DPOY Decision ’10: The Halfway Mark
Posted by Neil Paine on February 1, 2010
A little over halfway into the 2010 season, I thought now would be a good time to take a look at the leading candidates for the Defensive Player of the Year Award this year, using the various metrics we like to look at when judging a player's impact on defense. First up, defensive rating:
Player | Team | G | MP | DRtg |
---|---|---|---|---|
Dwight Howard | ORL | 48 | 1687 | 95.1 |
Gerald Wallace | CHA | 45 | 1894 | 98.1 |
Rasheed Wallace | BOS | 42 | 1005 | 98.3 |
Lamar Odom | LAL | 48 | 1490 | 98.6 |
Kendrick Perkins | BOS | 45 | 1306 | 98.8 |
Marcus Camby | LAC | 46 | 1433 | 99.2 |
Kevin Garnett | BOS | 34 | 1049 | 99.5 |
Joakim Noah | CHI | 44 | 1450 | 99.6 |
Andrew Bogut | MIL | 39 | 1282 | 99.9 |
Anderson Varejao | CLE | 47 | 1404 | 100.1 |
Josh Smith | ATL | 46 | 1550 | 100.4 |
Pau Gasol | LAL | 31 | 1132 | 100.5 |
LeBron James | CLE | 49 | 1889 | 100.6 |
Carlos Boozer | UTA | 45 | 1576 | 100.7 |
Shaquille O'Neal | CLE | 43 | 998 | 100.8 |
Andrew Bynum | LAL | 46 | 1467 | 100.8 |
Rajon Rondo | BOS | 44 | 1622 | 100.9 |
Tim Duncan | SAS | 43 | 1402 | 101.2 |
Thabo Sefolosha | OKC | 47 | 1406 | 101.4 |
Chris Andersen | DEN | 45 | 1009 | 101.6 |
(Note: The minimum requirement for all of the rate stats is 960 minutes.)
Defending DPOY Dwight Howard dominates this list, posting a defensive rating 3 full pts/100 poss. better than #2, Gerald Wallace. The Magic have the 5th-best defense in the league in 2010 (last year, they were #1), and Howard shows up very strong in 2 of the 3 individual categories that make up DRtg:
Player | Team | G | MP | DRb% |
---|---|---|---|---|
Marcus Camby | LAC | 46 | 1433 | 33.2 |
Dwight Howard | ORL | 48 | 1687 | 30.8 |
Samuel Dalembert | PHI | 47 | 1220 | 30.2 |
Carlos Boozer | UTA | 45 | 1576 | 29.3 |
Troy Murphy | IND | 38 | 1180 | 28.0 |
David Lee | NYK | 47 | 1736 | 27.9 |
Joakim Noah | CHI | 44 | 1450 | 27.3 |
Andrew Bogut | MIL | 39 | 1282 | 26.6 |
Lamar Odom | LAL | 48 | 1490 | 26.5 |
Tim Duncan | SAS | 43 | 1402 | 26.4 |
Chris Bosh | TOR | 48 | 1725 | 26.4 |
Gerald Wallace | CHA | 45 | 1894 | 25.7 |
Luis Scola | HOU | 47 | 1451 | 25.7 |
Kendrick Perkins | BOS | 45 | 1306 | 25.5 |
Al Jefferson | MIN | 47 | 1580 | 24.8 |
Shaquille O'Neal | CLE | 43 | 998 | 24.5 |
Emeka Okafor | NOH | 47 | 1431 | 24.4 |
Al Horford | ATL | 46 | 1579 | 23.8 |
Kevin Garnett | BOS | 34 | 1049 | 23.4 |
Ben Wallace | DET | 46 | 1383 | 22.9 |
Player | Team | G | MP | Blk% |
---|---|---|---|---|
Samuel Dalembert | PHI | 47 | 1220 | 6.6 |
Chris Andersen | DEN | 45 | 1009 | 6.1 |
Dwight Howard | ORL | 48 | 1687 | 5.7 |
Kendrick Perkins | BOS | 45 | 1306 | 5.3 |
Andrew Bogut | MIL | 39 | 1282 | 5.2 |
Roy Hibbert | IND | 48 | 1137 | 5.2 |
Josh Smith | ATL | 46 | 1550 | 5.0 |
