The voting for the NBA’s all-star game has concluded. And the starters for the 2007 game are as follows [Wins Produced and Wins Produced per 48 minutes in parenthesis].
- Western Conference
- Kobe Bryant, Guard (7.4; 0.241)
- Tracy McGrady, Guard (4.3; 0.174)
- Kevin Garnett, Forward (11.5; 0.351)
- Tim Duncan, Forward (9.6; 0.329)
- Yao Ming, Center (4.2; 0.216)
- Eastern Conference
- Dwyane Wade, Guard (8.1; 0.293)
- Gilbert Arenas, Guard (6.4; 0.185)
- LeBron James, Forward (9.0; 0.258)
- Chris Bosh, Forward (4.7; 0.209)
- Shaquille O’Neal, Center (-0.5; -0.213)
With the exception of Shaq, who only played four games this season (and was obviously hurt), all of these players are above average (average WP48 is 0.100) this season.
A Study of Fan Voting
It’s important to note that these players are selected by the fans, and it’s not clear how much weight fans place on current season performance (or how fans evaluate performance) in voting for these players.
Thus far I have not seen a specific study examining what factors explain fan voting for the NBA All-Star game. There was a nice study in Economic Inquiry of fan voting for baseball’s All-Star game by F. Andrew Hanssen and Torben Andersen. The Hanssen-Andersen work [Has Discrimination Lessened Over Time? A Test Using Baseball’s All-Star Vote] appeared in 1999 and explicitly studied the level of racial discrimination by fans in baseball. These authors found that fans did discriminate in 1970, but that behavior had disappeared by the mid-1990s.
Beyond the issue of race, these authors found that baseball fans considered such factors as current season performance, career performance, past playoff participation, team performance, and the popularity of the team. One suspects that the same collection of factors would be relevant to NBA voters.
Looking at Wins Produced
What if fans, though, only considered performance in the current season? And let’s just say that the performance metric considered was Wins Produced? And let’s say that only performance in the first 41 games was considered relevant? Well, if you can go along with all these statements, then — as the following tables indicate — the starters in this game would be somewhat different.
Table One: Evaluating the Western Conference All-Stars in 2007
Table Two: Evaluating the Eastern Conference All-Stars in 2007
In the Western Conference the starters would be include Steve Nash and Manu Ginobili at guard, Garnett and Dirk Nowitzki at forward, and Marcus Camby at center. For the Eastern Conference the starters would be Wade and Chauncey Billups at guard, James and Emeka Okafor at forward, and Dwight Howard at center.
Those would be the starters if we only considered players who finished in the top ten in voting. If we consider all NBA players, then David Lee of the New York Knicks would edge out Okafor. Lee, who appears on a quest to save the job of Isiah Thomas, had produced 9.8 wins and posted a WP48 of 0.386 through the first 41 games. In other word, Lee is currently one of the ten most productive players in the game. And Isiah doesn’t even start him.
The NBA’s MPP
Although Lee is playing very well, he is not quite the league’s Most Productive Player (MPP). For the past four seasons the MPP has been Kevin Garnett. So far this year, though, Jason Kidd is leading the way. If this continues, this will be the second MPP title for Kidd (he was also the MPP for the 1998-99 season). Kidd does benefit from playing for a generally poor team, but the effect is not so large that it would erase the difference between him and Garnett (who also plays on a bad team), Nowitzki, or Nash.
Of course MPP is not the same as MVP. Aju Fenn and I have a working paper (that we need to finish) that examines the media’s voting for the Most Valuable Player. The three factors that appear to matter most are scoring, assists, and team wins (and scoring tends to trump assists).
If we consider these three factors, the leading MVP candidates should be Nowitzki and Nash. Both of these players generate points and/or assists for the two best teams in the league. If the vote only considered Eastern Conference players, then LeBron James and Gilbert Arenas would be the top candidates.
Unfortunately for James and Arenas, the Western Conference does exist. Consequently, a third MVP for Nash is very much a possibility.
– DJ
Wages of Wins Journal 
anon
January 30, 2007
Is Manu Ginobili really that good? More productive than Wade, Nash, and McGrady? In terms of raw stats, Nash is a much better shooter and passer, but Ginobili is a better defender. Wade, however, has him beat in every category except turnovers.
