I've been in my hotel room crunching the last set of numbers and generating plots for my poster presentation at NCSSORS tomorrow, when I came across a most interesting piece of information from the Pythagorean table for the 2009 MLS regular season. (And by the way, there are a number of errors in the large results table on the MLS Wikipedia page, so beware. I'll point them out later.)
Most of the projected Pythagoreans were within that 3-5 point range with the exception of one team:
Real Salt Lake — (Projected) W13 D8 L 9 Pts47 with a differential of -7.
Much was written about RSL coming out from nowhere to win MLS Cup last season, and how they have used that momentum to play at a consistently high level in the league this season. (Recall that RSL did not have a single player on the MLS Best XI — the only time in league history that this has happened to a champion side.) One could look at the Pythagorean projection and claim that, if anything, RSL underperformed last season but managed to catch fire at the right time.
Another feature of the Pythagorean is that it has some utility in projecting team performance in the following season. The top five teams from last season, in terms of their Pythagorean point totals, were Houston Dynamo, Seattle Sounders, RSL, Columbus Crew, and LA Galaxy. Okay, maybe not in the case of Dynamo.
I'll reveal more interesting features on the other side of the NCSSORS conference.
UPDATE: As scaryice mentions in the comments, RSL had a 6-0 result in one of their league matches, which accounted for the +8 goal differential. If you remove that result and recalculate the Pythagorean from the other 29 games, you get:
Real Salt Lake — (Projected) W11 D8 L 10 Pts41 with a differential of -4.
which is a result that puts RSL's underperformance in line with Colorado and FC Dallas. The outlying result does have an effect on the Pythagorean projection, but not as much as one might think.
It should also be noted that RSL and FC Dallas had the two highest variances in goalscoring last season. (FCD has the lowest variance — offensive and defensive — in the 2010 MLS season, at least before their loss to RSL last Saturday.)