In 2003-04 the Phoenix Suns won 29 games. In the summer of 2004 the Suns signed free agent point guard Steve Nash. With Nash on board the Suns improved to 62 wins in 2004-05, one of the biggest leaps made by an NBA team in the history of the league.
The progress the Suns made was largely attributed to Nash, who was named league MVP in both 2005 and 2006. That Nash is very productive is supported by the data. In his ten year career Nash has produced 87 wins. The past two years he has been especially productive, producing 16.1 wins in 2004-05 and then 18.6 wins this past season.
But as the review of the 2005-06 Suns reveals, Nash is not the only top performer employed by Phoenix. Shawn Marion produced 22.6 wins last year, a mark that led the team. This performance was not unusual for Marion. In his career he has produced 127 victories.
Of course, some might argue that despite Marion’s career numbers, his productivity these past two seasons is a result of having Nash for a teammate. And there is some evidence to support that argument. A review of Marion’s career – posted HERE – indicates that prior to Nash arriving Marion had never recorded a point-per-shot [(PTS-FTM)/FGA] in excess of 1.0. In the two seasons with Nash, Marion has had a PPS of 1.04 in 2004-05 and 1.12 this last year. So Marion has become a more efficient scorer, and one suspects having Nash as the team’s point guard played a role.
Of course, one has to note that the causality probably runs both ways. Having a scorer like Marion also improves the productivity of Nash. After all, one only gets an assist if the scorer makes the shot.
If we look beyond scoring, we see some evidence that Marion improved independently of Nash. Consider Marion’s Win Score per minute. Although Marion has always been above average, his productivity these past two years has been extraordinary. Looking a bit deeper into the numbers, though, reveals that Marion’s improvement is tied to his rebounding. On a per-minute basis his offensive rebounding had improved by 12% with Nash as a teammate, and his defensive rebounding numbers have increased by 24%. And that last piece of evidence suggests that Marion’s improvement is not entirely linked to Nash. At least, it is hard to understand how Nash’s point guard skills have led Marion to improve his rebounding on the defensive end.
So does Nash make Marion better? Does Marion make Nash better? Does it matter? Both of these players have been productive NBA players. That is especially true when they played together, but it is also true when they played on different teams.
Although Marion and Nash are the two most productive players employed by Phoenix, they do have some help. Both Boris Diaw and Kurt Thomas were above average performers last season [average Wins Produced per 48 minutes is 0.100]. And it is possible that the team will get Amare Stoudemire back this year – although his health issues certainly raise a few red flags. When healthy, Stoudemire can be very good. In 2004-05 Stoudemire produced 13.5 wins and had a WP48 of 0.225.
If the Suns can add a healthy Stoudemire it can field a line-up of Nash, Raja Bell, Diaw, Marion, and Stoudemire, with Thomas also playing an important role. Such a line-up can be extremely formidable and clearly makes the Suns one of the legitimate contenders for an NBA title.
Will that happen in 2007? The problem for the Suns is the same problem for every other Western Conference team. Most of the top NBA teams play out West. And all these teams cannot win the Western Conference title. In fact, obviously all but one team out West will ultimately be disappointed.
The Suns can probably expect Marion and Nash to again lead the team. But for the team to surpass the Spurs and Mavericks it will need Stoudemire. And if he cannot be healthy this season, the odds increase that Phoenix will join the ranks of the disappointed.
– DJ
Our research on the NBA was summarized HERE.
Wins Produced and Win Score are Discussed in the Following Posts
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Wages of Wins Journal 
TK
October 21, 2006
Just asking: Is the rebounding stat based on percentage of boards per possession? Or just number of boards?
Because with Nash at the helm, Phoenix is pushing the pace and generating more possessions per game. (I’ve read it somewhere else, but I think that Phoenix generates 10 more possessions per game than the Pistons, for instance.)
So maybe Marion didn’t rebound at a vastly better rate since Nash’s arrival. Maybe he (and all other players on the court) just had more rebounds to grab! Looking at his stats, actually, it looks like if there were 12% more possessions per game, his offensive rebound stats reflect almost exactly that increase. But his defensive rebounding improved even if the possessions increase.
dberri
October 21, 2006
TK,
You are correct, Phoenix plays at a very fast tempo. And that would explain the increase in offensive rebounds. And you are also correct, that does not explain the increase in defensive rebounds.
By the way, Win Score is not adjusted for tempo. Wins Produced, though, is adjusted for the tempo the team plays.
joe
October 21, 2006
If wins produced is based on totals versus normal totals of a position, it seems to me that Diaw’s switch back to SF will cause his WP48 to rise. And once Stoudemire returns, it will mean that the Suns starting lineup will have 4 players with very above average WP48 numbers, which makes them my pick to win the NBA finals.
dberri
October 21, 2006
Joe,
You are also correct. If Diaw only plays small forward then his wins production increases. That assumes his statistical production stays the same.
Because of diminishing returns, adding a productive player like Stoudemire will reduce the productivity of other players. Still, that effect is small and one suspects with a healthy Stoudemire on the court the Suns should be a very good team.
Travis
October 22, 2006
Are there any metrics that evaluate a players performance when he is in a contract year, like Kurt Thomas will be this year? Analysts often conclude that players in the final year of their contract play better, for obvious reasons? Is there any truth to this rationale? I suspect that if a players contract is guarenteed for the upcoming season than they would have an incentive to play their very best in order to receive a lot of money.
dberri
October 22, 2006
Travis,
Anthony Krautmann and I published a paper this past summer in Economic Inquiry examining whether NBA players offer less productivity after signing a long term contract. We did not find much evidence that job security adversely impacted player performance.
Perhaps our next study should be a study of whether players step it up right before becoming a free agent.
dberri
October 22, 2006
Travis,
Okay, I went back to the data set Anthony Krautmann and I used in our study. Specifically, I looked at the relationship between player performance and whether or not a player was in the last year of his contract. The results indicate that the expected increase in performance is not in the data set. So it does not appear that players systematically perform better the year before becoming a free agent.
Ken
October 27, 2006
B.S.
Take a look at how the number of games played varies in a contract year vs. a non-contract year. Then tell us how # of games played fails to figure into player performance.