Brendan Haywood | WAS | 45 | 1501 | 4.7 |
Marcus Camby | LAC | 46 | 1433 | 4.6 |
Emeka Okafor | NOH | 47 | 1431 | 4.3 |
Brook Lopez | NJN | 46 | 1676 | 4.3 |
Joakim Noah | CHI | 44 | 1450 | 4.0 |
Tim Duncan | SAS | 43 | 1402 | 3.9 |
Andrew Bynum | LAL | 46 | 1467 | 3.7 |
Jermaine O'Neal | MIA | 43 | 1216 | 3.7 |
Shaquille O'Neal | CLE | 43 | 998 | 3.6 |
Marc Gasol | MEM | 46 | 1639 | 3.3 |
Andrei Kirilenko | UTA | 40 | 1159 | 3.3 |
Paul Millsap | UTA | 46 | 1225 | 3.2 |
Andray Blatche | WAS | 45 | 992 | 3.2 |
The only category in which Howard isn't a stud is Steal%, where Rajon Rondo has outpaced last year's leader, Chris Paul:
Player | Team | G | MP | Stl% |
---|---|---|---|---|
Rajon Rondo | BOS | 44 | 1622 | 3.5 |
Chris Paul | NOH | 38 | 1472 | 3.1 |
BaRon Davis | LAC | 47 | 1608 | 2.8 |
Manu Ginobili | SAS | 41 | 1099 | 2.8 |
Ronnie Brewer | UTA | 46 | 1490 | 2.8 |
Raymond Felton | CHA | 46 | 1508 | 2.8 |
Mario Chalmers | MIA | 46 | 1135 | 2.8 |
Rasheed Wallace | BOS | 42 | 1005 | 2.7 |
Stephen Curry | GSW | 46 | 1583 | 2.6 |
Dwyane Wade | MIA | 47 | 1710 | 2.5 |
Courtney Lee | NJN | 38 | 1232 | 2.5 |
Monta Ellis | GSW | 44 | 1841 | 2.5 |
Andre Iguodala | PHI | 47 | 1883 | 2.5 |
Jason Kidd | DAL | 46 | 1660 | 2.4 |
Nene Hilario | DEN | 47 | 1603 | 2.4 |
Stephen Jackson | TOT | 46 | 1793 | 2.4 |
Trevor Ariza | HOU | 46 | 1761 | 2.3 |
Josh Smith | ATL | 46 | 1550 | 2.3 |
James Harden | OKC | 47 | 1073 | 2.3 |
Louis Williams | PHI | 35 | 1095 | 2.3 |
All of these stats come together to help us calculate Defensive Stops, or the estimated number of opponent individual possessions directly stopped by a player through steals, blocks, forced misses, defensive rebounds, and forced turnovers:
Player | Team | G | MP | Stops |
---|---|---|---|---|
Dwight Howard | ORL | 48 | 1687 | 462.1 |
Gerald Wallace | CHA | 45 | 1894 | 442.6 |
LeBron James | CLE | 49 | 1889 | 409.4 |
Kevin Durant | OKC | 47 | 1874 | 388.0 |
David Lee | NYK | 47 | 1736 | 382.6 |
Marcus Camby | LAC | 46 | 1433 | 382.5 |
Andre Iguodala | PHI | 47 | 1883 | 379.2 |
Josh Smith | ATL | 46 | 1550 | 372.6 |
Kobe Bryant | LAL | 48 | 1852 | 371.6 |
Chris Bosh | TOR | 48 | 1725 | 370.0 |
Dirk Nowitzki | DAL | 46 | 1752 | 365.1 |
Nene Hilario | DEN | 47 | 1603 | 365.0 |
Carlos Boozer | UTA | 45 | 1576 | 361.7 |
Monta Ellis | GSW | 44 | 1841 | 357.0 |
Marc Gasol | MEM | 46 | 1639 | 356.3 |
Trevor Ariza | HOU | 46 | 1761 | 355.4 |
Stephen Jackson | TOT | 46 | 1793 | 353.0 |
Al Jefferson | MIN | 47 | 1580 | 351.9 |
Joakim Noah | CHI | 44 | 1450 | 349.7 |
Lamar Odom | LAL | 48 | 1490 | 349.3 |
Here at Basketball-Reference, we like to synthesize all of this information together into one defensive metric called Defensive Win Shares, which takes a player's DRtg and calculates how much impact his individual defense had on his team's defensive rating:
