BBR Rankings: 2011 Offense and Defense Ratings, Adjusted for Schedule (Week 2 Edition)
Posted by Neil Paine on November 5, 2010
Not sure if I want to make this the new BBR Rankings method or return to maximum likelihood, but these are your team schedule-adjusted offensive & defensive ratings so far in 2011 (negative = good for defenses):
| Rank | Team | Offense | Rk | Defense | Rk | Overall |
|---|---|---|---|---|---|---|
| 1 | MIA | 8.15 | 1 | -10.56 | 2 | 18.72 |
| 2 | ORL | 3.15 | 11 | -12.25 | 1 | 15.40 |
| 3 | DEN | 5.47 | 7 | -6.69 | 4 | 12.16 |
| 4 | LAL | 5.84 | 4 | -6.11 | 6 | 11.95 |
| 5 | NOH | 3.82 | 10 | -6.94 | 3 | 10.75 |
| 6 | DAL | 0.49 | 14 | -6.14 | 5 | 6.64 |
| 7 | ATL | 7.32 | 2 | 1.28 | 18 | 6.04 |
| 8 | BOS | 1.84 | 13 | -2.79 | 9 | 4.63 |
| 9 | SAS | 2.59 | 12 | -1.10 | 11 | 3.69 |
| 10 | POR | 4.57 | 9 | 1.32 | 19 | 3.25 |
| 11 | NYK | 0.29 | 16 | -2.39 | 10 | 2.68 |
| 12 | PHO | 5.73 | 6 | 3.14 | 21 | 2.59 |
| 13 | HOU | 6.23 | 3 | 4.31 | 25 | 1.92 |
| 14 | PHI | -3.70 | 24 | -5.56 | 7 | 1.86 |
| 15 | UTA | 0.45 | 15 | 0.13 | 16 | 0.32 |
| 16 | GSW | 4.87 | 8 | 5.11 | 26 | -0.25 |
| 17 | MEM | -1.58 | 20 | -0.72 | 13 | -0.86 |
| 18 | WAS | -1.18 | 19 | 0.72 | 17 | -1.89 |
| 19 | MIL | -7.18 | 26 | -3.28 | 8 | -3.90 |
| 20 | IND | -2.29 | 23 | 1.73 | 20 | -4.02 |
| 21 | TOR | -1.10 | 18 | 3.29 | 22 | -4.40 |
| 22 | CHI | -1.79 | 22 | 3.45 | 23 | -5.25 |
| 23 | SAC | 5.81 | 5 | 12.14 | 30 | -6.33 |
| 24 | CHA | -1.01 | 17 | 7.21 | 29 | -8.22 |
| 25 | OKC | -1.78 | 21 | 7.08 | 28 | -8.86 |
| 26 | CLE | -10.48 | 29 | -1.01 | 12 | -9.46 |
| 27 | LAC | -10.42 | 28 | -0.65 | 14 | -9.77 |
| 28 | NJN | -11.34 | 30 | -0.41 | 15 | -10.93 |
| 29 | MIN | -7.41 | 27 | 3.83 | 24 | -11.24 |
| 30 | DET | -5.39 | 25 | 7.04 | 27 | -12.43 |
| hca | 4.15 | |||||
| lgavg | 106.35 |
Here's how it works... For every game, you can set up a prediction for the home team's offensive and defensive ratings using the following formulae:
Home ORtg = Lg Avg ORtg + .5*HCA + Home Offense Score + Away Defense Score
Home DRtg = Lg Avg ORtg - .5*HCA + Away Offense Score + Home Defense Score
The ratings above represent the solution that produced the lowest squared errors for the predictions.
November 5th, 2010 at 1:05 pm
I like this, but can you switch to efficiency differentials for each game, rather than the score? Efficiency differential would be more accurate.
November 5th, 2010 at 1:22 pm
This is based on efficiency differential... Sorry, should have made that clearer. I added links to ORtg/DRtg which hoprefully explain it better.
November 5th, 2010 at 1:28 pm
I like this system better. Thanks Neil. Will you update it regularly?
November 5th, 2010 at 1:45 pm
Hornets with the defense! Of course, at such an early point in the season, quality of opponent is a larger factor than usual. Orlando's offensive rating must've taken a huge hit from the Miami game.
November 5th, 2010 at 1:46 pm
Re: #3 - Yeah, I'm leaning towards it too. It would be every Friday, just like the rankings last year.
November 5th, 2010 at 1:49 pm
Neil: can we get the SRS on this site changed to based on efficiency differential, rather than scores? That would be a wonderful improvement!
November 5th, 2010 at 1:50 pm
Unless it is already...
November 5th, 2010 at 2:16 pm
Trouble is, if we did that, SRS would only go back to 1987 (since it would require the game-by-game boxscore data to calculate possessions). We use raw scoring margin for SRS because, while admittedly less accurate (although not terribly so), it allows us to calculate SRS for all of pro basketball history. I'd rather SRS be slightly inaccurate for 60+ years than only have 23 years of ratings to work with.
November 5th, 2010 at 2:38 pm
I see.
A request for research: could you calculate the effect of rest days on team efficiency differential? Along the lines of NBA Stuffer's work: http://www.nbastuffer.com/2009-2010_NBA_Regular_Season_Rest_Days_Stats.html . It would take a little effort, but nevertheless it could be very interesting.
November 5th, 2010 at 6:53 pm
I like this. Hope you can continue.
November 6th, 2010 at 12:27 pm
This is great.
November 6th, 2010 at 1:35 pm
Cool, so it looks like basically everyone wants this to become the new BBR Rankings method of choice.
November 9th, 2010 at 12:29 am
This is a great site. Great methodology. Too bad the results place the the best team so far -- the Lakers -- behind.... DENVER? Better tweak the gears in a hurry, or this site will make like the Rockets and be gone. I invite responses at Obviousman@AOL.com.