Thursday, May 8, 2014

Competitive Balance in Major League Soccer

Today let's look at competitive balance in Major League Soccer (MLS).  Major League Soccer re-started in 1996, but for the first few years (1996-1999), there were no ties during the regular season and since 2000 ties are possible during the regular season, so I am truncating the overall data starting with the 2000 regular season.  I will use the Noll-Scully measure of competitive balance to measure competitive balance in Major League Soccer.

The Noll-Scully measure captures how far a particular sport or league’s distribution of wins deviates from a purely random outcome of wins and losses.  Thus a Noll-Scully equal to one would be one that is purely random.  Since the 2000 season, the average Noll-Scully in MLS equals 1.294, which is more competitive than the National Hockey League, the National Football League and the National Basketball Association.

(If you are interested in replicating these results, I got the data from MLS website and here is a step-by-step method to calculate the Noll-Scully measure of competitive balance - where I used the trinomial method under a standard deviation of the population.  I also used the probability of a tie to be the actual yearly average of games that were tied).

Here is a chart of how competitively balanced Major League Soccer has been since 2000.



For those interested in replicating the numbers, here is a table of the results.

Noll-Scully
Season
1.180
2000
1.757
2001
0.719
2002
1.189
2003
0.844
2004
1.834
2005
0.874
2006
1.287
2007
0.943
2008
1.160
2009
1.553
2010
1.396
2011
1.752
2012
1.631
2013