BBR Rankings: Final Regular-Season Schedule-Adjusted Offensive and Defensive Ratings
Posted by Neil Paine on April 14, 2011
2010-11 NBA power rankings through the games played on April 13, 2011:
| Full-Season | Post-Deadline | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Rk | Prv | Team | Cnf | Div | W | L | W% | Off | Rk | Prv | Def | Rk | Prv | Ovr | Off | Def | Ovr |
| 1 | 2 | Miami Heat | E | SE | 58 | 24 | 0.707 | 4.40 | 2 | 4 | -3.05 | 6 | 6 | 7.45 | 5.42 | -0.87 | 6.29 |
| 2 | 1 | Chicago Bulls | E | C | 62 | 20 | 0.756 | 0.85 | 13 | 12 | -6.41 | 1 | 2 | 7.27 | 3.30 | -6.70 | 10.00 |
| 3 | 3 | Los Angeles Lakers | W | P | 57 | 25 | 0.695 | 3.33 | 7 | 6 | -3.24 | 5 | 5 | 6.56 | 1.03 | -6.15 | 7.18 |
| 4 | 4 | San Antonio Spurs | W | SW | 61 | 21 | 0.744 | 4.23 | 3 | 3 | -2.14 | 10 | 9 | 6.37 | 5.33 | 1.06 | 4.27 |
| 5 | 5 | Boston Celtics | E | A | 56 | 26 | 0.683 | -1.02 | 21 | 18 | -6.39 | 2 | 1 | 5.38 | -3.59 | -5.63 | 2.04 |
| 6 | 6 | Orlando Magic | E | SE | 52 | 30 | 0.634 | 0.58 | 15 | 14 | -4.79 | 3 | 4 | 5.37 | -0.51 | -6.28 | 5.77 |
| 7 | 7 | Denver Nuggets | W | NW | 50 | 32 | 0.610 | 4.68 | 1 | 1 | -0.33 | 14 | 13 | 5.01 | 3.54 | -6.51 | 10.05 |
| 8 | 8 | Dallas Mavericks | W | SW | 57 | 25 | 0.695 | 2.05 | 9 | 9 | -2.77 | 7 | 10 | 4.83 | 1.98 | -4.59 | 6.57 |
| 9 | 9 | Oklahoma City Thunder | W | NW | 55 | 27 | 0.671 | 3.86 | 5 | 7 | -0.15 | 15 | 15 | 4.01 | 4.96 | -1.83 | 6.79 |
| 10 | 10 | Memphis Grizzlies | W | SW | 46 | 36 | 0.561 | 0.30 | 16 | 16 | -2.46 | 8 | 8 | 2.76 | 2.40 | -2.61 | 5.01 |
| 11 | 11 | Houston Rockets | W | SW | 43 | 39 | 0.524 | 4.02 | 4 | 2 | 1.47 | 20 | 20 | 2.55 | 3.76 | -1.65 | 5.41 |
| 12 | 12 | Portland Trail Blazers | W | NW | 48 | 34 | 0.585 | 1.52 | 10 | 10 | -0.58 | 13 | 14 | 2.10 | 3.38 | -1.68 | 5.06 |
| 13 | 13 | New Orleans Hornets | W | SW | 46 | 36 | 0.561 | -0.91 | 18 | 19 | -2.35 | 9 | 7 | 1.44 | 0.73 | 1.77 | -1.04 |
| 14 | 14 | Philadelphia 76ers | E | A | 41 | 41 | 0.500 | -0.59 | 17 | 17 | -1.68 | 11 | 11 | 1.09 | -1.57 | -1.90 | 0.32 |
| 15 | 15 | New York Knicks | E | A | 42 | 40 | 0.512 | 3.85 | 6 | 5 | 3.35 | 27 | 26 | 0.50 | 6.04 | 5.66 | 0.38 |
| Rk | Prv | Team | Cnf | Div | W | L | W% | Off | Rk | Prv | Def | Rk | Prv | Ovr | Off | Def | Ovr |
| 16 | 16 | Phoenix Suns | W | P | 40 | 42 | 0.488 | 2.36 | 8 | 8 | 2.80 | 22 | 21 | -0.44 | 1.19 | 1.89 | -0.69 |
| 17 | 18 | Milwaukee Bucks | E | C | 35 | 47 | 0.427 | -5.78 | 30 | 30 | -4.69 | 4 | 3 | -1.09 | -6.06 | -4.87 | -1.18 |
| 18 | 17 | Atlanta Hawks | E | SE | 44 | 38 | 0.537 | -0.97 | 20 | 20 | 0.26 | 16 | 16 | -1.22 | -2.30 | 0.89 | -3.19 |
| 19 | 19 | Indiana Pacers | E | C | 37 | 45 | 0.451 | -2.59 | 23 | 23 | -1.09 | 12 | 12 | -1.50 | -4.62 | -0.29 | -4.32 |
