed, 12/20/2006 - 3:48am

Peyton Manning is like Mickey Mantle in the early ‘60s — it's like he's getting left out of the MVP conversation because everybody agrees that it's time to give it to someone else.

Larry Bird, entering 1986-87. Here, Magic.

Tom Brady spontaneously combusts . . .

I don't believe in jinxes, but why risk it, brother? I'm still peeved the Patriots left Brady in as long as they did in the Minnesota game, chucking the ball against a frustrated defense.

I don't think it's an accident that Shananhan's team isn't getting flagged. Between his team respecting him and buying into his thing *and* the officials perhaps doing the same, I can see there's something to it. Meanwhile, Tom Coughlin *has* to be doing something wrong, no? Demanding people to come to meetings early and blowing your top a few times can't be enough. Of course I'm seldom invited to practice, so how can I say . . . but we'll continue to speculate, I suppose.

104
by hector (not verified) :: Wed, 12/20/2006 - 3:55am

I agree New England at 5 looks shaky. I'd love to hear a rational argument of how this team can go deep into the playoffs (and I'd love to be proven wrong, but sometimes you don't have the cards). Home field for a game will be nice, but I can't see them beating two good teams in a row, and certainly not three. The offensive line has dropped significiantly, the receivers are an embarrassment given the QB (can Brady have a star on the outside just *once*?), and the defense is missing key guys on every level.

Mind you, when your circle-the-wagons year is still a playoff and likely division winner, that aint' so bad.

105
by Kal (not verified) :: Wed, 12/20/2006 - 4:24am

#104: if NE goes deep into the playoffs, they'll do so because they have a good all-around team with no significant weaknesses other than an average running game, and they finally figure out how to not be so inconsistent from game to game.

That being said, you guys should look at not just the rankings but the actual percentages for DVOA. For instance, while they're 5th, they're 10% below SD at first. That's a big dropoff. The weighted numbers say that they're playing better - likely because early bad games against Denver weigh less now - but that's a dropoff.

Similarly in offense - while they rank 8th in offense, they're 20+% below the leader and in such company as KC, STL, Pitt and Baltimore right now.

So...I don't think they're going to go deep in the playoffs. The teams they're likely to face are all significantly higher in other categories. But they're definitely not a bad team, even with horrible losses to Miami in recent weeks.

106
by Kal (not verified) :: Wed, 12/20/2006 - 4:32am

Also, from the DVOA commentary - it's interesting that the Saints defense (from a year-to-year standpoint) is projected to actually improve next year, because of their poor 3rd-down conversion ranking. That's kind of a scary thought; Bush and Colston as 2nd year players, Henderson getting more time, and their defense actually improving...that might alone be enough to reign supreme throughout the NFC next year, and certainly would be better than what the Panthers, Bucs and Falcons have going on.

107
by Sean (not verified) :: Wed, 12/20/2006 - 7:04am

Not really related to this thread, but any guesses on what Al Davis would be willing to trade to get Mick Vick for 2007 and beyond?

108
by Pat (not verified) :: Wed, 12/20/2006 - 10:10am

82. Pat-Yeah that’s what is happening. It was actually down 15% (rather than 10%) but that is still just 3 victories fewer than expected out of 20 games. Easily explained by small sample size.

See, you can use data with a small sample size - you just need to properly put the errors in it. Doing that, I'd imagine you could easily get a monotonic function so long as the fit knew that those data points were complete garbage.

109
by dryheat (not verified) :: Wed, 12/20/2006 - 10:23am

#93...I had no idea Satan enjoyed fish so much. I'd be totally willing to sell my flounder and cod to the Devil in order to win the lottery and shack up in the penthouse of the Four Seasons with Adriana Lima. (Sorry...couldn't resist)

#104...In order to win the Super Bowl, the Patriots will only have to beat two good teams. The other one, by rule, has to be from the NFC. I wouldn't bet on them winning the Super Bowl, but that helps their odds some.

110
by Trogdor (not verified) :: Wed, 12/20/2006 - 10:30am

Perhaps he sold the bottom of his foot? While I think I'd also be willing to do make that trade, I would imagine that's a body part you don't realize how much you'd miss it until it's gone.

111
by Wanker79 (not verified) :: Wed, 12/20/2006 - 11:32am

Re: Dallas Commentary
One reason why the Cowboys and Eagles have virtually the same rating: The Cowboys have recovered 63 percent of fumbles, the best percentage in the NFC. The Eagles have recovered just 37 percent of fumbles, the worst percentage in the NFC except for New Orleans.

It's way too early for me to wrap my head around exactly what that's trying to say. A little help here?

112
by Travis (not verified) :: Wed, 12/20/2006 - 11:37am

#104…In order to win the Super Bowl, the Patriots will only have to beat two good teams. The other one, by rule, has to be from the NFC. I wouldn’t bet on them winning the Super Bowl, but that helps their odds some.

The Patriots can only get a first round bye if they win twice AND Indy and Baltimore lose twice. Assuming they win the division, their first round opponent will likely be a 10-win AFC team, and three of the four potential candidates (Jacksonville, Denver, the Jets) will have beaten the Patriots this year. It's going to be a tough road to the Super Bowl.

