Premier League power rankings through 10 Matches

We've played ten matches. My projections are still based significantly on previous season data. They're currently about 55% previous season data, 40% this season and 5% regression to the mean. but I think it might be fun to look at just the underlying numbers through ten matches. Who's been good and who's been bad? Spoiler: We're good, Fulham are bad.

The basic components of the ratings are expected goals scored, expected goals allowed and strength of schedule. The expected goals scored and goals allowed numbers are based primarily on shots on target inside and outside of the box, with adjustments for big chances created and allowed. The methodology is linked below.

I was inspired to put all this data together because Ted Knutson at Stats Bomb posted his power rankings through ten matches, and I found that mine appeared to mostly but not entirely agree. I definitely like Aston Villa a bit more than he does, and I like Fulham a lot less. I've got the underlying numbers to explain which clubs I think are best on both attack and defense, so I figured why not share it. Ten matches seems like a good point to data dump. Enjoy.

Power Rankings and Att/Def Ratings

• Expected Goals methodology
• The columns list expected goals scored and allowed, strength of schedule, then overall power rankings in Att+, Def- and Team+. The latter four numbers are scaled to 100 as league average, higher is better for Att and Team, lower is better for Def-. A higher strength of schedule means the team has played tougher than average opponents and vice versa.

• The big strength of schedule effects are for Aston Villa and Norwich City. Villa and Norwich have played very difficult schedules so far. Villa I have remarked on before. With Norwich, I hope Chris Hughton doesn't get sacked before the club have a chance to make up ground against decidedly lesser opposition. (Home to West Ham, away to Newcastle, home to Crystal Palace in their next three).
• Spurs and Southampton are fortunate to have allowed only nine goals combined, but they do have the two best defensive ratings in the league through ten games. Neither side has been particularly strong in the attack, but a goal prevented is equally valuable to a goal scored in the abstract, so they stand in good position in the ratings
• Swansea City are probably the biggest outlier here. They were really good early in the season but didn't collect expected points. They were unlucky to lose home to United in a match where the two sides created chances of roughly equal quality, for instance. But in the last couple weeks they've been falling off, producing little in the attacking third against Cardiff City and West Ham. I could hypothesize Europa fatigue here, but I don't really have any idea. Maybe they'll start scoring goals and winning, maybe they were just on a random hot streak early, who knows.
• And Manchester United have been playing as poorly as their results, maybe even a little worse. They are simply not playing defense like a side that will be competitive for the top four, let alone for the title. Liverpool are interesting in this regard as they have been allowing far too many good chances despite relatively weak opposition, but their attack is so dynamic and creative that they can still get results.
• Since I'm still using previous season data for the projections, Manchester United remain ahead of Southampton and Swansea City there. If I went to only 2013-2014 data, they'd fall behind the underdog sides. I think that for now, I'm better off not jumping that aggressively on to the damned United train. I would still pick them to take more points than Southampton and Swansea over the next several months. I guess this is how far United seem to have fallen, that it's an argument whether they project better than Southampton.