| 20 | 20 | Utah Jazz | W | NW | 39 | 43 | 0.476 | 1.05 | 11 | 13 | 2.62 | 21 | 24 | -1.57 | 0.70 | 4.50 | -3.80 |
| 21 | 21 | Golden State Warriors | W | P | 36 | 46 | 0.439 | 1.00 | 12 | 11 | 3.07 | 25 | 25 | -2.08 | 1.88 | 2.04 | -0.17 |
| 22 | 22 | Los Angeles Clippers | W | P | 32 | 50 | 0.390 | -1.83 | 22 | 22 | 1.12 | 19 | 19 | -2.95 | -2.85 | -1.39 | -1.46 |
| 23 | 23 | Detroit Pistons | E | C | 30 | 52 | 0.366 | 0.70 | 14 | 15 | 4.97 | 29 | 29 | -4.27 | 3.50 | 8.00 | -4.50 |
| 24 | 24 | Charlotte Bobcats | E | SE | 34 | 48 | 0.415 | -3.69 | 25 | 27 | 0.89 | 17 | 17 | -4.58 | -4.15 | 3.29 | -7.44 |
| 25 | 25 | Sacramento Kings | W | P | 24 | 58 | 0.293 | -3.90 | 26 | 25 | 1.09 | 18 | 18 | -4.99 | -2.39 | 1.12 | -3.51 |
| 26 | 26 | Minnesota T-Wolves | W | NW | 17 | 65 | 0.207 | -3.11 | 24 | 24 | 3.04 | 24 | 23 | -6.16 | -4.07 | 3.16 | -7.23 |
| 27 | 28 | Toronto Raptors | E | A | 22 | 60 | 0.268 | -0.93 | 19 | 21 | 5.90 | 30 | 30 | -6.83 | -1.33 | 7.87 | -9.21 |
| 28 | 27 | New Jersey Nets | E | A | 24 | 58 | 0.293 | -4.00 | 27 | 26 | 2.90 | 23 | 22 | -6.89 | -3.52 | 4.28 | -7.79 |
| 29 | 29 | Washington Wizards | E | SE | 23 | 59 | 0.280 | -4.66 | 28 | 28 | 3.08 | 26 | 27 | -7.74 | -4.87 | 4.30 | -9.17 |
| 30 | 30 | Cleveland Cavaliers | E | C | 19 | 63 | 0.232 | -5.02 | 29 | 29 | 4.49 | 28 | 28 | -9.52 | -6.20 | 1.44 | -7.65 |
| HCA | 3.42 | 3.47 | |||||||||||||||
| LgRtg | 108.06 | 108.52 | |||||||||||||||
To read more about the methodology and what these numbers mean, click here.
April 14th, 2011 at 10:56 am
Final SPM ratings:
http://www.basketball-reference.com/blog/?page_id=9221
April 14th, 2011 at 11:02 am
Haha, I love the post-deadline column in anticipation of the inevitable "but with the new players my team is way better/worse than the numbers say waaaaah" whining/curiosity.
Orien Greene is a monster!!
April 14th, 2011 at 12:29 pm
Orien Greene could turn the Cavs into the Heat. Orien Greene's tears would cure cancer... if he ever cried. Which he doesn't.
Michael Jordan considers Orien Greene a top-10 guard.
(We should start a #LegendofOrienGreene hashtag on Twitter.)
April 14th, 2011 at 1:09 pm
Quick question: Is the HCA for both sides? That is, home team gets +3.4 and away team gets -3.4?
April 14th, 2011 at 1:11 pm
No, it actually splits 50-50, so the home team gets +1.7 and the away team gets -1.7.
April 14th, 2011 at 1:16 pm
Thanks Neil. OK, so if I want to simulate outcomes based on these numbers I would split the HCA 50-50 and then, using the pythag formula, split it again on ORtg and DRtg? In effect, playing at home makes the home team score 0.85 more points on offense and give up 0.85 points less on defense?
