Adjusting for Position Played

Posted on July 9, 2007 by


Last week I posted a comment on the Orlando Magic’s signing of Rashard Lewis. This comment noted that Lewis might be worth his hefty contract (he signed for the league maximum) if he plays small forward. But if the Magic play Lewis at power forward – which they might be inclined to do if they truly pass on Darko Milicic – then the production offered by Lewis will be worth far less than the wage he is paid. In sum, the Magic can do much to justify the wage Lewis is paid by simply keeping Lewis at small forward.

Yglesias Disagrees

This analysis led to the following statement from Matthew Yglesias:

This is more than a little perverse. Good power forwards are hard to come by. That Rashard Lewis is capable of performing competently in that role is an asset he has as a player. But thanks to the WoW position-adjustment method, it registers as a problem for his game. If he was much, much worse at playing the 4, he’d never be asked to do it, and his WoW rating would look much better. But in the real world, he’d be a less valuable basketball player.

Where to start with this statement?

1. As I will demonstrate in a moment, Lewis is not a “good” power forward.

2. Lewis is capable of being an average power forward. But Lewis is being paid the league maximum. For a player to collect this kind of money and only be “average” suggests a problem.

3. The idea that Lewis would not be asked to play power forward if he is much worse at playing this position runs counter too much of the evidence we present about basketball in The Wages of Wins. Just a quick counter-example: Eddy Curry is not a competent center yet he is asked every night to play this position.

Beyond these points, Yglesias appears to be arguing that Lewis should be given extra credit for being able to play two positions.

Back to the Numbers

Let’s go back to the numbers and see the problem with the Yglesias “extra-credit” argument. The following table reports a side-by-side comparison of Lewis to both an average NBA small and power forward.

Table One: Comparing Lewis

Relative to the average small forward, Lewis is above average with respect points scored, rebounds, turnovers (he has less than an average small forward), and shooting efficiency. Given these marks, it’s not surprising that Rashard’s WP48 – when he is compared to a small forward — is well above average.

When he shifts to power forward, though, he should now be compared to the average at this position. Again he is above average with respect to points scored, turnovers, and shooting efficiency. But his ability to rebound is well below the mark we would expect from an average power forward. Consequently, when Lewis plays power forward, his lack of rebounding hurts the team.

An average NBA team captures nearly 42 rebounds per game. Nearly 24 of these, or 57%, come from power forwards and centers. When you put a player at this position who rebounds at a below average rate, a team’s chances to win are diminished.

Yglesias, though, wishes to give credit to Lewis for at least making the effort to play power forward. This strikes me as very much the same argument people offer in defense of Allen Iverson. The Answer takes a large number of shots, many of which miss the mark. People argued, though, that Iverson should get credit for “creating his own shot” or making the effort to take these shots. Following this argument, it must be the case that Wins Produced minimizes the “true” value of Iverson since players are not given credit in calculating Wins Produced for just taking shots. For a player to create wins, shots taken must actually go in the basket.

The validity of the Iverson argument was undermined by what we observed last season. The 76ers traded Iverson to the Denver Nuggets for Andre Miller. When we look at shot attempts on this team both before and after the trade we see little difference. In other words, Iverson didn’t so much “create shots.” Rather, he simply took shots his teammates were more than capable of taking. Furthermore, we saw before the trade that Miller had consistently been more productive than Iverson in terms of Wins Produced. Consequently, despite warnings that the 76ers were going to suffer without The Answer, I argued the team should dramatically improve. And this is indeed what happened.

Now we have the argument that the value of Lewis should not depend upon position played. The numbers tell us that playing Lewis at power forward will cost Orlando rebounds. But we should ignore this fact and simply give Lewis extra credit for making an effort.

Unfortunately, the numbers tell us a different story. If you play Lewis at power forward his team will suffer with respect to rebounds. And when rebounds decline, teams tend to win less.

In sum, position played matters in the evaluation of NBA talent. Centers and power forwards offer a different set of statistics than small forwards. Small forwards are different from guards. Given these differences, how much a player is helping a team win depends on the position the player is playing. To evaluate a player’s contribution these position differences have to be considered. If not, the contribution a player makes to team wins will be estimated incorrectly.

Metrics Without Position Adjustments

Although I should stop writing today and get to my “real” work, I want to make one more point about position adjustments. Metrics like NBA Efficiency (which is similar to TENDEX and Points Created) and John Hollinger’s Player Efficiency Rating (PER) do not have a position adjustment. This is because scoring dominates these measures (there is a strong correlation between a player’s NBA Efficiency and his scoring totals). And because scoring rates do not very as much by position, one does not have to adjust for position played when looking at NBA Efficiency. A similar argument applies to PERs.

Unfortunately, as noted last fall, both PERs and NBA Efficiency have a problem with inefficient scoring. Basically, a player can score inefficiently but raise his PERs and NBA Efficiency marks by simply taking more shots. Having inefficient scorers taking more shots does not help a team win, hence these metrics do not always evaluate a player’s contribution to team success correctly. Specifically, these measures tend to lead people to think that players like Allen Iverson and Carmelo Anthony are truly great players. In reality, though, both Iverson and Anthony are closer to the league average in player productivity.

Back to Lewis

Average is also how one would describe Rashard Lewis when he plays power forward. Just like it is not good enough to simply take shots in the NBA, it’s also not helping much to play a position as well as an average player. At least, if you are receiving the maximum salary, people should expect more from you than simply average production.

Basically, there is no “free lunch” in the NBA. You do not get extra credit for taking shots that don’t go in. You also don’t get credit on the scoreboard for playing out of position. At the end of the day, you win because your numbers beat your opponent’s numbers. And if the Magic do end up giving Lewis significant time at power forward, Orlando is going to find that their numbers – and their team wins – are going to suffer.

– DJ