Week 6 Efficiency Rankings

The ratings are listed below in terms of generic win probability. The GWP is the probability a team would beat the league average team at a neutral site. Each team's opponent's average GWP is also listed, which can be considered to-date strength of schedule, and all ratings include adjustments for opponent strength.

Offensive rank (ORANK) is based on offensive generic win probability is based on each team's offensive efficiency stats only. In other words, it's the team's GWP assuming it had a league-average defense. DRANK is is a team's generic win probability rank assuming it had a league-average offense.

GWP is based on a logistic regression model applied to current team stats. The model includes offensive and defensive passing and running efficiency, offensive turnover rates, and team penalty rates. A full explanation of the methodology can be found here. This year, however, I've made one important change based on research that strongly indicates that defensive interception rates are highly random and not consistent throughout the year. Accordingly, I've removed them from the model and updated the weights of the remaining stats.

10 ARI80.640.631319
11 ATL190.630.42924
24 BAL210.370.462812
14 BUF120.540.342114
4 CAR50.710.54101
6 CHI90.700.5485
27 CIN280.280.553018
28 CLE310.270.571928
7 DAL70.690.5049
16 DEN110.530.471125
32 DET320.120.563132
22 GB230.450.451815
21 HOU260.460.581630
19 IND200.520.471217
20 JAX250.500.581422
31 KC300.130.523231
9 MIA60.660.51723
18 MIN150.520.521510
26 NE220.310.512529
8 NO130.670.52611
5 NYG20.710.33216
15 NYJ100.540.50264
23 OAK160.410.572420
3 PHI40.780.5658
12 PIT140.610.43203
2 SD30.810.51113
25 SS270.360.432921
29 SF240.270.522726
30 STL290.250.662227
13 TB170.610.63176
17 TEN180.530.39237
1 WAS10.830.6032


  • 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

9 Responses to “Week 6 Efficiency Rankings”

  1. Brian Burke says:

    Oops. Just click on the Rank header to sort in order of team strength.

  2. Anonymous says:

    Why do you think the Rams stay at #1 despite losing to the 29th best team (from last week's rankings?) Was that loss not as bad as it seemed (not sure how) or is there something else going on here?

  3. Brian Burke says:

    I was surprised too. Thought they would fall dramatically, but they didn't for 3 main reasons: They still have pretty good eff stats, they still have zero ints and a very low fumble rate, and they have had one of the toughest schedules.

  4. Anonymous says:

    More surprising than the Skins staying on top was the Panthers moving up a notch after an awful showing in Tampa. What happened there?

  5. Anonymous says:

    There is something I don't understand in the model. I ran the regression with last year coefficients, and set the AHOME at 0.5 for neutral site. Then I took Washington's stats for the "A" coefficients, and the average's stats for the "B" coefficients. I found a GWP of 77% without adjustments. Then I reversed the A and B stats and found a GWP of 21%.

    So Washington playing an average team on neutral site has a GWP of 77% but an average team playing Washington on a neutral site has a GWP of 21%. Am I missing something?

    Second remark on this, why both percentages don't give 100%.

  6. Brian Burke says:

    Re: the Panthers. It could be a couple things--the teams they played previously may have done very well last week, improving their SoS. Also, even thought they lost, they might not have had terrible "internals" or things my model doesn't capture like punt returns or missed fld goals. Another thing could be that the team(s) ahead of them did even worse.

    Re:Washington vs avg team on a neutral site. Don't forget the constant. It's about -.3 or something. That might fix it.

  7. Anonymous says:

    Looks like the Panthers moved up despite a poor showing simple because the Giants were ahead of them and had a poorer one.

  8. Anonymous says:

    Do you think that the win probabilities are accurate at the tails? I find it hard to believe that any NFL team has a 5% or less chance of beating any other. Yes, I thought this even before the Rams and Browns won.

  9. Brian Burke says:

    SR-You might be right about that. That's one of the criticisms of logit regression in general. I've already added in pretty strong 'regression to the mean' corrections so that the game probabilities aren't overconfident based on short term outlier team performance.

    One way to check for overconfidence is to calculate the theoretical predicted accuracy of the regression and compare it to the actual accuracy. I've only done that for one season, 2007, and it was dead-on--72% predicted and 72% actual. So I'm not sure if further corrections are wise.

Leave a Reply

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