On December 14 the Charlotte Bobcats sent Walter Herrmann and Primoz Brezec to the Detroit Pistons for Nazr Mohammed. At that time, none of these players was averaging more than 14 minutes per game for their respective teams. Consequently, this trade didn’t create much of a stir among NBA observers.
At the time of the trade the Bobcats record was 8-12. Mohammed missed the next game (a loss to Orlando). Over the next 20 games, though, Mohammed has averaged nearly 25 minutes a contest and the Bobcats record has been…. okay, 8-12. So clearly this trade hasn’t helped much.
The Bobcats Improve
If we delve into the numbers, though, we see a different story. After 20 games the Bobcats were being outscored by 5.6 points per contest. In seven different games their opponent won by more than 15 points. If we considered points scored and points surrendered, we would have expected the Bobcats to only have won 6.3 of their first 20 games, or 26 wins over a completed season.
Since Mohammed has joined the roster this team is still below average. But with Mohammed in the line-up the team has only been outscored by 1.7 points per game (and only once has the team lost by more than 15 points). If the Bobcats could maintain a -1.7 point differential across a full season it could expect to win about 36 games. In sum, the Bobcats are better after the trade.
Evaluating the Trade
So exactly why are the Bobcats better? Much of the answer can be seen in the numbers of the players involved in the December 14 trade. Table One reports the career averages of Brezec, Herrmann, and Mohammed prior to the 2007-08 season.
Table One: The Career Performance of Brezec, Herrmann, and Mohammed
Let’s start with the player this column is praising, Nazr Mohammed. He began his career during the lock-out shortened campaign of 1998-99. In Mohammed’s first nine seasons he has demonstrated that he is a slightly above average center. With respect to most statistics he hovers around the average mark, with slightly below average scoring marks offset by an above average performance on the boards.
Although Mohammed is little better than average, he is still quite a bit better than the two players the Bobcats traded away. Brezec began his career in 2001. Across his first six seasons he has shown he can be an above average scorer. But he is deficient with respect to rebounds, steals, and blocked shots. When we look at the total package, via Win Score, we see that Brezec is well below average.
The other player, Herrmann, began his career last year and clearly demonstrated the ability to score efficiently. But his advantage in scoring is offset by a level of rebounding that is well below average for a power forward (his primary position). Herrmann is also below average with respect to blocks and steals. Hence, like Brezec, his Win Score is below par.
Although both Brezec and Herrman were below average entering this season, what they did in Charlotte across the first 20 games of 2007-08 was quite a bit worse.
Table Two: Brezec, Herrmann, and Mohammed before the trade in 2007-08
Table Two indicates across almost all statistical categories these players were below average with Charlotte in 2007-08. In contrast, Mohammed was basically his same average self in Detroit this season. Consequently, this trade allowed the Bobcats to swap two sub-par performers for one average player.
By combining the statistics of Brezec and Herrman, Table Two highlights what this trade gave Charlotte. Together, Brezec and Herrmann were occupying more than 20 minutes per game. When we compare the performance of Brezec-Herrmann to Mohammed, we see the Bobcats came out ahead on everything but free throw percentage.
We can actually go beyond these tables and look at how many wins the Bobcats could have expected had Mohammed played the minutes of Brezec-Herrmann from the start. Brezec-Herrmann combined to post a -0.135 WP48 [Wins Produced per 48 minutes] across the first 20 contests. Meanwhile, Mohammed’s mark in Detroit was 0.124 (quite close to his career mark of 0.123). If the Bobcats had been able to take the 443 minutes allocated to Brezec-Herrmann and given these to Mohammed, the team would have improved by 2.4 wins across the first 20 games. This translates into 9.6 additional wins over an 82 game season. If you go back to the beginning of this column you will see that this mark is quite close to the improvement we see from the Bobcats in the 20 games after this trade.
When the Bobcats defeated the Celtics, Nuggets, and Magic in one week in January, people began to think this team had improved. And people argued that this improvement was tied to the play of Gerald Wallace and Jason Richardson, the leading scorers in Charlotte. Although it’s the case that Wallace and Richardson have played a bit better recently, I think the numbers tell us that it was the Mohammed trade that allowed Charlotte to become a bit more respectable.
More Comments on Charlotte and Detroit
Now it’s important to note that this trade doesn’t solve all of Charlotte’s problems. Yes Mohammed helps at the center position. And he is actually giving this team a level of productivity that surpasses what he did in Detroit (a level I am not sure will continue). But beyond Mohammed, the team is only getting above average performances from Wallace (when he plays small forward), Richardson, and Emeka Okafor. Every other player on the team is below average this season. So we can see this team still has to add a bit more productive talent if it wishes to consistently contend in the Eastern Conference.
