Efficiency Rankings - Week 6

Two of the top three teams lost but still kept their spots. San Diego, in particular, lost to a far weaker team, but did so thanks to special teams play. One criticism the system receives is that it doesn't include special teams.

It's not that special teams don't matter or can't affect game outcomes. The reason they're excluded from the model is that they are wildly inconsistent from week to week. So while their explanatory power can be significant, they are not predictive. And because special teams performance is not predictive, is it really a mark of good or bad team? Or is it just randomness, subject to a handful of non-repeating plays in each game, of which there are only 16 in a season?

The NFC South looks particularly weak so far. The teams range from 21st to 27th. The NFC West might be even weaker, featuring the two bottom teams and led by 17th ranked Seattle.

As crazy as these rankings may seem (and they do to me to), they did predict several upsets (at least according to the point spreads). Several of these apparently qualify as shocking to many media outlets: ARI over NO, WAS over GB, PHI over SF, CHI over CAR. In the 5 games in which the model disagreed with the consensus, it went 4-1. In other words, consensus favorites were 5-9 last week, and the model was 8-6.

Click on the table headers to sort.

1 SD10.810.42110
2 NYG20.740.5442
3 KC30.660.56231
4 PIT80.640.49123
5 DAL40.600.58216
6 GB140.600.49514
7 IND130.600.57819
8 CHI90.600.53164
9 WAS50.590.52715
10 MIA110.590.481513
11 HOU60.580.57621
12 PHI150.550.471411
13 TEN160.550.55135
14 BAL70.540.50199
15 NE120.540.47330
16 MIN210.530.47208
17 SEA180.510.50277
18 NYJ100.500.511812
19 DEN200.480.51927
20 CIN170.460.462518
21 TB230.450.471032
22 SF250.440.512220
23 NO270.430.401131
24 ATL280.410.432423
25 CLE190.400.502124
26 JAC260.380.562829
27 CAR220.370.54326
28 DET290.350.511728
29 BUF240.340.522925
30 OAK300.300.492626
31 STL310.270.403022
32 ARI320.240.443117

Team efficiency stats are listed below.


  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

25 Responses to “Efficiency Rankings - Week 6”

  1. James says:

    San Diego's GWP is 0.81! That seems crazy high to me.

  2. James says:

    After looking back at last year's rankings, 0.81 isn't that high as the Colts and Saints varied between 0.77 to 0.87 for most of the season.

    The real surprise is the severe drop off between 1st and 3rd. Last year's rankings had a lot of teams in the 70s and a gradual decline.

  3. Brian Burke says:

    Yeah. SD should be a case study on why passing efficiency doesn't say it all. When I see a rating that high, I want to crank up the regression factor, but it's already insanely high. SD has just been putting up blistering numbers.

  4. Anonymous says:

    How are the Bears 8th overall?

    Drilling into the raw efficiency metrics (SR%, EPA/P, WPA, etc), CHI is in the bottom 5 in most offensive categories, yet ranked 16th above. Similar story with the Defense, 4th in the above chart, but never in the top five of the raw efficiency metrics.

  5. Anonymous says:

    Hmmm...Atlanta's drive stats from footballoutsiders is great. Fourth in net drive success rate and second in net turnovers per drive.

    Interesting they're so low here against league average teams.

  6. Brian Burke says:

    ATL has very average (slightly lower than avg actually) efficiency stats, but good turnover rates. Turnovers are strongly regressed at this point in the season. Their SoS has been anything but avg--28th easiest.

    I don't think FO's methodology regresses turnovers enough, and they're over-weighting red zone performance. That might be very explanatory of past outcomes, but it's not a stable predictor of future outcomes. So their rankings might jibe better with many people's intuitive sense of which teams are good and bad.

  7. Anonymous says:

    I don't pay much attention to FO's efficiency stats, I'm speaking about "drive stats" (percent of down series leading to first down or touchdown). I find it's a pretty good indicator.

    My mistake on their SOS, I accidentally looked at Cleveland's Opp GWP.

  8. Juri says:

    I thought the SD-OAK game was the perfect validation of your rankings. The only way Oakland could win that game was by crazy luck - and behold it came to pass.

    It does seem like the Norv Turner coached San Diego consistently underperforms - in that they don't get the amounts of wins they should according to their efficiency stats.
    Likewise New England consistently overperformed during their cheating days. Didn't you do a study on that Brian?

  9. Brian Burke says:

    Right. DSR or something. ATL must have a very good 3rd down conversion rate.

  10. Brian Burke says:

    Juri-That's interesting. Here's my 1-minute theory: Run Success Rate. Just looking at the past couple years--

    NE Run SR: #2 and #1
    SD Run SR: #28 and #23

    This year so far:
    NE: #1
    SD: #24

    Maybe I was wrong on Belichick's secret to his success. It might be that he has long understood what's really important in the running game.