113
by Ima Pseudonym (not verified) :: Wed, 12/20/2006 - 11:44am

109, 110: Perhaps the pun would have been more clear if I had included that Walter is out for the season with a foot injury.

114
by Rocco (not verified) :: Wed, 12/20/2006 - 11:53am

#93- If you're going to sell your soul for one good game, why would you waste it on a game against the Texans?

115
by Rocco (not verified) :: Wed, 12/20/2006 - 11:54am

Aaron- a minor quibble- the Jags' graph has them playing Philly twice. Week 2 was against Pittsburgh.

116
by Travis (not verified) :: Wed, 12/20/2006 - 11:56am

Re: 111

Dallas has been "lucky" in recovering fumbles; the Eagles have been just as unlucky. Those fumble recoveries affect the two teams' winning percentages, but not their DVOAs.

117
by Pat (not verified) :: Wed, 12/20/2006 - 12:16pm

#116: Or, otherwise stated: Philly's defense is better than you think it is - they're causing tons of fumbles. It hasn't translated into more turnovers yet, but it will. Also, Dallas's offense is worse than you think it is. They're fumbling the ball a lot, and while it also hasn't translated into more turnovers yet, it will.

(There's also the pair of "Philly's offense is better than you think it is, because they aren't really fumbling much, and Dallas's defense is worse than you think it is, because they're not causing that many fumbles.")

118
by Wanker79 (not verified) :: Wed, 12/20/2006 - 12:25pm

I get the part about Dallas having a ton of fumble recovery luck and Philly...not so much. It's the first sentence I'm having trouble with. Is the public perception really that Dallas is a far superior team to Philly even though Philly could take over the division lead with a win this week???

119
by MRH (not verified) :: Wed, 12/20/2006 - 12:29pm

Some observations on SD and Marty and variance:
SD variance ranking last three years: 29, 29, 32 (working backwards).

No other team has been that consistent in that span. The closest has been OAK (27, 32, 20) which proves that consistency is not necessarily a good thing. And maybe it’s not the coaching… or at least the head coach alone.

In the last three years, only two other teams have been as consistent as SD for a two year span: IND (10 this year, 31 last year, 28 in 2004) and JAC (1, 28, 30).

Besides JAC being very consistent last year, so was NE (27). Looking at other highly consistent teams over the last three years, consistency appears to rarely carry forward. This also seems to be true if I use the variances themselves and not the rankings.

Looking at Marty’s pre-2004 teams, their variance ranking has been much higher (i.e. less consistent): 18, 19, 1 (WAS, 2001), 3 (KC, 1998), 11 (KC, 1997). I’m sure Marty prides himself on getting equal effort each week from his teams but only the last three years has it been notable.

A last note: Marty’s 2001 Redskins were 3rd in weighted DVOA at 25.7%. I’m sure the improvement in the team is part of the reason for the high variance (they were awful in the 1st 5 weeks at 0-5 and then 8-3 the rest of the way). You have to wonder what would have happened in 2002 if Snyder hadn’t fired Marty and hired Spurrier.

120
by Travis (not verified) :: Wed, 12/20/2006 - 12:29pm

Also, Dallas’s offense is worse than you think it is. They’re fumbling the ball a lot, and while it also hasn’t translated into more turnovers yet, it will.

Not true. Dallas has fumbled the ball 14 times, tied for 2nd fewest in the league. (Philly, at 24, is tied for 9th most.)

True for defense, though. On defense, the two have forced about the same amount of fumbles (Dallas 25, Philly 23), but Dallas has recovered 4 more than Philly has.

121
by turbohappy (not verified) :: Wed, 12/20/2006 - 12:37pm

I was just looking at the stats pages - and it looks like Peyton has over 150 DPAR (including Monday night) and the next highest QB (Brees) has around 100. That's insane.

122
by Rich Conley (not verified) :: Wed, 12/20/2006 - 12:37pm

Travis, why do you think Jax is going to beat the Patriots?

Jaxonville does one thing well on offense: Run the ball. The patriots defence is absolutely phenomenal at stopping the run.

Honestly, I think this game is going to look a lot like Pats/Bears... except Jax has nowhere near the special teams the Bears do.

Jax/NE may be similar in DVOA, but I think this is a real bad matchup for jaxonville. Its going to look like last year's NE/Jax playoff game. Except this year's NE team is much better than last years, and this year's Jax team is pretty much the same as last year's.

123
by Pat (not verified) :: Wed, 12/20/2006 - 12:41pm

Not true. Dallas has fumbled the ball 14 times, tied for 2nd fewest in the league. (Philly, at 24, is tied for 9th most.)

Let me rephrase that, then. "Dallas is fumbling the ball more than you think they are. It hasn't turned into turnovers yet, but it will."

Romo has fumbled 5 times, and recovered all but one of them, for instance.

124
by GBS (not verified) :: Wed, 12/20/2006 - 12:42pm

I have to believe there is a mistake in the Colts' defensive rushing statistics. The Colts' rushing defense is woeful in every way, except that somehow they are 1st in adjusted line yards against "left end" runs. They are first by a LARGE margin. That just doesn't seem possible unless there have been very few runs attempted in that direction.

