FanPost

OAPD: Week 6 Update (With Point Ratio)

Otto Greule Jr

First I'd like to thank everyone for their feedback over the last week. I'm glad to see that the majority of you found OAPD to be useful (and only 3 voted for me to go away). I will update the rankings weekly and will try to incorporate suggestions as we go forward, starting this week with the addition of Opponent-Adjusted Point Ratio (suggested by Jason Drake).

For a brief refresher, OAPD stands for Opponent-Adjusted Point Differential. It is calculated by comparing a team's performance to their opponent's average points for and average points allowed. If Team A scores 30 points against Team B and Team B's average points allowed is 23, then Team A's offensive rating (PFD - Points For Differential) for that game is +7. Defensive rating (PAD - Points Allowed Differential) is similar, only it's good to have a negative result, as that indicates holding a team below their average output. These results are averaged over the number of games played to give us the final ratings.

For more on OAPD, and to see the 2013 regular season results, click here.

Results through Week 6

Week 6 sees a new team in the top spot, as the Seahawks struggled and the Ravens demolished the Bucs to move into first from last week's #4 rank. Denver stays put at number 2. The Seahawks, despite looking inept the majority of the game, only fell two spots and come in at #3 this week.

The relatively minor fall of the Seahawks brings up a weakness of this system - all points scored, regardless if they're Offensive/Defensive/ST are factored into PFD, which I then refer to as "offensive rating" in these articles. I probably shouldn't, but I will undoubtedly continue to do so. The same goes for PAD being referred to as "defensive rating" whether the points allowed are against the defense or result from a pick 6 or a kickoff return TD.

With the above limitations, the Seahawks PFD (offensive rating) was actually helped this week, scoring 23 against a Dallas team that only gives up 21 points on average. This result is hidden by their overall PFD going down, mostly due to the Chargers giving up 28 points to the Raiders which increased their PA/G and therefore decreased the Seahawks week 2 margin over that average. The cause and effect in week to week changes can be difficult to track since it is altered by every previous opponent's games. For that reason I generally won't attempt to analyze the causes of week to week changes.

Enough rambling, here are the numbers!

Team W L PFD PFD Rank PAD PAD Rank OAPD Last Week Change
1 Baltimore Ravens 4 2 2.07 11 -8.43 1 10.50 4 3
2 Denver Broncos 4 1 6.80 2 -3.09 8 9.89 2 -
3 Seattle Seahawks 3 2 5.34 3 -4.01 4 9.35 1 (2)
4 Indianapolis Colts 4 2 7.56 1 -0.68 15 8.24 6 2
5 San Diego Chargers 5 1 2.54 10 -3.96 6 6.50 3 (2)
6 Philadelphia Eagles 5 1 4.72 4 -1.53 13 6.26 16 10
7 Arizona Cardinals 4 1 1.77 12 -3.68 7 5.45 5 (2)
8 Kansas City Chiefs 2 3 1.18 16 -3.98 5 5.16 11 3
9 Cincinnati Bengals 3.5 1.5 3.27 8 -1.83 11 5.10 7 (2)
10 San Francisco 49ers 4 2 0.71 17 -4.34 3 5.05 13 3
11 Dallas Cowboys 5 1 3.03 9 -1.67 12 4.71 14 3
12 Green Bay Packers 4 2 4.29 5 0.48 19 3.81 10 (2)
13 New England Patriots 4 2 3.46 7 0.27 18 3.19 17 4
14 Cleveland Browns 3 2 3.56 6 0.52 20 3.04 19 5
15 Detroit Lions 4 2 -4.33 25 -7.25 2 2.92 12 (3)
16 Buffalo Bills 3 3 -0.19 21 -2.86 9 2.67 9 (7)
17 New York Giants 3 3 0.30 19 -1.06 14 1.36 8 (9)
18 Houston Texans 3 3 -1.69 23 -2.83 10 1.14 15 (3)
19 Miami Dolphins 2 3 1.77 13 2.25 23 -0.48 20 1
20 Chicago Bears 3 3 -0.17 20 1.19 22 -1.36 23 3
21 Carolina Panthers 3.5 2.5 1.23 15 3.17 25 -1.94 22 1
22 New Orleans Saints 2 3 0.37 18 4.41 27 -4.04 21 (1)
23 Atlanta Falcons 2 4 1.39 14 5.58 29 -4.18 18 (5)
24 Pittsburgh Steelers 3 3 -4.86 27 0.18 17 -5.04 24 -
25 Washington Redskins 1 5 -0.99 22 4.67 28 -5.67 25 -
26 Minnesota Vikings 2 4 -6.56 31 -0.63 16 -5.93 26 -
27 Tennessee Titans 2 4 -6.05 30 0.79 21 -6.84 27 -
28 New York Jets 1 5 -4.35 26 2.58 24 -6.93 28 -
29 Oakland Raiders 0 5 -5.76 29 3.60 26 -9.36 31 2
30 St. Louis Rams 1 4 -3.73 24 6.23 31 -9.97 29 (1)
31 Tampa Bay Buccaneers 1 5 -5.34 28 9.76 32 -15.10 30 (1)
32 Jacksonville Jaguars 0 6 -8.92 32 6.22 30 -15.14 32 -

Notable Rankings

  • The Eagles ranking (6) now aligns more with their record, thanks to their blowout victory over the giants. They are also the biggest riser, jumping 10 spots
  • The Giants had the largest fall, dropping 9 positions to #17
  • The Cowboys are still only ranked #11 despite beating the Seahawks. They had minor improvements in both PFD and PAD, but only rose 3 spots.
  • After their bye, the Chiefs remain the highest ranking team with a losing record at #8
  • The Panthers are the lowest ranked team with a winning record, at #21
  • Speaking of the Panthers.. Shoutout to them and the Bengals (#9) for making me alter formulas across several sheets to account for ties
  • The Dolphins (#19) now have the most neutral OAPD at -0.48
  • Jacksonville barely holds onto the #32 spot, being challenged by Tampa Bay
  • The Bucs were blown out again and have a significantly worse PAD (9.76) than the next closest (Rams at 6.23). One more point allowed or one fewer scored and they would have taken #32 overall.

