4 of the 6 best players in the league are white

[Wannabe MEP]

You wouldn't take Amare over Collison?

Nope. Not even close.

Amare has a negative APM for every single year at basketballvalue.com. In both of the multi-year APM/RAPM studies I have access to, Collison wins. In the most recent one, it's a total blowout: 6.2 to 0.2. That gap is enormous. That's like the difference between Dwyane Wade and Rodney Stuckey.

Also, I don't believe you can have an elite team without a solid defensive frontcourt. Amare will never win a championship because he can never be part of a championship-caliber defensive frontcourt. Collison absolutely could. (Although Amare has really cool highlight reels . . . so it's kind of a trade-off.)

[azuresou1]

You can win a championship if Amare is your second best player. You can't if Nick Collison is your second best player. Two year APM also indicates Nick Collison is better than Dwight, and actually any big man other than Dirk.

While I'd like to believe in APM/RAPM, there are too many flaws I see in the general methodology.

[Wannabe MEP]

You must have AN ELITE defensive frontcourt to win a title.
The numbers below are multi-year defensive RAPM.

Celtics '08
6.2---Garnett
1.8---Kendrick Perkins
1.2---Paul Pierce
3.9---Tony Allen
0.6---James Posey

Lakers '09
2.4---Lamar Odom
1.3---Andrew Bynum
1.0---Trevor Ariza
0.0---Pau Gasol
1.1---Luke Walton

Lakers '10
2.8---Ron Artest
2.4---Lamar Odom
1.3---Andrew Bynum
0.0---Pau Gasol
1.1---Luke Walton

Mavericks '11
2.9---Tyson Chandler
2.2---Dirk Nowitzki
1.3---Brendan Haywood
0.8---Shawn Marion

Amare Stoudemire: -1.7
(Carmelo Anthony: -1.6 . . . Knicks, we have a problem)

You can win a championship if Amare is your second best player.

No. You can't. Now, if he's playing next to Bill Russell 2.0 and Scottie Pippen 2.0: sure you can. He could even be leading scorer on that kind of team . . . but that doesn't mean he's the 2nd best player.

[jambalaya]

On Ilardi's 6 year traditional APM Collison ranks about 33rd among smaller minute players. On Englemann's 10 year RAPM Collison is about 55th overall. Collison"s big RAPM numbers are from just the last 2 seasons. Those are his only seasons clearly above +2 (and +4). The prior years were basically between neutral and +2. Let's see if he pulls more seasons over +2 or over +4.

The past 2 seasons may have been his peak, a great role, and / or influenced by error in the estimate including multicollinearity issues with Green. Or he could show that he has more great impact seasons left. Nick will probably do it again at least one more time but I am not sure how many more such seasons he has left in him. It could come down to how much of his value is intelligence based vs based on his athleticism and how much that athleticism changes from here and how fast. Obviously having enough of both could yield larger returns than if one slips below a desirable threshold.

Battier only pulled 2 seasons clearly above +2 on RAPM. Morey traded for him and Battier never duplicated those numbers in Houston. Either he peaked in Memphis or there were some multicollinearity issues that heightened his apparent RAPM value or both.

James, Notwitski, Nash, Ginobili and Duncan all have 6-9 years over +4. Those are clearly RAPM superstars. Few have achieved 4 or more seasons over +4 during the last deade. Wade now has 4. He might have joined the lower-end of that top-tier. Kobe has 5, but 2 years of decline and just barely above +4 last season. Is his +4 RAPM run about over or will he recover? Paul has 3 seasons over +4. Howard has 2. They are examples of guys pretty sure to be on their way to 4+ over +4. Amare... none over +2.

Amir Johnson has 1 but a very steady upward trend. Bogut none over +4 yet but the last 2 were close. Will be interesting to see what they do from here as with Collison. 4 years above +4 probably makes you a RAPM superstar. 3 maybe. 2 seasons probably doesn't provide quite enough evidence to me to be sure about designating that guy as a RAPM superstar. Could still be a short-term star though.

There would be math that could be done to estimate likelihood of true RAPM value over +4 based on a player's history if one wanted to try to estimate based on data available to date.

If you haven't already discovered it, click on a player's name at the RAPM and see there historical chart. Nick's:
http://stats-for-the-nba.appspot.com/players/172.png

Nash, Ginobili, Nowitski and Collison (and probably most players with high RAPM scores) all had primary subs with near team worst traditional Adjusted +/-. 3 of the 4 were near team worst on RAPM (Marion wasn't). All versions of APM even RAPM (those less so) are susceptible to overdoing the ranking separation of players who are primary subs.