Freeney must be a monster against the run after all.

125
by Travis (not verified) :: Wed, 12/20/2006 - 12:45pm

Travis, why do you think Jax is going to beat the Patriots?

I have no clue whether Jacksonville is going to beat the Patriots. I can't see how anyone could count on that, given Jacksonville's bipolar nature.

But the only likely way there's a Patriots-Jacksonville rematch in the first round of the playoffs is if Jacksonville gets to 10 wins, and the only way that happens is if they beat the Patriots this Sunday.

126
by Travis (not verified) :: Wed, 12/20/2006 - 12:47pm

I have to believe there is a mistake in the Colts’ defensive rushing statistics. The Colts’ rushing defense is woeful in every way, except that somehow they are 1st in adjusted line yards against “left end� runs. They are first by a LARGE margin. That just doesn’t seem possible unless there have been very few runs attempted in that direction.

There was a FO blog post on this last week.

127
by zip (not verified) :: Wed, 12/20/2006 - 1:28pm

This has very little do with anything else said here, but I really want to see Jax in the playoffs, just because they're so unpredictable. Of course on the road they seem to lay mostly eggs, so... maybe not. But I could see them crushing San Diego.

128
by Sid (not verified) :: Wed, 12/20/2006 - 1:33pm

RE: 18

If Atlanta beats Philly and Philly beats Dallas (and Atlanta wins their other game) and the Giants win their last two, it's a 3 way tie and Atlanta and Philly get the wild cards.
Thus, the Giants don't control their own destiny. But odds are, if they do win their last two games, they get in.

RE: Outside runs on the Colts

It makes sense that they wouldn't work. The ends are almost always rushing straight upfield. A run outside the tackle will run into that. A run between guard and tackle, though, will be effective...

129
by Travis (not verified) :: Wed, 12/20/2006 - 1:44pm

If Atlanta beats Philly and Philly beats Dallas (and Atlanta wins their other game) and the Giants win their last two, it’s a 3 way tie and Atlanta and Philly get the wild cards.

And Dallas has to beat Detroit. A near-given, but you never know.

130
by young curmudgeon (not verified) :: Wed, 12/20/2006 - 3:06pm

Re 30 & 38: yes, and serve it with a side of pasta and a nice bottle of valpolicella, please. Nothing fancy, you understand.

131
by dryheat (not verified) :: Wed, 12/20/2006 - 3:07pm

#113..It's a knee injury, which is probably why I didn't pick up on your intended pun. It also didn't help that I thought of fish instead of feet.

132
by M (not verified) :: Wed, 12/20/2006 - 3:08pm

#122 & #125 - Because the game is at home, and their opponent is not named "Texans". Seriously, the Texans seem to perform better against the Jaguars than any other team for their entire existence.

133
by young curmudgeon (not verified) :: Wed, 12/20/2006 - 3:16pm

Re 109: if you sold your flounder, being shacked up with Adriana Lima wouldn't do you any good! (Sorry, just thought "sold your flounder" sounded vaguely dirty and amusing in a Beavis and Butthead kind of way.)

134
by Crushinator (not verified) :: Wed, 12/20/2006 - 3:23pm

Fan watch - The following teams are who we need to root for this week.

Tennessee (obviously).
Baltimore
New England
Miami
San Francisco (obviously)
San Diego

If all those things happen, the dream of a Titans/Niners SB grow that much closer.

135
by Rich Conley (not verified) :: Wed, 12/20/2006 - 3:26pm

132.

The last time the Jaguars beat the Pats, Andy Katzenmoyer was playing middle linebacker.

Do you really think Garrard isnt going to turn the ball over 3+ times this week? And Del Rio isnt goign to convservative his way out of a win?

136
by Andrew (not verified) :: Wed, 12/20/2006 - 3:27pm

Wanker79:

Is the public perception really that Dallas is a far superior team to Philly even though Philly could take over the division lead with a win this week???

Yes. Dallas is still being spoken of as a likely NFC Super Bowl contender, while Philly is obviously an afterthought in that discussion right now.

137
by Sid (not verified) :: Wed, 12/20/2006 - 3:40pm

RE: 78
They'd have to beat Denver in Denver and their other game, have Seattle lose to San Diego and their other game, host the 5th seed (say, the Eagles), beat the Eagles in San Fran, then go to the 2nd seed and beat them on the road (Dallas? New Orleans?), and then they'd be home in the NFCC game if the 6th seed beat the 3rd seed on the road and then proceeded to go into Soldier Field in January and win.

:)

138
by Andrew (not verified) :: Wed, 12/20/2006 - 4:08pm

Pat and Travis:

Philly offense has 24 fumbles, with 14 lost, while Dallas offense has 14 fumbles, with 5 lost.

On defense, Philly has forced 23 fumbles and recovered 9, while Dallas has forced 25 fumbles and recovered 13.