April 14th, 2011 at 1:26 pm
I hate it when people say that Chicago isn't a title contender.
Also, a little off-topic: but even with all that was made about LBJ this season and his struggles out the gate, he finished the season in the top spot in both WS and WS/48. Again.
April 14th, 2011 at 1:34 pm
Well my first run thru at simulating the playoffs shows the Bulls doing quite well:
Eastern Conference Finals
Prediction: Chicago Bulls* in 7 games.
'Seed Team Diff %Win Home'
'1E Chicago Bulls* 8 51 Home'
'2E Miami Heat* 8.2 49 Away'
Western Confernece Finals
Prediction: San Antonio Spurs* in 7 games.
'Seed Team Diff %Win Home'
'1W San Antonio Spurs* 6.2 52 Home'
'2W Los Angeles Lakers* 6.7 48 Away'
NBA Finals
Prediction: Chicago Bulls* in 7 games.
'Seed Team Diff %Win Home'
'1E Chicago Bulls* 8 67 Home'
'1W San Antonio Spurs* 6.2 33 Away'
April 14th, 2011 at 1:42 pm
#6 - Right. Or alternatively, since a +3.47 HCA corresponds to 61% pyth for the home team, you could calculate each team's neutral-court pyth and for any matchup plug those pyth%'s into:
p(home W) = ((Home_pyth) * (1 - Road_pyth) * .61)/((Home_pyth) * (1 - Road_pyth) * .61 +(1 - Home_pyth) * (Road_pyth) * (1 - .61))
April 14th, 2011 at 1:56 pm
#9 - Yes, I wanted to turn the BBR numbers into ratings like ORtg (not sure the best way to do this) but the home and away pythag might work better. Although I have not really seen much of a difference in simulation outcomes using ORtg and DRtg vs. the BBR rankings rescaled. The main difference is linked to weighting the latter part of the season. When I use your post-deadline BBR values the Nuggets look really good. They barely lose to the Bulls in 7 in the Finals.
April 14th, 2011 at 2:08 pm
The way I've been doing it (and probably will do it for the Stat Geek Smackdown) is to take the BBR ranking numbers and convert them into neutral-court ORtg DRtg like this:
ORtg = off_BBRrate + lg_Rtg
DRtg = def_BBRrate + lg_Rtg
Then convert those into pyth%'s:
Pyth% = (ORtg^14) / (ORtg^14 + DRtg^14)
Then plug those ratings (and the HCA's pyth equivalent, in this case 61%) into the log5 formula I listed above. That should give you really good probabilities for each game. It's basically the methodology Ken Pomeroy uses at his NCAA site (obviously the extra steps for HCA aren't a concern in the NCAA tournament).
April 14th, 2011 at 2:23 pm
# 11 - My first take was to pace adjust the BBR values and add 100. I see the wisdom in the lg_Rtg, I somehow did not see that on the above table. And then I used:
p(A win)=(pA-pA*pB)/(pA+pB-2*pA*pB) where A denotes the home team and B the away team. p is the Pyth% based on adjusting the Off/Def values with the HCA and an exponent of 14. I think this is about what you suggested. Maybe I got tripped up on the Road_pyth, I was thinking of a split. In any event, this is what I've been using for some time, I was just curious how this would change using the BBR rankings.
April 14th, 2011 at 2:33 pm
Neil, Chris Paul is a better defender than LeBron and Wade?
I don't really see that as reasonable, DSPM has trouble rewarding the right aspects of perimeter defense in this instance. Chris Paul has throughout his career, not had great counterpart defense. Not just one season but throughout his career. Except for 2008-2009, he's always been overstated on this website. The usage-efficiency ratio Neil likes using seems to support this as well.
Kobe Bryant in 2008 actually had equal value to 2008 Chris Paul on 82games, because Paul continues to allow so much production at his counterpart position. Chris Paul in his prime (2009) is rated appropriately, aside from that he had little business leading the 2011 season in WS/48 for most of the year. He doesn't use enough of his team's possessions, and he's not a great individual defender. Very nice help defense though.
April 14th, 2011 at 2:38 pm
Anon, yeah seemed like he lead the league in most metrics again.