[Doctor MJ]

Good break down jambalaya. When we first started seeing +/-, I had an open mind to the idea that there could be secret (non-box score) superstars in basketball like there can be in other sports (offensive lineman in football, and heck, Xavi in soccer), it just doesn't seem like it happens. Players can be underrated by box score stats, but superstars play big minutes and tend to touch the ball frequently. Everyone else doesn't show huge APM numbers from year to year.

I still think it's possible to have Top 60 (roughly #2 on a team worthy) with shockingly weak box score stats simply because a lot of more talented players never figure out the nuances of the game, but Top 10 impact? Nah.

btw, re: Engelmann's historical charts. Aren't each of his years compiled through not-quite identical methods (lambdas, priors)? I'm hesitant to use his numbers that literally because it seems like he's trying to show the best estimate how good a player was at a given moment in time (given all that he had done to that point), not calculate year-to-year impact.

[jambalaya]

I'm a little afraid to ask: What's the average APM of white players in this league?

This is unweighted by possessions played so it is not the best possible answer (Jerry's site doesn't give the possessions played beside the player's rating) but it appears that for the 2010-11 season the unweighted average non-"black" players (could be of European ancestry, including Hispanic, or Asian or Middle Eastern) were a couple tenths of a point better on offensive RAPM, about a tenth better on defensive RAPM for a total RAPM edge of about .35 points per 100 own possessions over "black" players (African-American or African or black Caribbean).

[jambalaya]

Thanks Doctor MJ.

Yes Engelmann's data have different priors each year. Not sure if the lambda changes or not; it is not stated. They might. He has generally tried to find the best outcome for a year. Your concern about method change(s) affecting the year to year comparison are on point.

I agree that the numbers should be used with caution. Because of the details of the RAPM method he has declined to offer standard errors, saying it can't be done precisely or at least easily. But he has estimated that they are lower than traditional RAPM for both onw year and multi-season comparisons and I believe that.

[azuresou1]

You must have AN ELITE defensive frontcourt to win a title.
The numbers below are multi-year defensive RAPM.

Celtics '08
6.2---Garnett
1.8---Kendrick Perkins
1.2---Paul Pierce
3.9---Tony Allen
0.6---James Posey

Lakers '09
2.4---Lamar Odom
1.3---Andrew Bynum
1.0---Trevor Ariza
0.0---Pau Gasol
1.1---Luke Walton

Lakers '10
2.8---Ron Artest
2.4---Lamar Odom
1.3---Andrew Bynum
0.0---Pau Gasol
1.1---Luke Walton

Mavericks '11
2.9---Tyson Chandler
2.2---Dirk Nowitzki
1.3---Brendan Haywood
0.8---Shawn Marion

Amare Stoudemire: -1.7
(Carmelo Anthony: -1.6 . . . Knicks, we have a problem)

You can win a championship if Amare is your second best player.

No. You can't. Now, if he's playing next to Bill Russell 2.0 and Scottie Pippen 2.0: sure you can. He could even be leading scorer on that kind of team . . . but that doesn't mean he's the 2nd best player.

Jason Kidd
Landry Fields
Shane Battier
Amare Stoudemire/Nick Collison
Dwight Howard

Between those two, you honestly think Nick Collison and his efficient but limited offense has more value than Amare? Because no one else on that team outside of Dwight is going to be capable of creating any offense for himself.

[mysticbb]

I agree that the numbers should be used with caution. Because of the details of the RAPM method he has declined to offer standard errors, saying it can't be done precisely or at least easily. But he has estimated that they are lower than traditional RAPM for both onw year and multi-season comparisons and I believe that.

He offers out of sample testing, which is a good enough test in order to determine whether they are better or worse than standard APM. And RAPM is indeed doing better in out of sample tests.

Anyway, I agree with Doc MJ that we can't throw out minutes played here. Collison is very effectiv in his limited minutes, not necessarily also in increased minutes. Thus, we can't say that it is wise to play him more. We have to factor in multiple things like endurance, fit or matchups. Collison is a very good weakside and help defender, he is among the league leaders in charges drawn. On offense he is a good finisher around the rim and still a 70+ free throw shooter.

