Nobel Laureate Paul Samuelson once noted: “If you must forecast, forecast often.” As Samuelson stated, this quote is meant to emphasize the primacy of fact over theory. In other words, no matter what your theory might indicate, facts should rule the day. As the facts change, so should your forecasts of the future.
Samuelson’s words serve as a useful rule of thumb for economists who are prone to forecast. Economists, like those who predict the weather, often predict. Unfortunately, we are often wrong. In other words, the facts contradict our predictions. Nevertheless, armed with new facts we are happy to try and try again.
Given this record, we are immensely pleased when we manage to get something right. In The Wages of Wins we made the following prediction (p. 211):
Carmelo Anthony’s expected annual salary exceeds $13 million. When we look at Anthony’s points scored per game, we see why his salary in the future should be so high. If we look at Wins Produced and WP48, though, we again see evidence that the wages of wins in the NBA are not entirely consistent. Although Anthony in his first two seasons has only produced about one victory per year, his future salary is predicted to easily pass $10 million per season.
Okay, we were not exactly right. Anthony is going to receive a reported $80 million over five seasons, so our forecast is a bit off. We have a good explanation for this error. We haven’t forecasted Anthony’s salary often enough. Specifically, the forecast in the book was only based on Carmelo’s first two seasons (the book was mostly written in the summer of 2005). In Anthony’s first two seasons he averaged about 21 points per game. This last season, though, Anthony became a more prolific scorer, averaging 26.5 points per contest. Given the change in Anthony’s performance, let’s follow Samuelson’s advice and forecast again with our new facts.
Going back to our model with Anthony’s 2005-06 performance we now see a forecasted per-year salary of $16.9 million. The maximum contract Anthony is going to sign will pay him a reported $16 million per year, so our new forecast looks pretty good.
So we got one right. Alas, our forecast also indicates that the Nuggets may have gotten this one wrong.
It is important to note how we forecast NBA salaries. In The Wages of Wins we discussed several approaches to predicting how much money an NBA free agent would receive. The model that works best includes a player’s scoring, blocked shots, age, and position played. Additional stats (i.e. rebounds, turnovers, shooting efficiency) are not statistically significant and do not improve our ability to forecast. As we argue in the book, scoring dominates player evaluation in the NBA. And Anthony’s maximum contract is one more piece of evidence supporting our argument.
Unfortunately for the Nuggets, there is more to winning than scoring points. And this was clearly stated in our book and led to yet another prediction (p. 212).
Of course now that we have presented our evidence and made these statements, we should expect our predictions to be incorrect. With the publication of our book and subsequent publicity and fame that will obviously follow—yes, we are approximating humor again—we can expect people in the NBA to recognize the error of their ways. By the time the rookie contracts of these players expire the NBA will understand that points production does not necessarily equal wins production. So players like Carmelo Anthony will be paid less—sorry Carmelo—and Udonis Haslem and Josh Howard can expect much bigger paychecks.
Okay, this prediction was wrong. We clearly stated in the book that Anthony was not really worth a maximum contract. Yet that is what he’s signing. Clearly the decision-makers with the Nuggets didn’t read our book fast enough.
Where did Denver go wrong? Denver probably looked at Anthony and saw a player like Dwyane Wade and LeBron James. Anthony scores like Flash and King James. In fact, only Kobe Bryant, Allen Iverson, James, Anthony, and Wade managed to score 0.70 points per minute this past season. So Anthony is one of the NBA’s most prolific scorers. But being a prolific scorer does not guarantee prolific Wins Production.
Wade and James not only score, but also score efficiently and furthermore, make other contributions to their team’s success. Consequently Wade produced 18.2 wins this past season, while James produced 20.4. Not surprisingly, Wade and James are worth maximum contracts. In contrast, Anthony only produced 4.8 wins for the Denver Nuggets. And this level of wins production suggests that the Nuggets are paying Anthony just a bit too much.
Why does Anthony produce so little? Here is what I said back on May 30th.
Carmelo offers the type of production that we noted in the book and previously in this forum. Carmelo is a scorer, but not much else. He averaged 26.5 points per game, which is good. He went to the free throw line often and converted more than 80% of these shots. That is also good. From two-point range he made 51% of his shots. Again, that is good. From three point range he took 152 shots and made 37. That is a conversion rate of 24%. That would not be good. Given this performance from beyond the arc it is not clear why Anthony took nearly two such shots from three-point range per game. But he did and that lowers his value. When you factor in his three point shooting, Anthony’s points per field goal attempt was close to the league average. In other words, he accumulated high scoring totals but this is not because he offered high levels of shooting efficiency.
Beyond scoring, it is not clear what Anthony offers. As a small forward, on a per-minute basis, he was a bit below average in rebounding. He was about average in steals, assists, and blocks. With respect to turnovers, he was above average. Put it all together and what do you have? Carmelo is not LeBron James or Dwyane Wade, the two players he is often compared to from his draft class. James and Wade offer a team more than just scoring. Beyond scoring there was no other area that Anthony excelled. And as a scorer, he had one deficiency that reduced his effectiveness. Consequently, Carmelo did not offer many wins this past season.
Now does this mean Carmelo will never be a great player? No, it is merely an assessment of what he did in 2005-06, which actually was an improvement over what he did in 2004-05. Perhaps Carmelo will either learn to hit three point shots more consistently or stop taking so many shots from beyond the arc. Perhaps he will improve as a rebounder and cut down on his turnovers. Then again, perhaps not. Given what he did in 2005-06, though, it is not clear how his performance could be labeled All-NBA.
On last note on Anthony. The Nuggets did win 44 games this past season. If their leading scorer only produced 4.8 victories, where did the other wins come from? The answer can be found here, where the Wins Produced for each Denver player this past season is listed. The most productive players on the Nuggets were Marcus Camby (13.7 Wins Produced) and Andre Miller (10.4 Wins Produced).
Of course, maybe wins is not what Denver cares about. Carmelo is a “star” and he might be the one fans pay to see.
Okay, those who have read the book know this argument lacks much evidence. Specifically, we examined the determinants of gate revenue in the book and found that it is wins that drive gate revenue. Being a star matters, but only on the road. So Carmelo’s star power have value, but it really only produces revenue for the teams he plays. For the team who pays Carmelo, his star appeal has very little value.
– DJ
Wages of Wins Journal 
AJ
July 13, 2006
Okay, you’ve covered Melo, Flash and LBJ. But the other member of the ’03 elite, Chris Bosh, also considers himself a candidate for the max , and the Toronto Raptors seem to agree. But is he?
And I guess my larger question is: under your approach, could KG have been a 30 win player on the Raptors (last year)? Can even a great player’s teamates drag his WP48 score down?