Thoughts on a Sports Illustrated Column and Bloomberg Television

Posted on February 1, 2007 by


On Tuesday Ian Thomsen of Sports penned “To trade or not to trade?” In this column Thomsen explores why more trades do not happen in the NBA. Although he makes a few interesting points, he does offer one paragraph that I think is inconsistent with the empirical evidence:

What’s behind this lack of activity is a league-wide parity approaching that of the NFL’s, caused mainly by two things: the absence of franchise cornerstones due to suspension or injury, as well as the financial discipline imposed by the luxury tax, which has forced all but a few teams to spend virtually the same amount of money on players.

There are two problems with the above statement.

The NBA does not have — and it never has — a level of competitive balance similar to the NFL.

Thomsen argues that parity – or competitive balance – in the NBA is approaching the level we see in the NFL. In The Wages of Wins we discuss the Noll-Scully measure of competitive balance. This metric compares the standard deviation of winning percentage to the standard deviation that would exist if all teams were equal in playing strength. In the book we note that the average level of the Noll-Scully for 15 different sports leagues is 1.86. In other words, the actual standard deviation tends to be 1.86 greater than the idealized in the professional sports leagues we examined.

Okay, with the number 1.86 in our brain, let’s look at the NFL. The last time the NFL had a Noll-Scully that was greater than 1.86 was 1944. Basically the NFL has always been one of the most competitive sports leagues, with a historical average of 1.56. This past season the ratio of actual to ideal stood at 1.45 for the NFL (remember, a lower Noll-Scully means more parity or competitive balance).

Now consider the same metric for the NBA. The last time the NBA had a Noll-Scully below 1.86 was the 1976-77 season. The NBA has a historical average of 2.54 and in 2005-06 had a ratio of 2.47. If we look at the standings this year, we see a standard deviation of winning percentage that is 2.46 times the ideal.

In sum, the NBA – the league with historically the lowest level of competitive balance – cannot be compared to the NFL – the North American league with typically the highest level of competitive balance.

There is not much of a relationship between how much a team spends and their winning percentage.

The presumption Thomsen makes is that the NBA has parity because all teams spend basically the same. Clearly if teams spend the same, and teams “know” which players are best (or worst), then all teams would be roughly equal. We have already seen that teams are not all the same in the NBA and it is also not the case that NBA teams spend the same amount on their players.

Furthermore, there is evidence that NBA teams may not “know” best. In The Wages of Wins we report how much of a team’s winning percentage is explained by the team’s relative payroll (team payroll relative to the average in the league). We found – in an examination of teams from the 1990-91 season through the 2004-05 campaign – that only 12% of wins was explained by relative payroll. When we add the 2005-06 season to the sample, the explanatory power of relative payroll drops to 11%. And when we look at just the 2006-07 season, we fail to find a statistical relationship between relative payroll and the team’s current winning percentage.

This point can also be made if we look at the determinants of individual free agent salaries. We found that how many points a player scores explains much of the wage a free agent is paid. Factors like shooting efficiency and turnovers are not found to matter. Because scoring is over-valued in the NBA, the link between wins and pay is not as strong as one might believe.

I would add that other than that one paragraph, the Thomsen column is a good read. He basically accuses many general managers of not having the courage to pull off a trade. So that makes for interesting reading.

Bloomberg Television

Tomorrow I am scheduled to appear for about 10 minutes on Bloomberg Television. The interview is scheduled to take place a bit after 2:30pm EST (11:30am PST). The topic: The Super Bowl (and The Wages of Wins).

– DJ