Player | Team | G | MP | DWS |
---|---|---|---|---|
Dwight Howard | ORL | 48 | 1687 | 4.2 |
Gerald Wallace | CHA | 45 | 1894 | 3.9 |
LeBron James | CLE | 49 | 1889 | 3.4 |
Lamar Odom | LAL | 48 | 1490 | 3.0 |
Kevin Durant | OKC | 47 | 1874 | 3.0 |
Kobe Bryant | LAL | 48 | 1852 | 2.9 |
Rajon Rondo | BOS | 44 | 1622 | 2.8 |
Marcus Camby | LAC | 46 | 1433 | 2.8 |
Josh Smith | ATL | 46 | 1550 | 2.8 |
Carlos Boozer | UTA | 45 | 1576 | 2.8 |
Joakim Noah | CHI | 44 | 1450 | 2.8 |
Kendrick Perkins | BOS | 45 | 1306 | 2.6 |
Anderson Varejao | CLE | 47 | 1404 | 2.6 |
Andrew Bynum | LAL | 46 | 1467 | 2.6 |
Stephen Jackson | TOT | 46 | 1793 | 2.5 |
Raymond Felton | CHA | 46 | 1508 | 2.4 |
Dirk Nowitzki | DAL | 46 | 1752 | 2.4 |
Jeff Green | OKC | 47 | 1723 | 2.4 |
Tim Duncan | SAS | 43 | 1402 | 2.4 |
Andrew Bogut | MIL | 39 | 1282 | 2.4 |
As you can see, there's scarcely a boxscore defensive metric in which Howard doesn't rank 1st... But what about the plus/minus data? Which player is having the biggest on/off-court impact so far this season?
Player | Team | Min | On | Off | Net |
---|---|---|---|---|---|
Varejao, Anderson | CLE | 1404.4 | 99.28 | 108.71 | -9.43 |
Kaman, Chris | LAC | 1494.5 | 104.85 | 113.08 | -8.22 |
Williams, Marvin | ATL | 1342.0 | 103.44 | 111.66 | -8.22 |
Artest, Ron | LAL | 1424.2 | 99.21 | 106.65 | -7.44 |
Garnett, Kevin | BOS | 1015.8 | 98.12 | 105.43 | -7.31 |
Bell, Charlie | MIL | 1132.0 | 101.09 | 108.23 | -7.14 |
Andersen, Chris | DEN | 1009.3 | 102.20 | 109.30 | -7.09 |
Nowitzki, Dirk | DAL | 1752.5 | 103.84 | 110.69 | -6.85 |
Kirilenko, Andrei | UTA | 1158.8 | 101.48 | 108.24 | -6.76 |
Sessions, Ramon | MIN | 1017.1 | 106.70 | 112.91 | -6.21 |
Horford, Al | ATL | 1579.6 | 104.93 | 111.02 | -6.09 |
Wallace, Ben | DET | 1383.4 | 106.76 | 112.80 | -6.04 |
Marion, Shawn | DAL | 1398.3 | 103.12 | 109.12 | -6.00 |
Dalembert, Samuel | PHI | 1200.7 | 105.85 | 111.76 | -5.90 |
Jeffries, Jared | NYK | 1283.7 | 105.86 | 111.69 | -5.84 |
Odom, Lamar | LAL | 1490.1 | 100.03 | 105.82 | -5.79 |
Bibby, Mike | ATL | 1248.8 | 104.23 | 109.87 | -5.65 |
Howard, Dwight | ORL | 1687.6 | 102.02 | 107.59 | -5.57 |
Carter, Vince | ORL | 1234.0 | 100.97 | 106.50 | -5.53 |
Bryant, Kobe | LAL | 1851.7 | 101.00 | 106.38 | -5.38 |
Howard is still among the league's elite, ranking 18th, but Anderson Varejao headlines the list; Cleveland plays like the best D in the league while he's on the floor, and they play roughly like the Knicks' defense when he's on the bench. Naturally, not all of Cleveland's great on-court D can be attributed to Sideshow Bob, but since no one else on the Cavs with more than 820 minutes has a positive defensive on/off +/-, it's safe to say that Varejao has been one of the most integral parts of their 4th-ranked D.