At the end of the year and after the Heat's losing streak, he started playing quite well off the ball. People like to say other people should have the ball in their hands at the end of games, well guess what now LeBron is playing at a supreme level without the ball as much. March was a great example of him maximizing his offense without pounding the ball as much. I'd like to see more of that, he has been more layered this season.
AHL, there are indeed certain styles that translate better into the post-season, regardless of whether or not my team is in the top 5 post-deadline. Denver's depth for example, means less in the playoffs. Top heavy teams benefit more I think of the 2002 Lakers for example.
April 14th, 2011 at 2:47 pm
There seems to be a heavy correlation with good players named "Kevin" and their SPM: clustering around 5.0 or so.
Boy, the Atlanta Hawks are even worse than I thought.
April 14th, 2011 at 2:53 pm
#13 - DSPM overrated Paul a bit on the basis of steals, but his defensive impact is still better than Wade's:
http://stats-for-the-nba.appspot.com/4-year-ranking
Again, counterpart PER is similar to Oliver DRtg in the sense that it's a team stat masquerading as an individual defensive metric. Aspects of man-to-man are being captured in counterpart PER, certainly, but it also is heavily influenced by teammates, cross-matching, and the system.
Really the truest way to measure defense is to look at the adjusted +/- numbers, because they're isolating a player's on-court impact, independent of system, teammates, opponents, cross-matching, etc. By that measure, LeBron is superior to Paul.
DSPM comes in handy for years in which we don't have on/off data and the ability to run adjusted plus/minus. Its value is in approximation, a best guess as to what, say, Michael Jordan's plus/minus would have been in 1993 if basketballvalue.com existed back then. I would definitely go with RAPM over any boxscore metric for current player defense.
April 14th, 2011 at 2:57 pm
#12 - Pace-adjustment isn't necessary for these numbers. SRS requires pace adjustments, but here everything is denominated in efficiency differential.
April 14th, 2011 at 3:19 pm
#16 Well the question I would ask, is how much cross-matching does Chris Paul do compared to other perimeter stars? I would have to imagine his counterpart data is probably more useful than other versatile defenders.
Adjusted plus-minus defense is interesting, you are correct. Can it completely parse out teammate impact though? Is Chris Paul beating Wade in other defensive metrics? http://bkref.com/tiny/ghzUz
I don't think you can point to just one number, and say this is the reason Chris Paul is a better defender than Wade. Wade beats him in other helpful metrics, and could historically the 2 and 3 have been more dominant offensive forces in the league. LeBron, Jordan, Kobe, T-Mac, etc., I'd rather have Wade defend the most dominant perimeter players.
April 14th, 2011 at 3:26 pm
^and historically the 2 and 3 have had the biggest perimeter threats.
To make a title run I want a perimeter player that can slow down those freakish swingmen.
April 14th, 2011 at 3:36 pm
"People like to say other people should have the ball in their hands at the end of games, well guess what now LeBron is playing at a supreme level without the ball as much."
He's struggled in this area recently, but he's still above the league average in his career in GW shots.
Not that I get obsessed with it anyway. A player that is 50-50 with GW shots is VASTLY less valuable than the player who is more dominant for the rest of 47 min and 59 seconds of games. you better surround ab all-time team around "Mr. Clutch" if he plays like Brian Scalabrine during other parts of the game.
Also, Paul is a pretty good defender. He's pesky and his speed is an asset on that end of the floor.
April 14th, 2011 at 4:49 pm
As for Paul not cross-matching, more than once he's guarded Kobe at the end of games and done well against him.
How good are these post trade deadline numbers? The bad teams in the East became REALLY bad after the trade deadline with many others seeing their rating drop. Doesn't that show that these post trade deadline increases were somewhat inflated by playing teams that didn't give a crap?
April 14th, 2011 at 5:06 pm
I make no claims about how well the post-deadline numbers will predict the playoffs. If I had the time I'd look into their performance in past seasons, but the playoffs start Saturday and my TrueHoop picks are due tonight, so there's just no time to research it. (FWIW, I'm picking based on the full-season ratings.)
I think Daniel Myers has researched optimal weighting for recent vs. past games, so if he's around he probably would know how much weight to give the post-deadline ratings.
April 14th, 2011 at 6:22 pm
# 22 - I'm also intrigued by how much of the season one needs. I fixed the pace bit and used the full season BBR and post-deadline BBR. The probabilities are to win the title based on 1000 simulations. I was surprised ion how little they changed.