If we assume for the moment 50% fumble recovery, Philly is down a net of 7 fumbles to Dallas either lost or not recovered (4 on offense, 3 on defense). In other words, Dallas is gaining or retaining an extra possession every other game on average over Philly due to better fumble luck. This translates into 16 more points for Dallas so far this season.

Now, who has the fumble bug?

QB's - McNabb fumbled 3 times and lost 2, Garcia fumbled 5 times and lost 2, Feeley fumbled 1 time and lost it. Romo has fumbled 5 times only lost 1. Bledsoe fumbled 3 times and lost 1.

RB's - Westbrook fumbled 2 times and lost both, Buckhalter fumbled 2 times and lost 1. Jones fumbled 1 time and lost it.

WR's/TE's - Avant and Lewis each fumbled 1 time and lost it, Schobel fumbled 1 time and recovered. Crayton fumbled 1 time and didn't lose it.

KR's - Wynn fumbled 3 times and lost 1, Mahe fumbled 3 times and lost 2. Austin fumbled 2 times and lost none.

C's/LS's - Bartrum fumbled 1 time and recovered, Jackson fumbled 1 time and lost it.

P's - McBriar fumbled 1 time and lost it.

D's - Ayodele fumbled 1 time and lost it.

Most of the difference in recovery rate is from Bledsoe and Romo having their fumbles recovered while McNabb, Garcia and Feeley lost most of theirs. The rest of it is from the not at all missed Dexter Wynn and the unfortunately still with us Reno Mahe.

139
by NewsToTom (not verified) :: Wed, 12/20/2006 - 4:10pm

Re #134
I actually took a look at all the scenarios where the Titans could make the playoffs. Not all the individual games, but I do have a rooting guide. See link in name for 9-7 breakdown. Bonus neat fact: if NYJ, CIN, PIT, TEN, JAX, KC, & DEN all finish 9-7, the Bengals are the 5th seed and the Titans sneak in as the #6 seed.

Is it just my dreaming, or could we use an Open Discussion thread for wacky playoff scenarios?

140
by Kal (not verified) :: Wed, 12/20/2006 - 4:19pm

#135: Do you really think Garrard isnt going to turn the ball over 3+ times this week? And Del Rio isnt goign to convservative his way out of a win?

Just based on prior games - the Jags don't need a big offense to win at home, their defense does a great job of doing it all by themselves. One might ask the same question - do you really think that Brady isn't going to turn it over 3 times and Faulk won't fumble it a bunch, or that they'll drop it on special teams? Add to this that the Pats aren't the best at run defense (they're much better at pass defense than run defense) and it...

oh, to hell with it. I'm predicting space aliens from Mars implanting eggs in the Pats abdomens. It's about as likely as anything else that might happen with the Jags.

141
by Travis (not verified) :: Wed, 12/20/2006 - 4:31pm

Bonus neat fact: if NYJ, CIN, PIT, TEN, JAX, KC, & DEN all finish 9-7, the Bengals are the 5th seed and the Titans sneak in as the #6 seed.

This assumes the Titans beat the Steelers and Jets on strength of victory. It's likely (the Titans are currently ahead of the Steelers by about 6 wins, assuming both win out, and the Jets are well back), but the Steelers might come out ahead thanks to some favorite results (Miami, New Orleans, and Carolina winning, for starters).

Is it just my dreaming, or could we use an Open Discussion thread for wacky playoff scenarios?

I second this.

142
by Travis (not verified) :: Wed, 12/20/2006 - 4:34pm

That should be favorable results.

143
by Travis (not verified) :: Wed, 12/20/2006 - 4:48pm

Oh, and in that scenario, the Week 17 game between Cleveland and Houston has ridiculous importance, because the Steelers have beaten the Browns twice, while the Titans have done the same to the Texans.

144
by Rich Conley (not verified) :: Wed, 12/20/2006 - 4:58pm

140

Kal, any effect of the Jags playing much better at home is offset by how much better the Pats have played on the road. I think thats a wash.

145
by young curmudgeon (not verified) :: Wed, 12/20/2006 - 5:16pm

Rich, I'm not trying to be a jerk or anything, but aren't you drifting closer and closer to saying that it's unimaginable that Jacksonville beats NE this week? You started off with some salient points, but (unless I'm misreading your tone here) in each post you seem to up the ante. I agree with you that NE is more likely to win than Jax, but given the Jax track record this year, I'd say Garrard is just as likely to throw four TD passes, run for 100 yards, or levitate his body out of the stadium as he is to throw 3+ ints.

146
by NewsToTom (not verified) :: Wed, 12/20/2006 - 5:16pm

Re #141
Right, I should have realized the difference between .398 and .480 was only 8 wins, and it'd be 6 if the current record of the remaining opponents are taken into account. It's hard to believe CLE-HOU could decide who makes the playoffs, but such is the wonderful world of NFL tiebreakers.

And, I know I should be happy about this as a Titans fan, but it annoys me from a position of rightness that TEN would beat NYJ and PIT in a 3-way playoff, but lose to both teams head-to-head.

147
by GlennW (not verified) :: Wed, 12/20/2006 - 5:28pm

> "And, I know I should be happy about this as a Titans fan, but it annoys me from a position of rightness that TEN would beat NYJ and PIT in a 3-way playoff, but lose to both teams head-to-head."