Can turnovers and minutes played really be sufficient to make Ginobili more productive than Wade and Nash?
dberri
January 30, 2007
Ginobili has both a higher Win Score per-minute and NBA Efficiency per-minute than McGrady. So by either measure, Ginobili is preferred.
With respect to Wade, you have to remember that Ginobili has a much higher level of shooting efficiency (as well as other advantages).
Nash has slightly better stats, but he is hurt by playing on a team playing at a very fast pace. When you consider the team adjustment (which typically has virtually no impact on a player’s overall level of productivity as measured by Wins Produced), Nash dips a bit below Ginobili.
anon
January 30, 2007
How is shooting efficiency measured?
dberri
January 30, 2007
I use points-per-shot. (pts-ftm)/fga
Mavis Beacon
January 30, 2007
I’ve seen your formula and it’s a bit over my head. I noticed you critized Hollinger’s ratings for driving up a player’s rating who shoots more shots while making them at an average percentage. That seems like a fair critique.
In your system, if a player like Ben Wallace takes a shot, misses it, gets his own rebound, and makes the next shot, is he better off than if he’d made the first shot to begin with? Thanks.
dberri
January 30, 2007
Mavis,
The scenario you describe does not make Wallace better off. He loses from missing a shot the same amount he gains from the rebound. Then he gains from making the shot. So that would be equal to just making a shot.
hk
January 30, 2007
What is your take on usage vs. efficiency?
dberri
January 30, 2007
hk,
I really like Dean Oliver’s work (heck, Dean is a friend of mine so I like Dean, too), but I was not convinced by that chapter in Basketball on Paper. I just don’t think there is strong evidence that a player’s shooting efficiency is impacted by the number of shots he takes. If there is an effect, it must be quite small.
Jack Mott
January 31, 2007
Interesting, I had come here to ask how in the world Jason Kidd is so much better than Nash according to your Wins Produced numbers, and here I see the answer. I had not considered the fact that the Suns increased # of posessions per unit time would inflate his rebounds/assists/steals etc.
It is still amazing to me that Nash having such an incredibly high shooting percentage and assist rate isn’t more valuable than it is. Perhaps you can delve into the math on this a bit.
Jack Mott
January 31, 2007
Another thought. It would be interesting to see some charts of which players so far this year are generating the most wins for the dollar, and least wins for the dollar!
:)
The Franchise
January 31, 2007
I’ll second Jack Mott’s request—it seems that some sort of correlation between salary and WP could be very effective in analyzing the decision-making capabilities of General Managers. I don’t know if WP/salary is the right equation, because even the players that have negative WP are probably better than not filling that roster spot, but it’s a step in the right direction for that sort of analysis.
For instance, besides paying Kevin Garnett very well (which anyone that has watched him play would do) does McHale make any other good decisions in hiring players? My guess is that they are rare indeed. On the other hand, I’m sure that Buford and Popovich in San Antonio are very good at picking the right players to pay, even beyond Tim Duncan, who everyone knew was valuable coming out.
dberri
January 31, 2007
We did do this sort of analysis in The Wages of Wins. We found that salaries in the NBA are dominated by points scored (and that shooting efficiency, turnovers, etc… are not that important).
We also presented a chart with over-rated and under-rated players (which was updated for 2005-06 on Malcolm Gladwell’s blog). Perhaps when this season is over I will present that information again, as well as a table with players who are the best and worst bargains.
The Franchise
February 1, 2007
Right. So I’ll actually need to read the book, then. (Once I have the disposable income required to purchase it.)
This is the crux of your (correct) argument about how bad Thomas has been as a GM—that he traditionally signs scorers, and not role players, while teams like San Antonio or Utah have been more likely to have more well-rounded rosters, potentially in part because they can’t afford a roster full of scorers.
Not telling
March 10, 2007
The way i see it, i thought that nash should have definately been a starter. I think ginobili should have been a starter too or at least have been in the all-star game. I can’t believe parker made it instead of manu, that is so messed up. Ginobili does a lot more than parker, and definetely should have been in there.