What this trade has done for Detroit? Immediately after the trade the Pistons went on an eleven game winning streak. During that entire streak, Herrmann played 18 minutes. Across the last nine games, where the Pistons have gone 3-6, Herrmann has played 52 minutes.
Does this mean that Herrmann is responsible for Detroit’s troubles? No, the real problem can be seen in the recent play of Jason Maxiell, Rasheed Wallace, and Tayshaun Prince. The play of the two rookies, Rodney Stuckey and Aaron Afflalo, hasn’t helped much either. And I would add that Herrmann and Brezec have also struggled in Detroit just as they did in Charlotte.
In sum, the Mohammed trade hasn’t helped Detroit. But I think the team’s recent troubles can’t really be linked to Herrmann and Brezec. Again, these players are not part of the solution. But, they are not the problem in Detroit either.
– DJ
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
Wages of Wins Journal 
antonio
January 23, 2008
Don’t think its fair to label Herrmann as PF…
dberri
January 23, 2008
82games.com says Herrmann primarily played PF in Charlotte last year and this year. And when you look at the minutes on Charlotte, it also looks like Herrmann primarily was a PF.
Mountain
January 23, 2008
There is a lot going on in Charlotte- new players, playing small.
Mohammed is a good rebounder but since the trade Charlotte has slipped badly on offensive rebounds (-2.5 per game after trade vs before) and total rebounds (-5.5) and gave up more. Not simply Mohammed’s fault given the choice to play small. A more complicated minutes study would be required to fully understand what happened. If Charlotte expected the swap would add rebounds it didn’t. All things were not equal though. So I am not sure the ongoing debate about independence of rebounding gets new knowledge from this example. But maybe someone can work the data in a useful way.
Mountain
January 23, 2008
Available rebounds are affected by missed shots. So the study probably ought be cast as % of rebounds available.
Mountain
January 23, 2008
Mohammed’s rebounding per minute (total and offensive) are pretty constant in Detroit and Charlotte. But other Bobcats must have fallen. With one material exception every Bobcat rebounds less with Mohammed on the court with them than without. http://82games.com/0708/0708CHAP.HTM
Mountain
January 23, 2008
Team defensive rebounding is up with Mohammed on vs off so he helps on net. (I am seeing different things on offensive rebounding at 82games and basketballvalue.)
But given a 7 rebound per 48 difference between Mohammed and the average of Brezec and Herrman if everything was the same and unchanged over around 20 minutes a game it might have been predicted that Bobcat team rebounds would have increased by about 3 per game. The drop of -5.5 is mainly on Sam Vincent and the rest of the players but the idea that rebounds would transfer and be independent net adds to team rebounds does not get any help from this example.
Mountain
January 23, 2008
Unless a detailed minutes study somehow showed /asserted that Mohammed did add 3 rebounds but Vincent and the rest of players gave up 8.5 rebounds for reasons totally independent of Mohammed.
Mountain
January 23, 2008
A big part of the change is independent of Mohammed.
And part of it is him taking rebounds from others thru his efforts / stronger abilities than the guys he replaced. And part of the results may be that other Bobcats have changed their rebounding effort (by team direction or otherwise). This last theory is conjecture on my part but I have seen other teams were there is too much rebounding deference. It doesn’t have to happen but it sometimes does.
Jeff
January 23, 2008
With enough observations/games, gathering tedious info on in-game situations probably doesn’t offer much in the way of new information. Taken in aggregate, rebounding is dependent on missed shots. Having teammates or opponents that can’t shoot shouldn’t affect an individual’s rebounding performance though. Pretty much every regular player will face every other regular player over the course of a season. All of those game-specific dynamics can probably be safely treated as noise in an econometric model. Treating rebounding as explicitly dependent on missed shots is probably not going to do much more than make the model harder to estimate. If a player is consistently grabbing eight rebounds a game, that is likely better than only grabbing four a game, no matter the pace of game and number of total shots taken. Maybe I don’t know anything though: my dissertation research has nothing to do with sports; my econometric work is applied to land use change. I guess that is fundamentally different than basketball . . .
Daniel
January 24, 2008
Though Herrmann has inexplicably played PF in the NBA, nearly all his minutes on the Argentina national team come as a SF. Scola plays PF for Argentina, and Herrmann plays on the perimeter. It’s clear he’s a natural 3 watching him play. The reason the Cats traded him is because there wasn’t any room behind Gerald Wallace and Brezec is just terrible. Just because his coaches don’t know how to play him doesn’t mean he’s a PF.
I also think it’s unscientific to rely on this year’s stats so heavily when evaluating Herrmann as a player since he really hasn’t played much.
antonio
January 24, 2008
I agree. Thats what I mean regarding Herrmann. He may have been forced to play the 4, but clearly he is a 3 and almost any team for the future will use him at the 3