Football Outsiders does this with their defensive grades.
Point taken. Still, the FO methodology is mostly a black box: [
https://www.footballoutsiders.com/stats/teamdef] That's understandable. They don't want to be copied. Still, believing the conclusions of a black box requires a leap of faith.
Listed below are teams in order of their 2017 DVOA rank, best to worst, followed by their points-against rank, with many notable discrepancies.
Teams with a DVOA rank much better than their points against rank (10 or more ranking position difference): ARI, DEN, WAS, CLE
These teams ranked 25th. or worse in offensive points scored, 16th. or worse in offensive yards per game, 23rd. or worse in offensive turnovers, 17th. or worse in number of punts.
Teams with a DVOA rank much worse than their points against rank (10 or more ranking position difference): ATL, DAL, KC, NE, TB
These teams ranked 18th. or better in offensive points scored, 14th. or better in offensive yards per game, 4 of 5 were 14th. or better in offesive turnovers (TB 26th.), and 8th. or better in number of punts.
Conventional wisdom would say that a team with a good offense helps a defense keep points off the board while a bad one compromises the defense. If the offense moves the ball and scores points while limiting turnovers and punts, that helps the defense. A poor offense can make its defense look worse than it is in a points against ranking.
To FO's credit, their DVOA appears to capture this reality even if it doesn't incorporate an offensive production factor into defensive performance. It just works out that way.
1 JAC, 2 in points against
2 MIN, 1
3 BAL, 6
4 ARI, 19
5 PHI, 4
6 LARM, 12
7 CAR, 11
8 NO, 10
9 PIT, 7
10 DEN, T22
11 WAS, T27
12 LACH, 3
13 SEA, T13
14 CHI, 9
15 BUF, 18
16 CLE, 31
17 CIN, 16
18 NYJ, T22
19 DET, 21
20 GB, 26
21 TEN, 17
22 ATL, 8
23 HOU,32
24 NYG, T27
25 DAL, T13
26 SF, 25
27 IND, 30
28 MIA, 29
29 OAK, 20
30 KC, 15
31 NE, 5
32 TB, T22
