No preamble this week, but plenty of regular amble down below. If you didn’t read my thoughts from last week, they still generally apply, so maybe check those out before we continue.

Let’s just say that despite following the rankings week in and week out on my own, I did not expect two of the monster moves listed below.


I’ve highlighted the major movers in green and red. The two moves that I absolutely did not expect to see there were Chelsea going from 3rd down to 7th, and Fulham going from nearly last to middle of the table. Through 10 games, teams are still somewhat volatile, and a couple of bad games in a row can still propel a team up or down the rankings. That said, we’re a quarter of the way through the season, so these rankings as they currently exist are pretty much how you’d expect them to play for the rest of the year unless major changes are made.

Analytics guys have been talking fairly openly about how bad Fulham are for some time now, and given where they were ranked five weeks ago, I wholeheartedly agreed with them.  The surface numbers still peg them as one of the worst teams in the league. Their TSR rank is… well, last. And okay, defensively they are probably the second or third worst team in the league.

The big problem with this team is that they simply have no control in midfield. Like, zero. Steve Sidwell was bad last year. Sidwell plus the shell of Scott Parker is dire. Hangeland and Riether at least are quite handy defenders, and their keeping situation is generally okay. But because everything flows straight from the attacking end back at Fulham’s defense at pace, they really struggle against most teams.

The thing that’s saving them is that offensively they still have a lot of talent. Berbatov and Ruiz could easily play for better teams, and Darren Bent – despite being a player who just is never that involved offensively – can still score goals. Systemically they are shit, but good offensive players can still save a team.

Why is this team so much worse than the one who was actually quite good for the first ten matches last season? No fucking clue. Under the surface this year, they are now performing a bit better than people expect, and it keeps showing up in the results. Confusing stuff.

How does a Mourinho team go from being ranked 3rd to 7th in the space of five weeks?

Step 1) Get dominated at home by Manchester City. (And City were clearly better that match, even though the Hart gaffe gave Chelsea the win).
Step 2) Get dominated on the road by Newcastle.

That’s it. Chelsea went from rolling teams during a fairly easy stretch of schedule, to looking a bit off the pace against two decent teams in a row. (Newcastle at home are decent. Newcastle on the road… maybe not.)

The big problem for this club is that, like Fulham, they also don’t have good control of the match in the center of the pitch. Mourinho’s system works best with a runner + destroyer and a defensive regista like Xabi Alonso involved. Right now he kind of has the runner + destroyer in Ramires (who probably runs more and destroys less, but is quite talented), but they have absolutely no one who can play the distribution + control role. I highlighted this issue in my season preview – “Can Chelsea win the title with Obi Mikel and the ghost of Michael Essien as their defensive midfield?” but the loss of van Ginkel removed yet another important option for that role.

It’s interesting to compare how AVB and Mourinho tweak what is basically the same system. AVB prefers to go with two defensive holding midfielders, and has been strangling his team’s creativity with his team selection to the point that they simply aren’t scoring. Mourinho doesn’t have the personnel at DM that AVB does, but at this point in his career is more likely to deploy more offensive-minded players anyway. This results in Chelsea having problems shutting down better clubs defensively.

Basically, Chelsea’s midfield is a collection of older, slightly mis-matched parts to go along with their aging, but generally solid defense. The Premier League might be too good this year to allow him to get away with this group of personnel and still win the title. Or Mou might figure out the right combination of parts to steam roll the last 28 games, including adding what they need in midfield in January.

It’s impossible to rule his genius out.

I think they will recover a bit in the rankings, and obviously they are still second in the table, but I also feel like they will continue to have problems against the better teams in the league. The fact that there are actually nine other teams who can give them problems makes this league campaign a lot more random than it normally would be.

Manchester United
In the link at the beginning of the article, I wrote about Moyes the Muggle, and how he’s taken a lot of the magic out of United’s offense. What I did not expect was that United’s defense would suffer so much in the process.

Here are the baseline defensive stats for United this year under Moyes:

  • Tied for 13th in Goals Against
  • 7th in Shots Conceded Per Game
  • 12th in Shots on Target Conceded Per Game

That’s a fairly significant jump in rate between shots conceded and SOT conceded, which means United have some major issues with defensive organization. I say I’m surprised by this because it’s one of the things he did really well at Everton.