Post Deadline
Value Count Percent
Miami Heat* 219 21.90%
Orlando Magic* 65 6.50%
Chicago Bulls* 299 29.90%
Dallas Mavericks* 45 4.50%
Los Angeles Lakers* 149 14.90%
Boston Celtics* 67 6.70%
San Antonio Spurs* 110 11.00%
Oklahoma City Thunder* 16 1.60%
Denver Nuggets* 27 2.70%
Philadelphia 76ers* 1 0.10%
Portland Trail Blazers* 1 0.10%
Memphis Grizzlies* 1 0.10%
Full Season
Value Count Percent
Miami Heat* 229 22.90%
Orlando Magic* 75 7.50%
Boston Celtics* 79 7.90%
Chicago Bulls* 262 26.20%
Dallas Mavericks* 43 4.30%
Los Angeles Lakers* 150 15.00%
San Antonio Spurs* 105 10.50%
Oklahoma City Thunder* 16 1.60%
Denver Nuggets* 33 3.30%
Memphis Grizzlies* 4 0.40%
Philadelphia 76ers* 1 0.10%
Portland Trail Blazers* 1 0.10%
New York Knickerbockers* 1 0.10%
New Orleans Hornets* 1 0.10%
April 15th, 2011 at 12:05 am
Trick-or-Treat Tony does very well under SPM; is he the one the Celts are really missing?
April 15th, 2011 at 1:41 am
Neil,
Isn't HCA significantly greater in the playoffs? I did some research at one point and discovered it was ~66-67% (controlling for quality of opponent).
April 15th, 2011 at 3:26 am
I've done a bit of research from a gambling perspective and found that regular season performance is a poor predictor of potential playoff success. THE best indicator I've found is the opening point spread in the first playoff game.
The Bulls pose a particularly big problem in predicting playoff success. They were clearly the best team in the regular season; they had the best record and the best point differential. Yet they are, at best, the third most likely team to win the championship. Even with home court advantage, they will likely go off as underdogs in the conference finals and are only 60-40 to get that far in the first place.
To put it simply, the postseason is an entirely different animal from the regular season and must be analyzed very differently.
April 15th, 2011 at 3:34 am
Heat 16-0.The King for Finals MVP.
April 15th, 2011 at 8:42 am
Orien Greene is a monster!!
>>>>>>>>>>>>>>>>>>>>>>
And SOYLENT GREEN is people!
April 15th, 2011 at 9:32 am
#26 - Never mind the fact that I still haven't seen any evidence for these claims you continually make... If true, wouldn't that make Chicago the favorites to win it all?
Biggest point spreads of opening games:
Indiana at Chicago (CHI -11½)
Philadelphia at Miami (MIA -10½)
New Orleans at L.A. Lakers (LAL -10)
Atlanta at Orlando (ORL -8)
New York at Boston (BOS -6½)
Memphis at San Antonio (SAS -6)
Portland at Dallas (DAL -5)
Denver at Oklahoma City (OKL -5)
April 15th, 2011 at 10:33 am
Neil,
I'm sadly going to give #26 the benefit of the doubt, and assume he meant the moneyline odds to win the championship on day 1 of the playoffs when he says "third most likely".
Today, it's:
Lakers +250
Heat +325
Bulls +350
But Vegas isn't only in the game to be the most accurate. They are in the game of winning the most money. If the public is stupid about betting, they will adjust the lines even if the Bulls are the favorites. This is probably what's happening with the Lakers and Heat this year. "Oh, another three-peat hurf durf" and "Miami LeBron we already won hurf durf".
The key shouldn't be relying on what Vegas gives, it should be coming up with an independent method to beat it. After all, they set their lines starting from somehere.
April 15th, 2011 at 10:37 am
Oh, also it depends which site you use, some have Miami as favorites some have Miami 3rd behind Chicago, some have equal odds for the Lakers and Bulls.
April 15th, 2011 at 3:26 pm
#21 Matt, he doesn't cross match like Wade does. Let him cover Kobe over an entire game... Right of course you didn't suggest that but that is my larger point. Paul is a nice defender, Wade is better.
April 15th, 2011 at 4:54 pm
Dsong. Based on what, other than general sweeping statements about "regular season performance being a poor indicator" do you have to back up your statement. The fact that 15 of the last 31 champs have had the best regular season record and 22 of 31 have been #1 seeds in their conference would lead me to believe that regular season records are a tad better than "poor."