Finally, here's another boxscore metric that is based on adjusted plus/minus (similar to SPM)... It comes from an old Dan Rosenbaum post at APBRmetrics, in which he regressed boxscore stats on his pure adjusted offensive and defensive plus minus ratings. Here's what he found for defense:
Model: MODEL2 Dependent Variable: DEF1 Analysis of Variance Sum of Mean Source DF Squares Square F Value Prob>F Model 12 4432015.5876 369334.6323 48.579 0.0001 Error 1081 8218599.2007 7602.7744687 C Total 1093 12650614.788 Root MSE 87.19389 R-square 0.3503 Dep Mean 0.30028 Adj R-sq 0.3431 C.V. 29037.83409 Parameter Estimates Parameter Standard T for H0: Variable DF Estimate Error Parameter=0 Prob > |T| INTERCEP 1 -3.683703 0.66138060 -5.570 0.0001 PTS 1 -0.067574 0.06879300 -0.982 0.3262 TSA 1 -0.105195 0.06760132 -1.556 0.1200 FTA 1 0.179179 0.06810286 2.631 0.0086 TA 1 0.007954 0.04576327 0.174 0.8620 AS 1 0.035210 0.05237392 0.672 0.5015 OR 1 -0.126936 0.10860612 -1.169 0.2428 DR 1 0.393748 0.05988948 6.575 0.0001 TO 1 -0.382290 0.15430545 -2.477 0.0134 ST 1 1.080512 0.16107366 6.708 0.0001 BK 1 1.014717 0.11110280 9.133 0.0001 PF 1 0.309126 0.09203166 3.359 0.0008 MPG 1 0.057194 0.01251181 4.571 0.0001 Where: PTS = points per 40 minutes TSA = true shooting attempts per 40 minutes FTA = free throw attempts per 40 minutes TA = three point attempts per 40 minutes AS = assists per 40 minutes OR = offensive rebounds per 40 minutes DR = defensive rebounds per 40 minutes TO = turnovers per 40 minutes ST = steals per 40 minutes BK = blocks per 40 minutes PF = personal fouls per 40 minutes MPG = minutes per game
I used this equation and, similar to how I calculate SPM, I forced the weighted sum of a team's players in this category to equal the league's DRtg minus the team's DRtg. I'm going to call this metric DPA, or defensive points added, to avoid confusion (because it was calculated with an older, separate equation from SPM). Here are the 2010 leaders:
Player | Tm | G | Min | DPA |
---|---|---|---|---|
Dwight Howard | ORL | 48 | 1687 | 6.01 |
Marcus Camby | LAC | 46 | 1433 | 4.94 |
Chris Andersen | DEN | 45 | 1009 | 4.27 |
Samuel Dalembert | PHI | 47 | 1220 | 4.26 |
Gerald Wallace | CHA | 45 | 1894 | 3.86 |
Josh Smith | ATL | 46 | 1550 | 3.63 |
Joakim Noah | CHI | 44 | 1450 | 3.55 |
Anderson Varejao | CLE | 47 | 1404 | 3.38 |
Kendrick Perkins | BOS | 45 | 1306 | 3.37 |
Andrew Bogut | MIL | 39 | 1282 | 3.34 |
Lamar Odom | LAL | 48 | 1490 | 3.19 |
Ben Wallace | DET | 46 | 1383 | 3.03 |
Emeka Okafor | NOH | 47 | 1431 | 2.78 |
Rasheed Wallace | BOS | 42 | 1005 | 2.61 |
Nene Hilario | DEN | 47 | 1603 | 2.58 |
Pau Gasol | LAL | 31 | 1132 | 2.40 |
Brendan Haywood | WAS | 45 | 1501 | 2.38 |
