Wednesday, April 22, 2009

Pay and Performance in MLB - Part II

Last July I wrote a piece on pay and performance in MLB, and linked to a really neat Java applet by Ben Fry. As I am writing this he does not have the 2009 season started, but hopefully he will soon. You can look at the 2008 season (and the 2005, 2006 & 2007) seasons in terms of team performance and team payroll. As I said previously the way to interpret the lines is that, "[t]he lines in blue are for teams that perform better than their rank in payroll and the lines in red are teams that do not. While not a definitive analysis, it seems to show about half the teams have blue lines and half have red lines, which is a crude way of seeing correlation. Remember, correlation is where two variables move together." Now some have argued that we should use correlation as a way of measuring how well payroll explains performance. I have recently argued that this is wrong using the NBA.

Now, USA Today has announced that 14 of the 30 Major League Baseball teams have reduced their payroll. So since the correlation between payroll and performance is positive and if correlation is really the way to measure the link between MLB payroll and MLB performance, then these 14 teams will perform worse than in 2008. Since the correlation is positive, higher payroll is positively related to higher performance, thus it has to be the case that since correlation is positive and if correlation is the way to measure the relationship between payroll and performance that lower MLB payroll is related to lower MLB performance. You cannot make the argument that this is wrong, since it is the same argument some people make why higher payrolls result in higher performance.

Does anyone think that the Atlanta Braves, Baltimore Orioles, Boston Red Sox, Chicago White Sox, Cincinnati Reds, Detroit Tigers, Los Angeles Angels, Los Angeles Dodgers, Milwaukee Brewers, New York Yankees, San Diego Padres, Seattle Mariners, St. Louis Cardinals, and Toronto Blue Jays will all perform worse this year than last year? I sure don't.

If you are still not convinced, suppose all MLB teams payroll decreased from one year to the next. If all teams payroll decreased would all teams performance decrease? It cannot since the average performance for the entire league will still be 0.500. It is thought experiments like this that allow us to see why even if team payroll and team performance are positively correlated, they have little to do with each other in a statistical sense.

1 comment:

SEO Hyderabad said...
This comment has been removed by the author.