Tom Haberstroh has ranked the “Feeble Five” in the history of each NBA franchise (insider access required). His criteria were as follows:
Picking a franchise’s five best players is a wide-open layup compared to identifying the most detrimental players of an organization’s history. But we’ve accomplished the feat with a little help from John Hollinger’s player efficiency rating (PER), which provides a handy productivity measure using a player’s box score statistics.
It’s pretty obvious that the worst NBA player is the one who never plays, but such a list would be no fun to read or write. We’re targeting guys you actually might remember but wish you couldn’t (Nikoloz Tskitishvili, anyone?). So, in order to qualify for this list, a player needed to have played at least 10 minutes per contest over the course of at least 100 career games with the franchise. Furthermore, we’ve added the “Bruce Bowen Corollary” to exempt players who started for championship teams.
Haberstroh’s reliance on PER (a model that rewards inefficient scoring) is odd given what he argued last August. At that time, Haberstroh said he did not believe Carmelo Anthony was a maximum player. Haberstroh’s reasoning was summarized in a column I wrote at the Huffington Post.
Haberstroh is arguing that scoring totals are a poor measure of player performance. A better approach is to consider shooting efficiency and the other factors that determine wins. And in considering these other factors, one needs to consider what one might expect to see from an average player at a player’s position.
One should note that Haberstroh never mentions Wins Produced (a measure of performance introduced in The Wages of Wins in 2006 and utilized in Stumbling on Wins) in constructing his argument. But this argument is essentially the Wins Produced story. When we look at how offensive and defensive efficiency relate to wins, we see that success in the NBA is driven by shooting efficiency and factors that gain and keep possession of the ball (i.e. rebounds, steals, and turnovers). Assists, blocked shots, and personal fouls (what Justin Wolfers and Joe Price refer to as “help” factors) do matter. But not as much as the possession factors.
Once we understand what drives wins, we also have to note that position played also matters. Big men tend to get rebounds and not commit many turnovers. Guards tend to do the opposite. Consequently, in evaluating players one has to consider what an average player in the same position would offer.
With the Carmelo Anthony argument in mind I wondered what would happen if we re-visited Haberstroh’s discussion of the Feeble Five. This time, though, we will ignore PER and focus instead on Wins Produced.
When we look at the Pistons (the team I follow), we see that PER ranks the following five as the most feeble in team history:
- Eric Montross
- Michael Curry
- Earl Lloyd
- Aaron Afflalo
- Chuck Noble
Lloyd and Noble played prior to 1977, or the first year all the data needed to calculate Wins Produced was tracked by the league. So our analysis with WP48 [Wins Produced per 48 minutes] can only consider Montross, Curry, and Afflalo [PER is a per-minute measure, so we need to consider WP48].
The metric considered will be the only change we will make. Players still have to play 100 games and average at least 10 minutes per contest. And given his criteria, the new Feeble Five for the Pistons are as follows:
- James Edwards [-0.088 WP48]
- Tony Campbell [-0.061 WP48]
- Clifford Robinson [-0.050 WP48]
- Paul Mokeski [-0.039 WP48]
- Eric Montross [-0.035 WP48
Michael Curry – with a -0.016 WP48 – finishes 6th. And Afflalo – with a 0.058 WP48 – finished 27th. There are only 63 players on the list, so Afflalo is fairly close to the middle of the pack. But since he doesn’t score, PER does not think very highly of his contributions.
In contrast, James Edwards could score. Given Haberstroh’s last criteria (i.e. can’t be feeble if you started for a title team), Edwards shouldn’t be on the list. Furthermore, with a career PER of 14.7 [15.0 is average] – and a career points per 48 minutes of 25.2 – Edwards is not generally regarded as a player who is well below average. Wins Produced, though, tells a different story. For his career, Edwards produced -22.7 wins. In his best season – at the age of 25 – he only produced 3.4 wins with a 0.069 WP48 [with Indiana in 1980-81]. Yes, Edwards could score. But he didn’t do much else to help his team win.
Let me close by noting the top twenty feeble Pistons. As the following table reveals, a few familiar names rank among the least productive Pistons since 1977. Allan Houston, Lindsey Hunter, Kelly Tripucka, and John Long were considered to be “star” players for the Pistons when they played. This quartet, though, only produced 41.5 wins in 30 year of action.
Allan Houston’s performance is especially interesting. Houston could score. And as a consequence, he was paid more than $100 million in his career. But he didn’t produce wins. Now there is talk he might someday be the general manager with the Knicks. Given his career, what sort of players do you expect Houston to choose if he ever become a general manager?