While we’re here though, check out the real problem.


Elite teams do not give up shots on target at that rate. Neither do good teams. In fact, Opta’s Devin Pleuler and our own Colin Trainor have both noted that the shots United are conceding in the first 10 matches of the year have the highest likelihood of yielding opposing goals of anyone in the league. Compare that to what the same stat looked like under Fergie and you see the size of the problem.


The really odd thing is that Everton were only a small touch worse than United defensively last year, obviously with much cheaper personnel. Moyes knows defense, but this team shows no evidence of that so far.

So yeah, things are not at all right for United at the moment. To make matters worse, they face the league’s best team at creating chances this weekend. Things like this have already been happening to United this season.

Watch out.

Doom Patrol
If the model has its way, then Cardiff, Sunderland, and Crystal Palace are getting relegated, but Norwich are bad enough that they’re in the mix. Obviously the 7-0 hammering at Manchester City isn’t exactly going to help. Sunderland are showing signs of being less terrible than under Crazy Paolo, despite giving up 25 shots against Hull. I don’t think it will save them, but it’s enough to make Chris Hughton uncomfortable.

After them you have a very grindy set of teams that should be safe, but won’t exactly be exciting. This includes West Ham, who would probably be better if they had any semblance of a forward line. Nice Andy Carroll purchase there, Allardici.

Random Opinions

  • Chelsea will recover and could certainly win the league, but they need to reinforce their midfield in January. This was a fairly obvious problem before the season started. I can only assume van Ginkel was thought to be enough to solve it.
  • Spurs have the personnel to blow teams out. AVB still isn’t playing those personnel. Will he figure this stuff out before it starts to cost them points? They also could be historically good defensively.
  • United is into their easier run of games, yet their coefficient has dropped. This is me, on record, saying that I don’t think Manchester United will make the Champions’ League this season. *Cue Rocky theme and United ripping off ten wins in a row.*
  • Everton have simmered down after a hot start, and are playing at about the same level as they were last season. Kone and the Barry loan were such bad moves financially for a team on a strict budget, but Martinez seems like a good hire.
  • Swansea are good, but the Europa League eats small squad teams alive. It’s a shame, because they could really do some damage if they weren’t in that competition.
  • As Grayson points out here, Aston Villa have had one of the toughest schedules in the league, and Benteke was injured for a while. Expect them to start churning out results again.
  • Liverpool’s numbers have been creeping up since Suarez returned. They may continue to do so with Coutinho back now as well. They still have midfield issues, but I think FSG will be willing to make a big transfer or two if the team is still in good shape at Christmas time.
  • 5 points clear of Chelsea. 6 points clear of City. In the top 3 of the model rankings… this is the first time all year I have been willing to admit that, yeah okay, maybe Arsenal actually could win the title this year. Gulp.


  • 47ruedroite

    Great work Ted, nicely done, as always, keep it up.

  • Phil

    Would like to have seen more talk about City in terms of their prospects for the year, unless you’ve already covered that in previous weeks, nothing has changed for them and they move up because THFC moves down under AVB’s ineptitude…

  • MCofA

    Cool stuff. Do you know how Fulham improved so much in the model? I rated their performance against United quite highly, but I still have their 10g numbers in the toilet.

    My ratings through 10 games (different from my power rankings numbers which include also previous season data. I’ve also listed a Team+ rating (how good the team is compared to a theoretical average team, scaled to 100). I’ve bolded what look like our two primary disagreements.

    1) 163 MCI (1st)
    2) 142 ARS (3rd)
    3) 137 TOT (2nd)
    4) 128 CHE (7th)
    5) 126 SWA (4th)
    6) 123 LIV (8th)
    7) 122 SOT (5th)
    8) 121 EVE (6th)
    9) 108 MUN (9th)
    10) 106 AVL (13th)
    11) 100 NEW (10th)
    12) 87 STO (12th)
    13) 82 WBA (14th)
    14) 79 WHM (16th)
    15) 74 HUL (15th)
    16) 69 NOR (17th)
    17) 62 CAR (20th)
    18) 58 FUL (11th)
    19) 56 SUN (19th)
    20) 56 CRY (18th)

    Everything else is close enough to basically confirm we’re seeing similar stuff, but I have Fulham as still quite terrible, and I rate Aston Villa as roughly average so far because of the strength of their opposition.