That is strange. The NFL should probably apply the "one team sweeps or is swept by" application to all the successive 2-team tie-breakers and not just to head-to-head games. But admittedly, the tie-breaker of record against common opponents (when only four opponents are involved) is just as arbitrary as strength of victory, if not more so.

FWIW, if you consider remaining games in which TEN's defeated opponents play each other (there are 3) as well as PIT's (2), TEN's SoV advantage over PIT moves from +6 to +7. Realistically PIT is not going to overcome that deficit.

148
by Travis (not verified) :: Wed, 12/20/2006 - 5:54pm

FWIW, if you consider remaining games in which TEN’s defeated opponents play each other (there are 3) as well as PIT’s (2), TEN’s SoV advantage over PIT moves from +6 to +7. Realistically PIT is not going to overcome that deficit.

Except that a NYJ-PIT-TEN scenario only comes about with a bunch of other results, which likely include at least some of the following:

- KC, a PIT opponent, winning twice, including over JAX, a TEN opponent
- MIA, a PIT opponent, beating NYJ
- CIN, a PIT opponent, beating DEN

Throw in CLE, a twice-beaten PIT opponent, beating TB, and a PIT strength-of-victory win begins to seem plausible.

149
by Rich Conley (not verified) :: Wed, 12/20/2006 - 6:09pm

". I agree with you that NE is more likely to win than Jax, but given the Jax track record this year, I’d say Garrard is just as likely to throw four TD passes, run for 100 yards, or levitate his body out of the stadium as he is to throw 3+ ints."

No, I'm not saying that Jacksonville has no chance, I just had a problem with Travis basically saying it was a foregone conclusion (which isnt what he meant, but his post sounded like that), and people just looking at DVOA and saying JAX has 2% higher DVOA, and is at home, so they'll win. THats lazy, its ignorant, and its against everything this site is for.

Its going to be a good game, but I think that just like Jax is a bad matchup for Indy, NE is a bad matchup for Jax.

You've got 2 good defences, and one good offense. The team with the good offense has better special teams. Jax can only run the ball, so NE is goign to stack the line, and Garrard is going to be forced to stay in the pocket, and throw the ball a LOT.

150
by CaffeineMan (not verified) :: Wed, 12/20/2006 - 6:18pm

I can't feel confident at all about the Pats-Jags game. What a freakin' crapshoot. Jags home record. Pats road record. Jags rising on defense. Pats sinking on offense. Jags inconsistent on offense. Pats rising on defense. Jags crappy passing game and great running game makes them one-dimensional. Pats can't put the ball in the end-zone against a bad defense (Texans). Both teams' variances are off the charts. WTF is gonna happen?

I bet if either teams gets up by 2 TD's, for whatever reason, the game's over. Neither team can be successful in do-or-die passing situations at this point, because both defenses are good enough to shut down a one-dimensional attack. I guess I'd have to say edge to the Jags unless Wilfork is back at NT and healthy enough to be effective. Wright is a nice flexible sub, but he got pushed around by the Texans and Vrabel-Bruschi is worse than Seau-Bruschi against the run.

151
by dryheat, posing as Sean Salisbury (not verified) :: Wed, 12/20/2006 - 6:33pm

Right there with you, #150. If Wilfork plays, I feel a lot better about NE's chances. Of course, if Watson and Maroney play as well, I feel better to the point where I'd throw down significant cash.

I expect the Pats to load up against MJD, and only rush four vs. Garrard. If Garrard can be successful throwing the rock against seven covering (or six + a spy), Jags should win.

In contrast, the best chance for the Pats would seem to be to test the Jags on the edges (hence Maroney) and to try to exploit their LBs/backup safety in coverage against Faulk and the tight ends.

If the O-line can keep Brady clean, the Pats should win this one.

152
by chris clark (not verified) :: Wed, 12/20/2006 - 6:51pm

wacky playoff scenarios:

I'd love a thread on that. And, from some of the comments, others have looked into the scenarios at some depth. For example, last week any of the 4 AFC divisions could have sent 3 teams to the playoffs. I still think that is true, but I haven't verified it yet. I do know that NE winning (over JAC), allows NYJ to control their own destiny. I haven't found any other win/loss combinations that have given teams control over their destiny (which don't already have it). In fact, given that currently CIN, DEN, and JAC are the 3 teams with control over their destiny and one of CIN or DEN loses control over their destiny this week. It seems likely that next week only two AFC teams will have control over their own destiny--the winner of CIN/DEN and either JAC or NYJ depending on the outcome of NE/JAC. Is that the case, or is their a combo of games this weekend where JAC/KC determines which of them go to the playoffs that I haven't found yet? As far as I can tell there isn't for this weekend, because among other things NYJ needs to lose twice and BUF and TEN each lose once for the KC to go to the playoffs, and of course DEN and CIN cannot both go if KC is to go.

153
by NewsToTom (not verified) :: Wed, 12/20/2006 - 7:03pm

Re #152
Off the top of my head...
CIN beats DEN and PIT to finish 10-6. In as a WC. 1 WC spot left. PIT no better than 8-8.

