Team Efficiency Rankings - Week 16

The team rankings below are 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 offensive generic win probability, which 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, defensive interception rates, and team penalty rates. If you're scratching your head wondering why a team is ranked where it is, just scroll down to the second table to see the stats of all 32 teams.

Click on the table headers to sort:


RANK
TEAM
LAST WK
GWP
Opp GWP
O RANK
D RANK
1
IND
2
0.79
0.50
5
4
2
NO
1
0.78
0.48
1
17
3
SD
3
0.77
0.49
2
19
4
PHI
4
0.75
0.51
9
3
5
DAL
8
0.75
0.54
3
13
6
NYG
6
0.74
0.57
6
9
7
DEN
5
0.71
0.55
16
2
8
NE
7
0.70
0.53
4
20
9
PIT
10
0.64
0.44
7
15
10
BAL
12
0.62
0.49
13
5
11
GB
9
0.62
0.39
11
7
12
NYJ
11
0.60
0.49
23
1
13
HOU
13
0.59
0.48
8
21
14
TEN
15
0.58
0.53
12
16
15
MIN
14
0.55
0.39
10
23
16
WAS
16
0.54
0.51
19
11
17
CIN
17
0.54
0.48
18
6
18
JAC
19
0.47
0.49
14
26
19
ARI
18
0.47
0.43
15
18
20
ATL
21
0.43
0.56
17
28
21
CAR
24
0.43
0.53
21
12
22
BUF
22
0.41
0.50
26
8
23
MIA
23
0.40
0.57
20
24
24
SF
20
0.38
0.48
22
10
25
CHI
25
0.31
0.45
25
14
26
TB
27
0.30
0.56
27
22
27
SEA
26
0.27
0.42
24
25
28
KC
28
0.24
0.55
29
29
29
OAK
29
0.21
0.61
30
27
30
STL
30
0.16
0.48
28
31
31
CLE
31
0.14
0.52
32
30
32
DET
32
0.13
0.49
31
32

And here are the sortable raw team efficiency stats. Passing, running, and penalties are in yards per relevant play. Fumbles and interception stats are in turnovers per relevant play.


TEAM
OPASS
ORUN
OINT%
OFUM%
DPASS
DRUN
DINT%
PENRATE
ARI
6.5
4.2
2.9
1.6
5.9
4.6
3.2
0.45
ATL
6.0
4.1
3.0
0.7
7.1
4.2
2.3
0.32
BAL
6.4
4.5
2.6
0.7
5.9
3.6
4.4
0.54
BUF
5.2
4.3
4.4
0.8
5.4
4.8
5.8
0.43
CAR
5.5
4.7
4.7
1.6
6.0
4.6
4.5
0.37
CHI
5.6
3.9
5.3
1.5
5.8
4.4
2.6
0.44
CIN
6.0
4.1
2.5
1.1
5.7
3.8
3.4
0.45
CLE
4.2
4.2
4.1
0.8
7.2
4.5
1.4
0.34
DAL
7.3
4.9
1.5
0.9
6.1
4.0
2.0
0.45
DEN
6.2
4.2
2.0
0.8
5.1
4.1
3.1
0.37
DET
4.9
4.0
5.4
0.9
7.5
4.5
1.9
0.40
GB
6.8
4.4
1.4
0.4
5.7
3.6
4.9
0.56
HOU
7.5
3.3
2.9
1.3
6.1
4.4
2.6
0.45
IND
7.6
3.7
3.0
0.4
5.5
4.1
3.0
0.28
JAC
6.0
4.5
1.5
0.8
7.0
3.9
3.0
0.30
KC
4.9
4.0
2.7
1.9
7.0
4.8
2.6
0.30
MIA
5.5
4.4
3.3
1.3
7.0
4.1
3.0
0.32
MIN
6.9
4.1
1.5
0.9
6.0
4.0
1.9
0.38
NE
7.2
4.0
2.3
0.7
6.1
4.4
3.3
0.38
NO
8.1
4.5
2.3
0.8
6.1
4.4
4.6
0.40
NYG
7.3
4.2
2.3
1.4
6.1
4.0
2.9
0.39
NYJ
5.8
4.5
5.9
1.2
4.7
3.8
3.3
0.35
OAK
4.5
4.2
3.5
1.3
7.1
4.5
2.0
0.45
PHI
7.0
4.3
2.5
0.6
5.5
3.9
4.6
0.47
PIT
7.2
4.2
2.5
1.0
5.8
3.8
1.6
0.41
SD
8.0
3.3
2.0
0.1
5.9
4.4
2.5
0.30
SF
5.2
4.5
3.0
0.6
6.3
3.7
2.9
0.40
SEA
5.5
3.9
2.6
1.5
6.5
4.2
2.3
0.39
STL
4.9
4.4
3.7
1.0
7.3
4.5
1.9
0.43
TB
5.2
4.1
5.4
1.8
6.5
4.7
4.1
0.35
TEN
6.3
5.3
2.8
1.7
6.5
4.2
3.5
0.38
WAS
6.2
4.0
3.3
0.7
5.8
4.1
2.1
0.37
Avg
6.2
4.2
3.1
1.0
6.2
4.2
3.0
0.39

  • 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

19 Responses to “Team Efficiency Rankings - Week 16”

  1. James says:

    If anyone had any question about the best division in the NFL, I think this says it all.

  2. Anonymous says:

    Brian,
    Any chance you can write an article similar to your:

    Dec 23, 2007
    Week 15 QB Ratings Adjusted for Pass Defense


    - Brandon in New Jersey

  3. Anonymous says:

    Is there an post explaining how these efficiency rankings are calculated. As a Chargers fan, it's interesting that you rate them as an elite team whereas football outsiders rates them as an average team.