I would also take Collison over Stoudemire, because Stoudemire's weakness on defense can only be compensated by a really good front court defender like Howard or Bogut. Stoudemire needs to play defensively as PF, because he is unable to defend his position against stronger players. Stoudemire can be easily moved around underneath the basket and is rather trying to challenge a shot instead of defending the position. The latter is much more needed from a frontcourt player, because even if someone can block 2 or 3 shots a game he will still give up more by giving the opponents frontcourt the space to operate effectively. That is an issue I can't easily compensate and one of the reasons the Suns defense as well as the Knicks defense suffered with Stoudemire on the court. On offense Stoudemire occupies the room of an agile center, someone who would be very valuable next to Stoudemire on defense. Thus, it is hard to find a fitting center to make sure the defense works.

And when I have a guy like Howard or Bogut I can still settle for cheaper frontcourt options than Stoudemire. It makes especially no sense to give max money to a player like Stoudemire when I have Dwight Howard already. Much better would be stretch 4 like Ryan Anderson, who is also much cheaper. And as backup Collison is the much better and more versatile solution.
Stoudemire makes only sense, if I want to settle for mediocricy, but still want to earn money. Stoudemire simply attracts more people willing to spend money than Collison. For a winning team Stoudemire is not the best option, because he demands 1st option money, but is not good enough to carry a team like other max money options (including the names listed in this thread already like James or Nowitzki).

[Wannabe MEP]

Jason Kidd
Landry Fields
Shane Battier
Amare Stoudemire/Nick Collison
Dwight Howard

That's a pretty interesting scenario. I don't know what I'd do there. Take Collison, and damn, you can enforce your will defensively. Also, I can see Collison working well with guys like Fields and Kidd on offense, like he does with Harden and Maynor. Not your typical NBA iso-offense, but they might be able to get things going like the '04 Pistons with ball-movement.

But you're right: Stat brings an offensive spark that's not otherwise there. Is that worth his clueless defense? I honestly don't know in that situation. I'd probably vote no, but maybe that's partly because I'm a jaded Suns fan who thinks the Suns would have a couple titles by now if Amare's defense was only moderately terrible instead of epically terrible.

[Wannabe MEP]

Some rather good discussion happening here. Thanks for the well thought-out contributions, all.

One thing I keep asking is: Are there any "fit" issues happening with the Thunder?? What I mean by "fit" is that a player's impact is greatly affected by the particular roles of the players on the court. I've been looking for situational impact in the units on basketballvalue.com: I'm not finding anything. But the reality is that I don't know OKC's personnel very well, so I'm curious about whether there's something I'm missing.

For example: in 2009-2010, of the 9 Suns players with the most minutes, the four lowest APM scorers were the four guys who don't shoot three pointers: Amare Stoudemire, Lou Amundson, Grant Hill, and Robin Lopez. The Suns usually had 2, 3 or 4 3-point shooters on the court. Almost without exception, the Suns were much more effective with four 3-point shooters on the court. Stat, Lou, and Fropez were all decently effective if they were the lone paint-dweller, but not if two of these three were out there together. APM doesn't distinguish between those role differences, so Lou gets poor APM numbers even though he was actually very, very effective when he was paired with Frye.

[Collison's] past 2 seasons may have been his peak, a great role, and / or influenced by error in the estimate including multicollinearity issues with Green. Or he could show that he has more great impact seasons left.

First of all, Green's impact has to be limited because:
-good basketball is good basketball. One way or another, OKC was playing really, really good basketball with Collison on the court: 12 of his 14 units in the 2010-2011 season have a positive APM score, and his raw on-court rating was +11.13. That is not effected by Jeff Green. That +11.13 was 4th in the NBA, behind only Garnett, Ginobili, and Bosh, and the only other OKC player in the top 45 is Maynor.
-Collison's raw on/off numbers are superb for the 2011 playoffs, after Green was no longer on the team.

Now let's suppose that Collison had a "great role" for the past two seasons. Is that a bad thing? APM/RAPM is ALWAYS, ALWAYS, ALWAYS situational. But that's ok: a player's impact is situational because basketball is situational. Every stat is situational. That doesn't matter if you're a ball-dominating "star" or a behind-the-scenes defensive "star"--you have to be around the right people to be effective.

Battier only pulled 2 seasons clearly above +2 on RAPM. Morey traded for him and Battier never duplicated those numbers in Houston. Either he peaked in Memphis or there were some multicollinearity issues that heightened his apparent RAPM value or both.