Chuck Hayes | HOU | 47 | 1021 | 2.37 |
Thabo Sefolosha | OKC | 47 | 1406 | 2.29 |
Marc Gasol | MEM | 46 | 1639 | 2.24 |
Again, Superman flies high above the rest, combining great individual boxscore defensive stats with great team defensive performance. And based on the evidence above, it seems pretty clear that the consensus statistical pick for DPOY is the incumbent. However, all of these metrics measure who should win based on their performance... what about the question of who will win? Well, luckily I also created a model based on past voting results to predict DPOY winners, and it gives these rankings (along with the variables found to be significant -- and yes, FG40 and FGA40 were significant):
Player | Team | G | MP | tmDRtg | MPG | FG/40 | FGA/40 | Stl/40 | Blk/40 | p(DPOY) |
---|---|---|---|---|---|---|---|---|---|---|
Dwight Howard | ORL | 48 | 1687 | 103.0 | 35.1 | 6.5 | 10.9 | 1.3 | 3.1 | 0.247 |
Gerald Wallace | CHA | 45 | 1894 | 102.0 | 42.1 | 5.9 | 12.1 | 1.5 | 1.2 | 0.153 |
Rajon Rondo | BOS | 44 | 1622 | 101.7 | 36.9 | 6.6 | 12.4 | 2.6 | 0.1 | 0.061 |
Nene Hilario | DEN | 47 | 1603 | 106.1 | 34.1 | 6.2 | 10.6 | 1.9 | 1.1 | 0.031 |
Marc Gasol | MEM | 46 | 1639 | 110.0 | 35.6 | 6.3 | 10.5 | 1.2 | 1.8 | 0.029 |
Josh Smith | ATL | 46 | 1550 | 106.1 | 33.7 | 6.9 | 13.7 | 1.8 | 2.6 | 0.026 |
Jason Kidd | DAL | 46 | 1660 | 105.5 | 36.1 | 3.6 | 8.4 | 1.9 | 0.6 | 0.016 |
Pau Gasol | LAL | 31 | 1132 | 102.0 | 36.5 | 7.1 | 13.3 | 0.5 | 1.7 | 0.015 |
Kendrick Perkins | BOS | 45 | 1306 | 101.7 | 29.0 | 6.4 | 10.2 | 0.4 | 2.8 | 0.014 |
Marcus Camby | LAC | 46 | 1433 | 107.7 | 31.2 | 4.2 | 9.0 | 1.7 | 2.5 | 0.013 |
Chris Paul | NOH | 38 | 1472 | 108.3 | 38.7 | 7.8 | 15.4 | 2.3 | 0.2 | 0.011 |
Brendan Haywood | WAS | 45 | 1501 | 109.4 | 33.4 | 4.6 | 8.3 | 0.5 | 2.5 | 0.010 |
Al Horford | ATL | 46 | 1579 | 106.1 | 34.3 | 6.5 | 11.5 | 0.8 | 1.4 | 0.008 |
Joakim Noah | CHI | 44 | 1450 | 104.2 | 33.0 | 5.2 | 10.6 | 0.8 | 2.1 | 0.008 |
Ronnie Brewer | UTA | 46 | 1490 | 104.2 | 32.4 | 5.1 | 10.0 | 2.2 | 0.3 | 0.008 |
Ben Wallace | DET | 46 | 1383 | 109.0 | 30.1 | 3.0 | 5.8 | 1.6 | 1.5 | 0.007 |
LeBron James | CLE | 49 | 1889 | 102.7 | 38.6 | 10.4 | 20.6 | 1.6 | 1.1 | 0.007 |
Anderson Varejao | CLE | 47 | 1404 | 102.7 | 29.9 | 4.4 | 8.3 | 1.3 | 1.3 | 0.006 |
Andrew Bynum | LAL | 46 | 1467 | 102.0 | 31.9 | 7.9 | 13.9 | 0.6 | 2.0 | 0.006 |
Andre Iguodala | PHI | 47 | 1883 | 108.8 | 40.1 | 6.0 | 14.1 | 1.9 | 0.6 | 0.005 |
Which means the voters are likely to choose D-12 again as well, with Gerald Wallace possessing an outside chance at taking the hardware. And given the stats we saw earlier, they'll definitely get no argument from me if they pick Howard -- the man has simply been dominant this season.