DEN beats SF to finish 9-7.

TEN/BUF winner loses in Week 17. TEN/BUF 8-8/7-9.

NYJ loses twice (MIA, OAK), to finish 8-8.

JAX beats NE, goes into Week 17 9-6.

KC beats OAK, goes into Week 17 8-7.

Under this scenario, if KC wins, KC finishes 9-7 with DEN and JAX. KC beats DEN based on division tiebreaker, and beats JAX H2H. If JAX wins, JAX finishes 10-6 with CIN and gets the second wild card.

I think if DEN loses to SF to finish 8-8, it doesn't matter if JAX beats NE or not. Because, then, the only 9-7 would be the winner of JAX-KC.

154
by young curmudgeon (not verified) :: Wed, 12/20/2006 - 7:12pm

Re 149: Rich, thanks for your response. Rational discussion and analysis is always interesting--now, watch the game be decided on a funny bounce, a missed call, and a fluke deflection.

155
by Travis (not verified) :: Wed, 12/20/2006 - 7:31pm

Scenarios where each AFC division gets 3 teams into the playoffs. There may be other scenarios than the one given, and as always, it is important to remember to break division ties first (which is why DEN might not make it, even with an 8-4 conference record):

East
- NYJ wins twice.
- BUF wins twice.
- DEN beats CIN, but loses to SF.
- KC beats JAX and OAK.
- CIN loses to PIT.

NYJ, at 10-6, is the #5 seed. BUF, at 9-7, wins tiebreakers against KC (and possibly PIT and TEN) due to their 7-5 conference record.

North
- CIN beats DEN.
- PIT beats CIN and BAL.
- JAX loses to NE and KC.
- KC beats OAK.
- BUF loses at least once.
- NYJ loses at least once.
- TEN loses at least once (or PIT has strength of victory tiebreaker).

CIN, PIT, and KC (and possibly TEN and DEN) are all 9-7. CIN has the tiebreaker over PIT and all others, thanks to their 7-5 conference record, and is the #5 seed. PIT has the tiebreaker over KC, thanks to their 6-6 conference record, and is the #6 seed.

South
- DEN beats CIN, but loses to SF.
- CIN loses to PIT.
- PIT beats BAL.
- JAX beats NE, but loses to KC.
- TEN wins twice.
- NYJ wins once, and loses once.
- KC beats OAK.
- TEN has the strength-of-victory tiebreaker over NYJ and PIT.

NYJ, PIT, JAX, TEN, KC, and DEN are all 9-7. TEN beats JAX thanks to their 4-2 division record, and beats NYJ and PIT on strength-of-victory (KC is eliminated due to a 5-7 conference record). JAX beats the others on conference record.

West
- DEN beats CIN
- PIT beats CIN
- BAL beats PIT
- KC beats JAX and OAK.
- TEN loses at least once.
- NYJ loses twice.
- BUF loses at least once.

DEN has 9 or 10 wins, and KC has 9, while JAX has at most 9, and no other AFC team has more than 8. KC has the tiebreaker on JAX due to their head-to-head victory, and DEN has the tiebreaker due to conference record.

156
by chris clark (not verified) :: Wed, 12/20/2006 - 7:32pm

wacky playoff scenarios:

153: yes, that's roughly what I've come up with, so the game can decide which team makes the playoffs, but not without other week 17 games going specific ways.

I couldn't find a JAC + TEN playoff scenario. I could find CIN|DEN|NYJ|KC|PIT + TEN scnerios. Perhaps, if CIN and DEN tie (and then both lose the following week), the software I'm using to calculate playoff positions won't left me enter ties.

157
by Travis (not verified) :: Wed, 12/20/2006 - 7:34pm

A couple of corrections (why don't I ever notice these when previewing?):

East - Buffalo would beat Jacksonville on head-to-head, if it (eventually) came down to just those two teams.

North - Pittsburgh would beat Kansas City on head-to-head, if it came down to just those two teams.

158
by GlennW (not verified) :: Wed, 12/20/2006 - 7:43pm

> "Except that a NYJ-PIT-TEN scenario only comes about with a bunch of other results, which likely include at least some of the following:"

Right, some (not all) of those results are required (DEN beating CIN is more likely to end with a NYJ-PIT-TEN tie than the other way around, for example). But there are only 15 relevant games left which include PIT and TEN opponents, 5 of those head-to-head. Even if PIT's opponents win all 5 of those games (including the double-bonus CLE-HOU game, for a +6 gain), PIT still needs a 6-4 favorable result in the remaining 10 games (11-4 total) to take the SoV tiebreaker. Move one of those head-to-head games over to the TEN column, and PIT needs an 8-2 result in the remainder (12-3 total) to win SoV (11-4 to tie), etc.

Making up a net -7 SoV over only 15 games is a huge longshot unless there's a big quality differential between the teams' opponents (and here TEN's opponents have been better so far this season)-- in this case basically like winning 11 of 15 coinflips including a majority of the pre-defined double-credit ones. Even if you assume a couple results as givens, I think TEN can rest easy on this tie-breaker (if on nothing else).