Opponent-Adjusted Point Ratio

In last week's comment section Jason Drake suggested evaluating the point ratio instead of just the differential. I'll let him explain the concept:

I’d also be curious to see an opponent-adjusted points ratio. The rationale being that a 20-10 victory is just as dominant as a 40-20 victory (either one doubles the opposition score). The Broncos, by way of interest, lead the league in points differential last year (606 – 399 = 207) ahead of Seattle (417 – 231 = 186). But the Broncos only scored 152% of their opposition’s points (606/399) whereas the Seahawks scored 181% of their opposition’s points (417/231).

Naturally, one would use a logarithm of the differential for comparison, so that doubling an opponent (ln 2 = 0.693) is comparable to being double by one’s opponent (ln 1/2 = -0.693).

Nice pun, Jason.

To adjust for opponents, when looking at Team A, we use Team A's PF/G and the average of Team A's opponent's PA/G. The natural log of this becomes PFR, which we will use to rate the points scored of Team A. PAR is used to rate the points allowed of Team A and is calculated similarly. These two results are added to gauge the overall quality of a team (OAPR). I have also included comparisons between the two different systems, OAPD and OAPR in the table below.

Team W L OAPD vs OAPR PFR PFR Rank PFD vs PFR PAR PAR Rank PAD vs PAR OAPR
1 Baltimore Ravens 4 2 - 0.079 12 (1) -0.420 2 (1) 0.499
2 Denver Broncos 4 1 - 0.263 2 - -0.138 8 - 0.401
3 Seattle Seahawks 3 2 - 0.224 3 - -0.163 6 (2) 0.387
4 San Diego Chargers 5 1 1 0.097 10 - -0.232 3 3 0.329
5 Indianapolis Colts 4 2 (1) 0.292 1 - -0.029 15 - 0.321
6 Philadelphia Eagles 5 1 - 0.168 5 (1) -0.073 13 - 0.241
7 Arizona Cardinals 4 1 - 0.080 11 1 -0.160 7 - 0.240
8 Kansas City Chiefs 2 3 - 0.051 16 - -0.180 5 - 0.231
9 Detroit Lions 4 2 6 -0.202 25 - -0.426 1 1 0.224
10 San Francisco 49ers 4 2 - 0.031 17 - -0.192 4 (1) 0.223
11 Cincinnati Bengals 3.5 1.5 (2) 0.130 8 - -0.078 11 - 0.208
12 Dallas Cowboys 5 1 (1) 0.117 9 - -0.077 12 - 0.194
13 Green Bay Packers 4 2 (1) 0.174 4 1 0.023 19 - 0.151
14 New England Patriots 4 2 (1) 0.139 7 - 0.012 18 - 0.127
15 Cleveland Browns 3 2 (1) 0.143 6 - 0.023 20 - 0.120
16 Buffalo Bills 3 3 - -0.010 21 - -0.128 10 (1) 0.118
17 New York Giants 3 3 - 0.014 18 1 -0.045 14 - 0.059
18 Houston Texans 3 3 - -0.074 23 - -0.132 9 1 0.058
19 Miami Dolphins 2 3 - 0.076 13 - 0.095 23 - -0.019
20 Chicago Bears 3 3 - -0.007 20 - 0.051 22 - -0.058
21 Carolina Panthers 3.5 2.5 - 0.054 14 1 0.129 25 - -0.075
22 New Orleans Saints 2 3 - 0.014 19 (1) 0.170 27 - -0.156
23 Atlanta Falcons 2 4 - 0.052 15 (1) 0.219 29 - -0.167
24 Pittsburgh Steelers 3 3 - -0.211 26 1 0.008 17 - -0.219
25 Washington Redskins 1 5 - -0.044 22 - 0.185 28 - -0.229
26 Minnesota Vikings 2 4 - -0.321 31 - -0.026 16 - -0.295
27 Tennessee Titans 2 4 - -0.299 29 1 0.032 21 - -0.331
28 New York Jets 1 5 - -0.240 28 (2) 0.103 24 - -0.343
29 St. Louis Rams 1 4 1 -0.170 24 - 0.233 31 - -0.403
30 Oakland Raiders 0 5 (1) -0.311 30 (1) 0.144 26 - -0.455
31 Tampa Bay Buccaneers 1 5 - -0.237 27 1 0.338 32 - -0.575
32 Jacksonville Jaguars 0 6 - -0.507 32 - 0.225 30 - -0.732

Wrapping It Up

Overall the differences between the two systems are minor, with PFD/PFR differing on 13 teams and with a max change of 2. PAD/PAR differ on only 7 teams, but with a max change of 3. The most interesting result is in the overall rankings, with Detroit rising 6 spots despite staying in the same position in PF and only rising 1 position in PA. Although the 31 (Bucs) and 32 (Jags) spots didn't change, the gap is significantly wider in OAPR than it is in OAPD.

It is harder to glean meaning from the OAPR ratings, since the numbers don't directly correspond to the score, whereas the OAPD results represent points. That being said, the adjustments OAPR made to the overall ratings seem like, in my expert helplessly amateur opinion, an improvement over OAPD. I'm inclined to provide both sets of data in the future, but let me know what you think in the comments!