February 1st, 2010 at 3:06 pm
Dwight's got to take it again I'd think. I wouldn't be surprised if Bron and Rondo came in second and third of G. Wallace because of reputation and team wins.
February 1st, 2010 at 3:39 pm
What's the advantage to going by 40 minutes instead by game? If you foulzilla like Dwight Howard is, and you can only be out there 35 minutes a game, doesn't seem like you should be rewarded for that. Wallace can be out there for 42 minutes - as a good defender that 7 minutes is very valuable.
February 1st, 2010 at 5:19 pm
That's a good point... Here's the regression re-run using per game metrics:
February 1st, 2010 at 5:59 pm
Looks like a Bobcats fan was just put in their place.
February 1st, 2010 at 6:01 pm
Is there a direct measure of how many points the guy they're supposed to guard scores?
Rondo has been getting torched.
http://www.boston.com/sports/basketball/celtics/articles/2010/01/25/opposing_point_guards_present_problem_for_celtics/
February 1st, 2010 at 7:11 pm
Negatory, Patrick. I'm rooting for the Knicks and Bucks this year.
February 2nd, 2010 at 3:19 am
Can you run the year that Hakeem Olajuwon torched Dennis Rodman in blocks and rebounds, yet Rodman won the award? That seems a travesty to me; though Rodman's Pistons had the best defense in the league, Hakeem was a better defender by the conventional stats.
Anyway, with Howard leading the league in blocks and rebounds, it seems that he is a prudent choice until the voters get tired of voting for the same person every year.
February 2nd, 2010 at 4:58 am
I wonder how accurate your model on who should win based on past results is.
I mean, the one you did on the MVP voting (NBA MVP Award Tracker) does not seem well calibrated: Mo Williams, Andrew Bynum and Lamar Odom in the MVP top 10? Doesn't it seem like you put too much emphasis on team success?
February 2nd, 2010 at 1:57 pm
Any idea why Rasheed Wallace is so high on DRtg? He's had his moments, but he's looked terrible often -- particularly on help defense. I don't know enough about how DRtg is calculated; would weak help defense not penalize a player?
February 3rd, 2010 at 9:54 am
Dwight Howard is underrated. He will become the first player ever to lead the league in rbs and blocks two seasons in a row after the year is done. Go Magic!
February 4th, 2010 at 5:23 pm
Jared Ras
"Can you run the year that Hakeem Olajuwon torched Dennis Rodman in blocks and rebounds, yet Rodman won the award? That seems a travesty to me; though Rodman's Pistons had the best defense in the league, Hakeem was a better defender by the conventional stats.
Anyway, with Howard leading the league in blocks and rebounds, it seems that he is a prudent choice until the voters get tired of voting for the same person every year."
Actually the Rockets had the best defensive efficiency that year and they beat out the Pistons by .1 percent. However, the Pistons had the lowest points per game allowed (because of their slow pace) and Rodman's perimeter D, even at that time, was becoming legendary. However, Olajuwon's performance was better, in fact it was arguably the best defensive season by any player since the merger. He lead the league in defensive rebounding percentage, was 3rd in the league in block percentage (while leading the league with the fifth highest bpg average ever at 4.59), and added 2.1 STEALS PER GAME, his third consecutive year in a row finishing in the top ten in steals. Sometimes defensive stats lie, but not in the case of The Dream. He was that damn good on defense.
February 24th, 2010 at 3:33 pm
What about charges taken? Or is that factored into forced turnovers? A charge taken is much better than a blocked shot. A blocked shot might go out of bounds back to the offensive team or if might get blocked back to a player on the offensive team. A charge taken is both a foul on the offensive player and a guaranteed switch in possession. Andrew Bogut is 2nd in blocks and 4th in charges taken plus he's top 10 in rebounding. That sounds like a solid case to be a bit higher on the list than he currently is.