Or, there were some multicollinearity issues that REDUCED his apparent RAPM value in Houston. Thing is, Houston was really good with Battier. In 2008-2009 the Rockets could have won the title if both McGrady and Yao Ming hadn't gone down with injuries; they still pushed the Lakers to seven games. The raw on-court data shows that the Rockets were playing superb basketball with Battier on the court . . . but Ron Artest was also there. You couldn't have a closer imitation of Battier's skillset. Raw on-court team leaders goes like this:
1) Yao +8.63
2) Artest +6.48
3) Battier +6.47

So Artest and Battier stole each other's thunder, but that was really, really good basketball. Now Battier isn't an offensive spark, so he needs to play next to guys who can create offense . . . which hasn't really happened since those McGrady and Yao injuries in 2009 (although Conley-Mayo-Battier-Randolph-Gasol was starting to look scary good).

If Steve Nash and Chris Paul were on the same team, the offense would hum but each player's RAPM value would suffer. The exact same thing happened with LeBron and Wade: In basketballvalue.com's one-year APM, LeBron and Wade went from #2 and 4 in the league in 2009-2010 to 13 and 24 in 2010-2011.

On Ilardi's 6 year traditional APM Collison ranks about 33rd among smaller minute players.

I'm looking at Ilardi's 2007-2008 multi-year APM results posted on 82games.com: on that one Collison is 37th in the big minutes grouping. Is Ilardi's 2008-2009 study considerably improved for some reason? Why isn't it posted on 82games.com? And why are you guys ignoring the 2007-2008 results?

I guess the big question going forward is: Where does Collison land in RAPM after the 2011-2012 season (assuming there is one)?

[azuresou1]

Jason Kidd
Landry Fields
Shane Battier
Amare Stoudemire/Nick Collison
Dwight Howard

That's a pretty interesting scenario. I don't know what I'd do there. Take Collison, and damn, you can enforce your will defensively. Also, I can see Collison working well with guys like Fields and Kidd on offense, like he does with Harden and Maynor. Not your typical NBA iso-offense, but they might be able to get things going like the '04 Pistons with ball-movement.

But you're right: Stat brings an offensive spark that's not otherwise there. Is that worth his clueless defense? I honestly don't know in that situation. I'd probably vote no, but maybe that's partly because I'm a jaded Suns fan who thinks the Suns would have a couple titles by now if Amare's defense was only moderately terrible instead of epically terrible.

04 Pistons had excellent ball movement, but they also had 3 guys who were capable of creating fairly consistently for themselves, and they needed to at times. Take Collison, and only Dwight can really regularly create for himself.

Stat's a pretty bad defender, but with Dwight there to cover up for him, I can't see how you don't take the sheer massive advantage in offensive production when your defense is already spoken for. IMHO with Collison that team barely makes it into the playoffs because there's absolutely no offense, while with Amare that team is a championship contender.

[jambalaya]

Thanks for the replies Los Soles. Here are some in return:

Yes OKC played really good basketball with Collison on the court. But his Adjusted +/- tries to make sense of all minutes of the team (and league), with him on the court and off (not just on) and the fact that the team was barely above neutral with him off the court affects how much credit Nick got. (I think.)


No Thunder player had as strong a raw +/- on / off as Nick. Green's was the second most lopsided negative. Westbrook's was the most negative. Only a little more than 1/4th of Westbrook's minutes were with Nick. Nick was about 50 /50 with Westbrook and with Maynor. Maynor-Collison was about 3 times as positive as Westbrook-Collison. Nick's raw on/off was influenced by playing with Green about 1/3rd of the time, 2/3rds not. Nick on had Westbrook half the time while Nick off had Westbrook almost 3/4s the time.

Harden with Westbrook was +0.6 per 48 minutes on raw +/-; without Westbrook, +9. Durant with Westbrook, +4.3; without him, +12.8. Collison, Harden and Durant aren't the only ones dramatically worse with Westbrook and his 2010-11 monster usage rate than without him but Harden and Durant are the most important cases.

Many say Westbrook had a great year, deserved to an All-Star, should keep shooting as much as he did, should get the max... I would not put it as strongly or as simply positive on all of that. Talented, rare player to be sure, who made some improvements including on shooting / scoring efficiency, but the team +/- numbers with him on the court are not were you'd want and expect for an elite team's starting PG.

Him shooting less is one option to perhaps help with that. Or shooting better from mid-range or 3 pt. land. Would he also play harder and better on defense if his offensive usage was lower, back at or closer to 26% from 2009-10 instead of the 32% in 2010-11? The team did better on +/- overall with him on the court back then when he was at the lower usage. He could probably still do more to help his teammates beyond the direct assist opportunities (i.e. giving up the ball early or making the pass that has time for another pass and the opportunity for a good one). Less turnovers would be helpful too (he was #1 in the league in total turnovers for 2nd time in 3 years and was 4th the other time). Maybe a bit less minutes would be useful as well to allow him to play hard at both ends and give more room for Harden to get shots and maybe Ibaka too (as long as Maynor's on the court +/- and personal RAPM stay at last season's level).