  • Rich

    Fulham against Utd will be deceptive. we woke up when 3-0 down which just gives toilet numbers because United were seeing the game out while we huffed and puffed.

    • MCofA

      Fulham produced a bunch of good chances against United – 9 shots inside the box, 7 from close/central positions, 5 shots on target inside the box, 4 from close/central positions. If Berbatov in particular had his finishing boots on, it could very easily have ended 3-3. United were sitting back, but the purpose of sitting back is to force the opposition into long-range pot shots, not to hope the opposition striker has an off day.

      I have Fulham’s overall numbers still terrible, but their performance against United was perhaps the club’s best of the season.

  • KC_Gunner

    Ted — good post as usual. However, I have strong doubts about Tottenham’s ranking at #2. Do you really feel confident the model is correctly evaluating them? I asked you on Twitter whether any Top 4 (or Top 6) side had ever gone through a 10-game stretch of scoring 9 or fewer goals. I don’t think I have the data to look that up, but I decided to look at the goals per game of the Top 4/Top 6 (and relegation) over the last 4 seasons.

    Spurs have scored 9 goals in 10 games. For the last 4 seasons (2009-10 through 2012-13), the average of the goals scored by the 4 champions was 90 goals, or 2.36 g/pg. Among the 16 top 4 finishers, the average goals scored was 77.4 goals, or 2.03 g/pg; the average of the top 6 finishers over the same period was 71.67, or a 1.89 g/pg clip.

    For Tottenham to reach 77 goals, they need to score 68 goals over the next 28 games, or a 2.43 g/pg clip — better than the average champion has done over 38 games. To reach 71, they’d need to score 2.21 g/pg.

    Meanwhile, the 12 relegated teams over this period had an average goals scored of 41.58, or 1.09 g/pg — better than Spurs have managed through 10 games this year.

    Granted, Spurs have had very good defensive performances good so far, but a team has to score goals to be in contention to finish top 4 / top 6. And sure, 10 games may not quite be a meaningful enough sample size. And probably measures other than goals scored will produce a higher correlation re finishing position, but this was a quick & easy thing to chart & tally up. From the results, it’s hard to believe Tottenham are really #2 in the league when they are scoring goals at a lower per game rate than the last 12 relegated sides.

  • Toshack

    Good stuff Ted!
    So Everton is slowly sliding downwards.
    Due to a grim upcoming fixture list Liverpool will be on 6th – 8th spot by the half was mark of the league (was my prediction before I read this – yeah really!). Even though I hope otherwise.
    Playing vs. a lightning fast Arsenal midfield is perhaps not the correct “measure” on Liverpool’s centre midfield, but if Rodgers want to play a more fluid, and fast, game in the future he needs to rebuild the centre or the park. Soon.
    And perhaps considering a solid left full/wing back? First Newcaste, now Arsenal. Just giving the goals away due to lax attitude and lack of concentration.

  • Derek

    interesting, always love looking at others models to compare mine. Don’t quite understand Chelsea falling to 7th especially the comment about them being dominated by City and Newcastle. they had higher Expected Goals in each game. they took more shots from areas where there is a 10%+ shooting percentage in both games (6-4 in City game, 8-5 vs Newcastle). they weren’t far and away the better team in either game but certainly weren’t dominated, I’m interested in what metrics showed they were.

    Fulham jumped in my model, also finally off the bottom to 19th. 11th seems strangely high for a team with essentially 70 minutes of good chances when down 0-3 vs ManU.

    I got
    tier 1-City

    tier 2-Chelsea, Arsenal

    tier 3-Liverpool, Tottenham, Southampton

    tier 4-West Ham, Everton, Swansea, Stoke, Man U, West Brom, Newcastle

    tier 5-Aston Villa, Hull, Norwich, Palace

    tier 6-Sunderland, Fulham, Cardiff

Improve Performance and Productivity in Your Club:
State-of-the-art Football Analytics