159
by Travis (not verified) :: Wed, 12/20/2006 - 7:44pm

Another correction:

South: Jacksonville would beat New York and Pittsburgh because they beat both head-to-head (KC would be eliminated on conference record).

160
by Rich Conley (not verified) :: Wed, 12/20/2006 - 7:45pm

154.

It totally could come down to a fluke like that, especially if that fluky play happens early in the game. If NE comes out and scores a TD, and then Jacksonville fumbles the kick, or Garrard throws a pick, and NE scores again, the game is pretty much over, and it'll end up being a blowout, because Jax will have to throw all day.

That being said, if Jax hops up 14 points, theyre just going to run the ball down NE's throat for the rest of the day, and may force NE to go pass wacky.

I do think NE has a slightly better chance of coming back than Jax though.

161
by chris clark (not verified) :: Wed, 12/20/2006 - 7:47pm

155: thanks!! (and fast too). I tried setting the games you listed upto the TEN strength-of-victory condition and got ~90% TEN + JAC, 10% PIT + NYJ, which I assume accounts for the strength-of-victory probabilities.

162
by Travis (not verified) :: Wed, 12/20/2006 - 8:05pm

Re: 158

Yup, not probable, but plausible (the 10% figure quoted in 161 sounds about right). I mentioned it because we're discussing wacky playoff scenarios, and what's wackier than a Week 17 Houston-Cleveland matchup that actually means something?

163
by Crushinator (not verified) :: Wed, 12/20/2006 - 8:44pm

NYJ, at 10-6, is the #5 seed. BUF, at 9-7, wins tiebreakers against KC (and possibly PIT and TEN)

Just an addendum, there's no way for Buffalo and Tennessee to both go 9-7, so a tiebreaker with them wouldn't matter.

164
by Kal (not verified) :: Wed, 12/20/2006 - 9:28pm

Boy, do I want the week 16 games to get here - if only to deal with these insane tiebreaker issues and weird-ass playoff scenarios.

165
by Stillio (not verified) :: Fri, 12/22/2006 - 5:04am

The reason Jax will beat NE: 191-64. That's how much the Jags have outscored their opponents at home this year. They've beaten Dal, Pit, NYJ, NYG, Ten and Ind and haven't allowed more than 17 points at home all year. Also, they 've been leading every one of those games by at least two scores with 2 minutes to play. The only glitch is that they did manage to lose to the Texans, but they still only gave up 13 points. As usual for the Jags, this game is all about the offense; if they show up and don't turn it over then the Jags win big. If Garrard has 3+ turnovers again, then the Jags lose. Considering that Garrard only has two such games out of eight this year, I'll pick the Jags.

166
by Mike H (not verified) :: Fri, 12/22/2006 - 12:14pm

A new version of the Playoff Odds Report is up, with the tiebreaker problems fixed. As a bonus, "Sorry I Screwed Up" feature, the report now links to a tiebreaker breakdown page, which shows every scenario and its frequency. That explanation leaves a lot to be desired, I know, but just check it out and you'll get what I mean.

167
by Pat (not verified) :: Fri, 12/22/2006 - 12:21pm

NFC chance of winning the Super Bowl: 44.8%
AFC chance of winning the Super Bowl: 55.2% (obviously)

Sigh. Yup. NFC sucks.

168
by dryheat (not verified) :: Fri, 12/22/2006 - 2:56pm

Oh....and the Pats just signed Todd Sauerbrun. Which is a huge upgrade on-field.

169
by chris clark (not verified) :: Fri, 12/22/2006 - 3:57pm

166: Impressive, but I didn't think you had screwed up previously.

Just to make sure I understand the tiebreaker scnario parts. Each line represents a unique scenario where the team in column 1 (e.g. CIN) gets the playoff seed in column 2 (e.g. 5) beating out the other teams with the same number of wins in column 3 (e.g. NYJ,DEN) using the tiebreaker rule in column 4 (e.g. h2h_winner). The number of wins those teams all had is column 5 (e.g. 9) and the number of simulations out of 10k where this occured is in column 6 (e.g. 45).

(Apologies KAL, since I know this bores you, but....)

So, from this, we can determine in those 45 simulations JAC (for example) won at most 8 games, because CIN got the 5th seed and JAC can't get higher than 5th and since JAC isn't listed in the OTHERTEAM column, it must have gotten fewer wins. Therefore, for those 45 simulations, either NYJ or DEN must have been the 6th seed, because they were the only other 9 game winners. Then, by looking at the secnarios where DEN and the NYJ are listed at 6th seed and 9 wins, we find only DEN listed with a NYJ opponent, so we can conclude that those 45 secnarios are all CIN 5th and DEN 6th seed scenarios. One could probably do that for all the lines in the table (doesn't that sound like a fun task) and thus figure out what all the playoff combinations are and what their probabilities are. I guess I'll pass on that.

Hmm, you could get your software to do it though. Not that what you have done isn't wonderful, but there are *always* improvements. I bet the number of playoff combinations is roughly the same as the number of lines in your current chart (at least for the AFC), maybe the combinatorics explode for the NFC.

