MLS Power Ratings: June 1, 2020

Most MLS teams have seven games – just over 20% of the season – under their belts, and we’re beginning to build a solid base of data for what this all means, man. It is time for the Power Ratings to make their glorious return.

Methodology Note

If you’re not interested in the math, or in checking to make sure there’s some logical consistency with the ratings (that’s putting a lot of trust in me, dude), feel free to skip this section. But the annual remainder of how this works follows:

Teams are what their résumés say they are. For that purpose, an average of gamescores home and away (and then the home and away numbers simply averaged together) form the basis of the power ratings. If you’ve played three games, the average performance across those three games is what this rating system believes of you.

What are gamescores, you ask? They are xG performance (as published by American Soccer Analysis) compared against the “average” opponent of each team you’ve played – a normalized distribution. If you accrue two expected goals at FC Cincinnati, that’s a below-average performance (the average team gets 2.12 xG in the Queen City), whereas if you were to do the same at Seattle, it’d be an outstanding performance (the average team gets just 0.93 xG there). Because of the wide disparity in home and road performances in the league, they’re considered separately.

“Luck” measures the difference between teams’ xG performances and their performances in strict goals scored terms – you know, the ones that actually matter. Performances in the goal department are descriptive, whereas xG is more predictive (backward-looking versus forward-looking). Own-goals aren’t included only because they’re not in the ASA data, but probably given my methodology I should add them back in.

The Ratings

I’ve previously published this only in an image, but for ease of consumption, I’ll do it in print, as well. What the above all boils down to: 0.00 is a bang-average team, +1.00 would be consistent performances a full standard deviation better than average, while -1.00 is terrible. Below -1.00 is territory only one team knows.

1. Nashville SC (+0.88)
2. Seattle Sounders (+0.87)
3. Sporting Kansas City (+0.77)
4. New York City FC (+0.68)
5. Los Angeles FC (+0.50)
6. DC United (+0.40)
7. San Jose Earthquakes (+0.34)
8. Atlanta United (+0.21)
9. Philadelphia Union (+0.12)
10. Chicago Fire (+0.10)
11. New England Revolution (+0.07)
12. New York Red Bulls (+0.04)
13. Colorado Rapids (+0.03)
14. Orlando City SC (-0.00)
15. Portland Timbers (-0.04)
16. Austin FC (-0.07)
17. Houston Dynamo (-0.30)
18. Minnesota United FC (-0.33)
19. Montreal Impact (-0.33)
20. Real Salt Lake (-0.36)
21. Toronto FC (-0.40)
22. FC Dallas (-0.47)
23. Columbus Crew (-0.48)
24. Inter Miami CF (-0.53)
25. Los Angeles Galaxy (-0.61)
26. Vancouver Whitecaps FC (-0.86)
27. Fußball Club Cincinnati (-1.36)

Das Analyse

I was hesitant to post this week, with Nashville jumping to No. 1 (the three teams behind them now were, in the same order, ahead of them before the weekend’s games). Ain’t no homers over here.

However, it does make some sense: Nashville has racked up incredible xG margins in a number of games (Cincinnati, Montreal, and Austin, in particular), and even come away with the xG edge in both road games. But the luck factor explains the results: Nashville has underachieved its xG performances this year (with nine goals scored on 12.6 xG, that probably doesn’t surprise you). If you look at the descriptive rather than the predictive, NSC is No. 13 in the league.

The other hyper-unlucky team is Chicago Fire, which… may be more likely to continue, since this team has been a massive underachiever compare to xG the last couple years, as well? (Nashville was somewhat lucky last year with largely the same roster).

Bigtime lucky guys include two Nashville SC opponents (probably not a surprise, given those luck margins had to build somewhere), as well as Orlando City, Columbus Crew, and Vancouver Whitecaps. Some of that is just because it’s hard for bad teams to be unlucky (there’s only so much more “down” out there), and some is small sample sizes skewing the data just a bit. Let us go to the chart:

There’s a fairly visible downward trend there, though without Cincinnati and Nashville it would be a little less pronounced (and again: should level off as more data rolls in).