I don't consider myself to have ignored Collison's traditional APM in 2007-8 but it was a long time ago and 1 year traditional APM is the least reliable (highest error) form of APM. It was also the 2nd and last year of him in an over 2000 minute role. Only Durant remains from the team. I put more emphasis on the multi-season estimates and / or the most recent single season RAPMs.


Yes it is also possible that Battier's Adjusted +/- was limited by the good +/- performance of Artest and then Budinger. I was probably guilty of leaning my narrative toward an interpretation.

New York definitely needs to figure out who to put beside Amare, different from their current options if they are to have any chance to go beyond the 1st or 2nd round of the playoffs. The Suns got very good regular season performance with him beside both the offensive-minded Frye and the defensive-minded Lopez so at least that much can be done. The playoffs are a different level of challenge though. Frye had success beside Amare in the 2010 playoffs, Lopez didn't.

[Wannabe MEP]

No Thunder player had as strong a raw +/- on / off as Nick. Green's was the second most lopsided negative. Westbrook's was the most negative. Only a little more than 1/4th of Westbrook's minutes were with Nick. Nick was about 50 /50 with Westbrook and with Maynor. Maynor-Collison was about 3 times as positive as Westbrook-Collison. Nick's raw on/off was influenced by playing with Green about 1/3rd of the time, 2/3rds not. Nick on had Westbrook half the time while Nick off had Westbrook almost 3/4s the time.

What's interesting though is factoring in the playoffs:
-No Jeff Green
-Maynor's minutes plummeted, and so did his raw impact.

Meanwhile, Collison's impact was superb again. The other players with good raw numbers were again players who had a high correlation of minutes with Collison. Factoring in the playoffs, it seems pretty clear that Collison was definitely effecting his teammates numbers, and I see much less of a trend in the reverse.

For example here are units from the playoffs that played at least 10 minutes against Memphis, from BV. The first group below is with Collison, the second group without.

Westbrook - Harden - Durant - Collison - Perkins +26.87
Maynor - Cook - Harden - Collison - Mohammed +15.20
Westbrook - Sefolosha - Durant - Collison - Perkins +8.51
Westbrook - Harden - Durant - Ibaka - Collison +14.20
Maynor - Harden - Durant - Collison - Mohammed +15.65

Westbrook - Sefolosha - Durant - Ibaka - Perkins -4.87
Westbrook - Harden - Durant - Ibaka - Perkins +0.25
Westbrook - Cook - Harden - Durant - Perkins -35.41

The only trends in there that I can see are:
1) OKC was simply more effective with Collison on the court (at least against Memphis).
2) Collison was more effective with Harden (but he's still effective without him).

Playoff raw +/-
. . . . . . . . . w/ Collison . . . without
Harden: . . . . . +81 . . . . . -1
Westbrook: . . . +64 . . . . -35
Durant: . . . . . +60 . . . . . -9
Nazr . . . . . . . +24 . . . . . -2
Cook . . . . . . . +27 . . . . . -8
Maynor . . . . . +3 . . . . . -12
Ibaka . . . . . . +10 . . . . . -35
Perkins . . . . . +43 . . . . . -77
Sefolosha . . . . +4 . . . . . -54

[Wannabe MEP]

For example here are units from the playoffs that played at least 10 minutes against Memphis, from BV. The first group below is with Collison, the second group without.

Westbrook - Harden - Durant - Collison - Perkins +26.87
Maynor - Cook - Harden - Collison - Mohammed +15.20
Westbrook - Sefolosha - Durant - Collison - Perkins +8.51
Westbrook - Harden - Durant - Ibaka - Collison +14.20
Maynor - Harden - Durant - Collison - Mohammed +15.65

Westbrook - Sefolosha - Durant - Ibaka - Perkins -4.87
Westbrook - Harden - Durant - Ibaka - Perkins +0.25
Westbrook - Cook - Harden - Durant - Perkins -35.41

I looked at this for the other series. Exact same thing against Denver. Very similar against Dallas, except that these two units . . .