  4. ben says:

    Here is the team efficiency method:
    http://www.advancednflstats.com/2009/01/how-model-works-detailed-example.html

    I believe that he figured out how each factor correlates to wins and found that passing efficiency (offensive and defensive) is the most important factor. Since offensive passing efficiency is the Charger's best aspect they rank much higher than on football outsiders where each element is considered to be equal.

  5. James says:

    Brian, could you take a look at Nick Folk's FG kicking this year, since he is now unemployed? Knowing that % is based on the distance, how likely was it that he would only hit 64% of his attempts this year?

  6. James says:

    I ask because I wonder if this could just be a severe regression to the mean (87% over his first two season to 79% career) or a result of his off season hip surgery.

  7. Brian Burke says:

    Great question. I think the biggest differences come from what I intentionally ignore/downplay. I don't use special teams and defensive fumbles or recoveries (forced or otherwise). I also under-weight defensive interceptions.

    I under-weight defensive turnovers because these things are functions of the offense far more than abilities of the defense. Defensive interceptions have everything to do with 1) who's throwing, and 2)randomness. The defense (with some exceptions) is "just there."

    The FO models are primarily explanatory, while the one I use is designed to be predictive, which focuses primarily on "repeatable" skill. Teams with good stats in non-repeatable, context-dependent situations, like special teams or defensive turnovers, will be over-ranked in FO's models. Those with bad stats will be under-ranked.

    SD has below average defensive turnover rates, which explains past performance, but doesn't predict future performance. That's probably why they're ranked lower at FO.

    I bet that FO's rankings would better estimate past W-L records, and my rankings would better estimate future W-L records. Which one is of better? That's up to you and what you're looking for from the model.

    Bill Parcels will go down in history for saying "you are what your record says you are." But I think he was a little off. In my mind, you were what your record says you were. You are what your next game says you are.

  8. Brian Burke says:

    64% in a partial year is not statistically that different from a long-term NFL average 83%. It also depends on his attempt distances.

    His head coach is on the hot-seat. Step 1 in the Coaches Manual in that situation is to "Fire the kicker."

  9. James says:

    I've been debating this with a friend so I looked it up myself. Before this year he was 17 of 18 in the 40-49 yard range, or 94%. This year he is 5 of 12 in that range, making his career average 22 of 30 or 73%. According to your site, the NFL average for that distance is about 70%. Everything else seems consistent. Looks like he is just crashing back to the mean.

  10. Anonymous says:

    How are the Redskins (4-10) ahead of the Bengals (9-5)?

  11. Vince says:

    5/12 is significantly different from 17/18 (p = .003 by a Fisher exact test). Both rates are significant different from 73% (p = .022 and .027, respectively). His overall accuracy this year (18/28) is also significantly different from his overall accuracy during his first two years (46/53), p = .024.

  12. Becephalus says:

    -Anonymous idiot
    "How are the Redskins (4-10) ahead of the Bengals (9-5)?"

    Because not only are wins and losses not the main thing going into these rankings, they do not actually appear in them at all. The factors used to make these rankings are right there on the page for you to look at...

    Before you say "but wins and losses are what determine how good a team is!". You might also ask how my fantasy team could be ranked ahead of its opponents in all my friends' minds when it finished 7-7 and lost the first round of playoffs. If you are sophisticated and look beyond W-L at the actual statistics that matter (scoring) you would see that I ranked 2nd in scoring and was very consistent (great signs of success), but had an abnormally high pts against me scored. This is how I can have the best or second best team and finish in a three way tie for 4-6th, and how WAS can finish above CIN in these rankings.

    WAS plays in the toughest division in the league.

  13. Anonymous says:

    LOL, Brian doesn't like saying retrodictive!

    Actually, VOA is intended to be retrodictive and DVOA is predictive. Maybe someone should do a bake-off prediction contest of DVOA vs. TER.

    But I doubt either system is more predictive than the Vegas line over the long run.

  14. Brian Burke says:

    Not everyone here knows what retrodictive means.

    I thought the only difference between VOA and DVOA was an opponent adjustment. Someone can straighten me out on that if I'm off base.

    Presuming a typical week, this will be the 4th straight year (out of 4) my system has beaten the Vegas line in predicting straight-up winners. I count only weeks 4-16. I realize that's not long-term enough for some people, but it's promising.

  15. Anonymous says:

    DVOA is not predictive. It is only adjusted for opponent.

    All of FO's work is retrodictive even thought they like to think they have a crystal ball.

  16. Drew says:

    Now....if you could beat the Vegas spreads you'd really be on to something :)

  17. DAve says:

    I was pretty suprised by seeing the Redskins in front of the Bengals. However when you actually look at most of their rate stats its close to a dead heat.

    The main difference has been turnovers. The Redskins have committed more of them via ints and fumbles and have also generated less of them leaving them dead last in the league in turnover differential.

    But I also think sacks have a more significant effect than maybe the opass and dpass are weighting them.

    Avoiding and generating pressure is definitely a repeatable skill. Also defenses that sack or get to the qb more often are more likely to generate fumbles since fumbles happen on a high% of sacks (~20%).

    The formula could *possibly* be underrating this aspect of the Bengals and Redskins.

  18. Danny says:

    Thanks Ben and Brian.

    -The chargers fan

  19. James says:

    Vince, I see what you are saying but those sample sizes are way too small to really mean anything. That's why I was looking at his career as a whole.

Leave a Reply

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