In fact, I suspect you could add a couple of columns to the current chart and list what happened to the other teams (i.e. which seeds they got if any), or better yet (for presentation purposes) you could add their seedings to the OTHERTEAMS column. Thus, for the line I was explaining above, the OTHERTEAMS column could read DEN(6th), NYJ(out) and that would capture everything. The only complicating case is where the remaining teams had a variety of tiebreaker resultsfor the same CIN 5th over DEN and NYJ scenario, for example if DEN won some of the tiebreakers scenarios and NYJ won some, for example if it was due to point differentials. Then, you would need to split the current line into 2 lines. One for the scenarios where DEN won the subsequent tie breaker and the other for when NYJ won the tie breaker.

That's not the only presentation possibility and may or may not be the easiest to implement. If you want to discuss the idea further, let me know and I can drop you an email line, and discuss how you the presentations might look and some implementation ideas. (If you don't want to do it, or want to do it on your own, I won't be offended if I don't hear.)

170
by Andrew (A.B.) (not verified) :: Fri, 12/22/2006 - 5:28pm

Mike -- Found at least one error in your tiebreaks. In a three-way tie, Philly is given the tiebreak over GB & Atlanta because of "h2h_winner". But there is no way that Philly and Atlanta can be tied without Atlanta beating Philly in week 17.

171
by Pete C (not verified) :: Fri, 12/22/2006 - 5:53pm

Is it just me, or has anyone else noticed a huge pro-Denver bias by the ESPN's of the world this week? Thankfully this week's DVOA brings some semblance of reality to the table. Denver beats up on the atrocious Cardinals, and all of a sudden Jay Cutler is tremendous.

172
by B (not verified) :: Fri, 12/22/2006 - 6:10pm

168: Is there a pool on how Saurbraun will meet his demise? I'll put $10 on "chocked to death on vomit, but not his own."

173
by Mike H (not verified) :: Fri, 12/22/2006 - 6:37pm

Re: 170

The largest flaw in the presentation currently is that, in a case where:

1. 3 Teams are Tied
2. A tiebreak is used to narrow down to 2 teams, and the tiebreakers reset
3. The 2-team tie is broken

The "tiebreaker used" column will only list the method used to break the 2 team tie.

What I assume happened in this case was that some other tiebreaker was used to narrow down to Philly and GB, which was then broken by h2h. I would have to check further to verify.

Sorry for the confusion.

P.S. Will be responding to Chris Clark next.

174
by Travis (not verified) :: Fri, 12/22/2006 - 7:17pm

Mike:

I think you mean "strength-of-victory," instead of "score_margin_winner." Score margin used to be the wild card tiebreaker after common opponents, but the NFL changed the order of tiebreakers during realignment before the 2002 season.

175
by Travis (not verified) :: Fri, 12/22/2006 - 7:22pm

What I assume happened in this case was that some other tiebreaker was used to narrow down to Philly and GB, which was then broken by h2h. I would have to check further to verify.

Conference record. Philly and GB would be 7-5 in the NFC, while Atlanta would be 6-6.

176
by Travis (not verified) :: Fri, 12/22/2006 - 7:37pm

Also, some of the tiebreakers still depend on what happens in future games.

Assume the Giants and Green Bay tie for the last spot in the NFC at 8-8. If the Giants beat New Orleans and lose to Washington, they'll beat Green Bay based on common opponents (Philly, Chicago, New Orleans, Seattle), 2-3 vs. 1-4. But if the Giants lose to New Orleans and beat Washington, then they go to strength-of-victory.

Similar things also happen in tiebreakers involving the 8-8 Eagles vs. the 8-8 Giants; 8-8 Atlanta vs. 8-8 San Francisco or St. Louis (dependent on whether ATL beats PHI), and maybe a couple others.

177
by Travis (not verified) :: Fri, 12/22/2006 - 7:40pm

Re: 176

Never mind, I just noticed that you present both potential tiebreakers, albeit a few lines away from each other.

178
by Travis (not verified) :: Sat, 12/23/2006 - 12:42pm

Other Chris:

NFL-Forecast.com shows that the Panthers have a non-zero chance of making the playoffs at 7-9. I can't think of a scenario where this occurs - their potential 5-7 conference record (the Giants already have 6 conference wins), plus their head-to-head loss to the Giants seems to make this impossible. Am I missing something?

179
by John (not verified) :: Sat, 12/23/2006 - 11:55pm

I dunno, I think the Lions are still ranked way too high at #31

180
by the other Chris (not verified) :: Sun, 12/24/2006 - 3:58am

Travis - You're right. I traced down the case that was giving the error. It involved 6 teams at 7-9. After breaking ties within the divisions, that left 4 teams: NYG, GB, Carolina, and SF.

At the conference tiebreaker, NYG, GB, and SF all had 6-6 conference records, while Carolina should have been eliminated at 5-7. My first glance at the code seems to indicate that I forgot to handle the case where 3 of 4 teams are tied and only the bottom team is eliminated, or maybe I left out the 4-team tiebreaker all together.

Thanks, I do appreciate finding the bugs so that I can fix them. Probably won't try to fix it until Tuesday or so.