Maynor, Eric - Westbrook, Russell - Harden, James - Durant, Kevin - Collison, Nick
Westbrook, Russell - Sefolosha, Thabo - Durant, Kevin - Ibaka, Serge - Collison, Nick

. . . lost the Collison magic against the near-perfection that is Kidd-Terry-Marion-Dirk-Chandler.

[jambalaya]

I usually have been Nick's advocate and an advocate of +/-. You had that had covered, so I explored the other side and noted some cautions and possible factors that could be involved with his strong ratings. But Collison's raw and Adjusted +/- in the last 2 seasons (including playoffs) are really good and don't look anything like a fluke.

Kidd-Terry-Marion-Dirk-Chandler was generally outstanding for them / tough for the opponent in the playoffs including against Nick. But in the series against Dallas when Nick was on the court against any Chandler lineup (excluding lineups played less than 1 minute), the Thunder was only -1 in about 78 minutes while the rest of the team minutes were -31 in 162 minutes.

Nick played about 24 minutes a game in the playoffs, up about 15% from the regular season. Could he have played more minutes effectively? In the regular season he played 28+ minutes 9 times and had an average raw +/- of +10 per 48 minutes. In the playoffs he played that long 9 times but against the tougher average competition only average about +3. Still positive in every game. Against Dallas his minutes went from 17 in the first game to 26 in the first victory in game 2 and up to 38 minutes in the last game. His +/- was positive in all but one game. If he had played more in games 1 or 3 they might have had a better chance to win an extra game but they were probably going to lose the series eventually anyways. With Westbrook on the court against Dallas the team was -30. So without Westbrook on the court they were nearly even with Dallas but lost by about 7.5 pts per 48 when he was on the court. The only game they won, he played by far his least minutes, because of the late benching. Westbrook had positive +/- thru the first 2 rounds but it went drastically south against Dallas. One series. But against the Lakers in 2010 Westbrook's +/- was basically all of the point margin loss to the Lakers too and the bench again played the champ's bench even. Maybe they needed to change the starting lineup each time (hopefully without losing the edge the bench had). More minutes for Nick with both the starters and bench was one option. More Harden and Cook and less Perkins (or less Perkins with most or all of the starters) were the other main promising options this time, especially when the Dallas series turned against them. But of course they stayed completely rigid with the starting lineup both times.

[Krodis]

As a Thunder fan, I don't think there's any particular "fit" issue with Collison that would be inflating his numbers. He essentially plays the same role as all our other bigs.

[jambalaya]

True Krodis, the role for bigs for the Thunder is mostly rebounding and defense with very limited shooting & scoring. In fact they were 3rd lowest on shots from PFs & Cs and 2nd lowest on points. They probably would have last (and maybe by a sizable margin) without the shots & scoring from Green for most of the season. They will be last next season unless the design is to get Ibaka at least several more shots.

Can you win in the playoffs with extremely low scoring from interior players? Title winning Detroit was 19th on big man points in 2003-4 but every other champ since the Jordan lead Bulls has been above average on interior player scoring.

The Thunder hope to compensate with a big perimeter driving game, like the 90s Bulls had. But if you are going to emphasize defensive bigs you better get that defensive efficiency down. With Perkins late last regular season they were better but then in the playoffs they had a below playoff average defensive performance. Against 2 top offenses but that is what you often have to go thru in the west to get to the Finals and a title.

[Wannabe MEP]

Playoff raw +/-
. . . . . . . . . w/ Collison . . . without
Harden: . . . . . +81 . . . . . -1
Westbrook: . . . +64 . . . . -35
Durant: . . . . . +60 . . . . . -9
Nazr . . . . . . . +24 . . . . . -2
Cook . . . . . . . +27 . . . . . -8
Maynor . . . . . +3 . . . . . -12
Ibaka . . . . . . +10 . . . . . -35
Perkins . . . . . +43 . . . . . -77
Sefolosha . . . . +4 . . . . . -54

I still think those numbers are absolutely unreal. Thought I'd check the 2010 numbers as well:

. . . . . . . . . w/ Collison . . . without
Harden: . . . . . +12 . . . . . -8
Westbrook: . . . +31 . . . . -43
Durant: . . . . . +35 . . . . . -35
Maynor . . . . . +2 . . . . . . -2
Ibaka . . . . . . +27 . . . . . +4
Sefolosha . . . . +8 . . . . . -32
Green . . . . . . +12 . . . . . -43
Krstic . . . . . . . +5 . . . . . -46

[brenner]

I am not an expert on stats so I am just following the conversation as a bystander, but wow, those Collison playoffs +/- numbers are scary.