Economists like to think people are rational. It is expected that they won't make the same mistake over and over again. If something doesn't work, people will learn and try something else. And, eventually, the most efficient solution will be chosen to fix a problem.
Economists, though, are often disappointed. Nowhere is this more evident than in the NBA, where teams (and their managers) often pin their hopes on inefficient scorers. A classic example of an inefficient scorer is Allen Iverson, a player who currently ranks 7th in NBA history in career points per game but had a career shooting efficiency – from both the two-point and three-point range – that was decidedly below average. So although teams gave Iverson more than $150 million, his employers only received a career level of wins production that was less than average.
Occasionally, though, teams come to the realization that hitching their hopes to an inefficient scorer doesn't work, and they take rather drastic action: instead of trying to convince the player to change, they subtract the player entirely from the roster.
The most recent instance of this subtraction happened in Detroit. A few days before Christmas, the Detroit Pistons – a team that had only won five of their first 28 games – waived (or essentially fired) forward Josh Smith, who was in the midst of a four-year, $54 million contract. Because the Pistons waived Smith (and he was not claimed by any team), they still had to pay the remaining $26 million on his contract.
To the casual observer, this might have seemed somewhat insane. The Pistons have volunteered to pay a player $26 million to not work. In addition, when this move was made, Smith was leading the Pistons in field goal attempts and minutes per game. So it seemed clear the Pistons were choosing not to play – but continue to pay – one of their better offensive players. It was not a surprise to see people describe this move as "blowing it up and rebuilding."
It is true that, if one simply looks at scoring totals, it is easy to conclude that a double-figure scorer like Smith is helping his team. But scoring totals are partially determined by shot attempts, and as noted, Smith was very good at taking shots with the Pistons. But he was not necessarily good at making them. When we examine Smith's shooting efficiency, we see that Smith was well below average. An average NBA player makes 48.7% of his shots from two-point range and 35% of his shots from beyond the three-point arc. With the Pistons this season, Smith shot 40.7% from two-point range. And he only hit 24.3% of his three-point shots.
OK, but why take such a big step? Why not encourage Smith to shoot less? It would make sense: by allowing a more efficient scorer to take more shots, the Pistons probably would have improved.
The problem the Pistons would have faced is the problem of the incentives facing NBA players. The study of player evaluations in the NBA – whether it be where a player is drafted, free agent salaries, post-season awards, or the coaches allocation of minutes (examined in Stumbling on Wins) – all indicate that the more a player scores, the more likely the player will be considered a "good" player (and get paid accordingly).
Smith couldn't shoot but led the Pistons in both shot attempts and minutes. He was in the midst of a contract paying him more than $50 million. Why would Smith conclude that his poor shooting was a bad idea? And since Smith probably didn't believe his shooting was a problem, why would he agree to stop?
Thus the nuclear option: the Pistons chose to pay Smith $26 million to not play – or take shots – for Detroit.
And what happened?
First, Smith learned that his inefficient shooting was apparently not an issue. After clearing waivers, the Houston Rockets – a title contender in the NBA – signed Smith and initially inserted him in their starting line-up. Yes, Smith learned that even being cut from a very bad team does not necessarily lead everyone to think you are "not good".
As for the Pistons, the loss of Smith led to a significant improvement. Since this move, the Pistons have won 12 of their first 19 games without Smith taking any minutes or shots.
This is not the first time that this has happened in recent NBA history. The unequal trades of the following inefficient scorers are all examples of teams that apparently were subtracting more than they added:
- In December of 2013, the Toronto Raptors had only won seven times in 19 tries when they traded Rudy Gay to the Sacramento Kings in what was considered a one-sided deal. Before the trade, Gay was leading the Raptors in field goal attempts. But, like Smith, Gay was below average with respect to shooting efficiency (45.2% from two-point range, 33.6% from beyond the arc). The team won 66.1% of its remaining games.
- In February of 2011, the Denver Nuggets sent Carmelo Anthony to the New York Knicks(again, this was considered a one-sided deal). Prior to the trade, the Nuggets had won 56% of their games. Without their leading scorer, it was expected the Nuggets would struggle on offense. But like Smith and Gay, Anthony was a relatively inefficient scorer (47.4% from two-point range, 33.3% from beyond the arc). Without Anthony, the Nuggets became a much more efficient offensive team, winning 72% of its remaining games that season.
- In December of 2006, when the Philadelphia 76ers traded the aforementioned Allen Iverson to the Denver Nuggets (and like we saw with Gay and Anthony, this was not considered a good deal for the teamlosing its star). Prior to the trade, the Sixers were the worst team in the NBA with only five wins in 23 tries (with Iverson making only 44.4% of his two-point shots and 22.6% of his three-pointers). Across the final 59 games, though, the Sixers managed to win 30 games. Again, a big reason was losing a player who led the team in field goal attempts but was an inefficient scorer.
To understand why losing an inefficient scorer leads to more wins, we need to review what causes wins in the NBA. Dean Oliver notes that wins are determined by four factors: shooting efficiency, rebounds, turnovers, and free throws. In essence, wins in the NBA are determined by a team's ability to get and keep the ball (i.e. rebounds and turnovers) and the ability to turn those possessions into points (shooting efficiency and free throws).
But in the NBA, players are not evaluated primarily in terms of what statistically determines wins. In fact, it appears that player evaluation is not often based on statistical analysis, but rather visual observation of the players. Because the "eye-test" dominates, scorers – who are often the most visible players on the court – tend to dominate player evaluations. The consequence of this approach is that players have an incentive to focus on scoring and not on the factors that primarily determine wins.
And because players have a clear incentive to take shots, it is very hard for coaches to get players – especially those who are perceived to be "stars" – to stop behaving this way.
This often leaves teams little choice: they have to subtract the "star."
By the standards of the rational world of math, it feels wildly incorrect. However, once we understand the conflict between player incentives and the factors that determine wins, we can see why subtraction is often addition in the NBA.