  11. coldbikemessenger says:

    "I don't think FO's methodology regresses turnovers enough,..."

    what do you mean by regresses turnovers?

  12. Brian Burke says:

    CBM-The Falcons have a defensive int rate of 6.0%, but that's unsustainable. Turnovers are highly random (about 80%), meaning that there is almost no consistency from one game to the next, or from the early half of the season to the latter half.

    If you want to estimate ATL's most likely future interception rate, you need to regress to the mean by a considerable degree. It's almost certainly not going to be 6%. It's far more likely to be closer to the league average than where it is now.

    By the end of the season it's probably going to be above average, but that's only because they've already got 5 games of a 6% rate under their belt. It's not important what their final rate is; What's important is what their rate will be between now and the end of the season.

    It seems to me, just from occasionally observing their weekly rankings, that FO is either unaware of the tendency to regress, or does not regress nearly enough, especially with turnover stats. But I can't say that for sure because I don't know what their recipe is.

  13. coldbikemessenger says:

    didn't you do a post a while back on incompletions and interception rates? Last year there was one INT for every 12.7 incomplete passes
    this year there is one INT for every 12.4 incomplete passes. Atlanta is leading the NFL with one INT for every 5.6 incomplete passes, so yeah, they are probably not gonna keep that up.

  14. Jim Glass says:

    Yes, the 2-3 Chargers may be bearing witness to the idea that running success matters after all, leaving Vince Lombardi smiling in his grave.

    They are indeed blowing through yardage, to quote PFR.com...

    "... the absurdly inefficient San Diego Chargers, who have piled up the yards but left the wins behind. The Chargers have the second best yardage different in league history through five games. They've gained 1,078 more yards than they allowed, more than twice as high a differential as any team, ever, with a losing record after five games."

    But they are only 24th at run success rate. Speaking of regression to the mean, what's going to move most during the rest of the season -- that success rate up towards those yardage stats, or that yardage down towards normality?

  15. Jim Glass says:

    Although looking at the game logs I see the Chargers were murderized in their three losses by *six* huge kick returns, two blocked punts and three lost fumbles.

    All probably a bit more significant to the Ls than that run success rate. Sorry, Vince.

  16. Jim Glass says:

    Not to overdue a point (or maybe to) but FOers ranks the SD special teams performance as far below #31 Miami's as Miami is below #13 New Orleans.

    So we could consider the SD special teams as being ranked #49. That will probably regress towards the mean a bit during the rest of the season.

  17. Chris says:

    The thing about SD is that with average special teams play, they would most likely be 5-0, and then no one would think twice about their high efficiency ratings. If they can coach up their special teams just a little, I think their wins will start matching their efficiency ratings.

  18. James says:

    Ok, so special teams regression/improvement should fix San Diego's problems. Now how do we figure out what to do about the 5th ranked 1-3 Cowboys?

  19. Anonymous says:

    The Jets are running the ball better than SD. Flip the running games and their records would probably flip, too.

  20. Anonymous says:

    How does Baltimore drop 7 places on a week where they completely dominated their opponent in every phase of the game.

  21. Brian Burke says:

    Baltimore had a solid win against an inferior opponent, just like we'd expect. But it was not a domination.

    The Ravens dropped primarily due to their SoS. It's not that Denver was so weak, it's that their previous opponent's ratings dropped--CIN and CLE. Plus, the Ravens were at the head of a very tightly grouped pack of teams, some of which leapfrogged them.

  22. Unknown says:

    Brian, can you comment on why the AFC East teams' ratings (other than Buffalo) seem so wacky? They are very different from FO and very different from what I would expect based on watching the 3 teams. I'm guessing part of why the Jets are so low is because their TO margin has been unsustainably good?

  23. Anonymous says:

    I think an interesting post to put up along with your rankings is how many wins a team would have if it played against all other teams on a nuetral field. I did this with my system here:


    This is the first week my rankings will be on Todd Becks prediction tracker. I created them 2 weeks ago, and the last 2 weeks it is 18-10.

    Do you scale a teams chances of winning based on recent performance? Like if the NYG are doing well the last 5 games, but did horrible the 5 games before that, do you take that into account and predict that that trend continues, or do you let their entire season performance be the sole predictor from week to week?

  24. Andy says:

    Brian, do you think you could (or did you already) post something about the run/pass SR and winning correlations? Do the run/pass SR's overtake efficiencies as winning indicators?

  25. Brian Burke says:

    Andy-I'm working on that. For passing no, but it's close. For running yes. I'm experimenting with some hybrid models that might improve on the current model.

Leave a Reply

Note: Only a member of this blog may post a comment.