clock menu more-arrow no yes

Filed under:

Beyond the table: Expected goals vs. actual goals, schedule strength, top six breakdown

New, comments

Taking stock of the Premier League after Matchweek 31

Chelsea v Tottenham Hotspur - Premier League Photo by Catherine Ivill/Getty Images

So how’s that Premier League looking right now? All depends on which club you support, of course. But the show must go on, regardless.

Let’s take a brief look at how the league and the Top 6 are doing, although it’s more and more difficult to actually call it a Top 6 as the season grinds to its eventual outcome.


First things first, here’s how the table looks visually, with lines for CL Qualification and Relegation:

As you can be seen:

  • City have continued their constant runaway-with-the-league performances
  • The next three are in a very tight race to see who has more bragging rights next season
  • Chelsea and Arsenal are slowly drifting towards that bottom-of-the-top-half table cluster (sigh)
  • That relegation line is at 29 points right now, under goes down and over is safe. But notice even Bournemouth in 10th place are only at 37 points, or three losses from Relegation-ville. It’s incredible how close 10-19 are right now in the Premier League.
  • The last several seasons, and how high 8 points from safety got:
    2016-17 - 35 = safety, 13th place Stoke at 44 points.
    2015-16 - 38 = safety, 12th place Swansea at 47 points
    2014-15 - 36 = safety, 13th place West Brom at 44 points
    2013-14 - 34 = safety, 12th place Swansea at 42 points
    2012-13 - 37 = safety, 10th place West Ham at 46 points
    2011-12 - 37 = safety, 14th place Stoke at 45 points
    2010-11 - 40 = safety, 9th place Aston Villa at 48 points


Back after Matchweek 26, I set up a table of the matches to come, and my predictions (based on pure gut instinct, not any sort of complex algorithm) for how the Top Six would do for the rest of the season. It was a pivotal point, as we still had a one point lead for 4th place, and were on the inside looking out, but still had matches at United, at City, home to Spurs and home to Liverpool yet to come.

In fact, the Top Six all had quite a few matches against each other left. I identified these Top 6 matches, as well as Bottom 9 matches, and at the time what I considered to be “Trap Matches” meaning more dangerous than the table position implies.

As the matchweeks have gone on, I’ve indicated the wins and losses (surprisingly no draws in the last five matchweeks for any of the Top 6? Odd anomaly) as well as how the actual results plus my predictions-still-to-come play out for the season.

We all know it’s quite unlikely at this point that we are going to nab that Fourth Place slot, given how the match over the weekend went. And even if we win out, I still think it highly unlikely, as we are simply so far back, and both Liverpool and Spurs are playing on great form right now.


I’ve found a site I really like for sourcing “expected” data, even though it’s a black box, simply because I don’t have raw data feed access myself, so don’t really have any good options at building a useful ‘expected’ model myself right now, and this site publishes all of their results. It’s, and that’s where I’m getting the data for the next couple of sections.

I wanted to see how the clubs in the Premier League are performing with respect to Goals For and Goals Against, and with respect to what their “expected” number is. In the chart below, I label each quadrant, but here’s a bit of explanation:

  • Underrated Attack - This means they are scoring more than they are “expected” to score.
  • Underrated Defense - They are giving up fewer goals than they are “expected to give up.
  • Overrated Attack - They are scoring less than they are “expected” to score.
  • Overrated Defense - They are giving up more goals than they are “expected” to give up.

Most “expected” or “x” models are similar enough to be confident in Understat’s values.

Some observations:

  • Very few teams are giving up fewer goals than expected. Five, in fact. The last two don’t surprise me, but the other three do, quite a bit.
  1. Newcastle
  2. Swansea
  3. Bournemouth
  4. Burnley
  5. Manchester United
  • How sad is it that Chelsea are actually scoring more goals than they are expected to?
  • If West Ham could only play defense they’d probably be in the Top Half of the table.
  • Crystal Palace are so bad offensively I accidentally clipped them completely out of the chart. Their numbers: xG Delta = (15.51), xGA Delta = 2.58. Yeah, they are over 15 goals UNDER their expected value. Shocking.
  • Bournemouth are somehow overperforming massively on both ends. Imagine where they’d be otherwise.
  • Same with Manchester United. Almost 8 goals over expected For, and over 12 Goals under expected Against.
  • Poor, poor Arsenal. Their Death Row Box:
    20th West Brom
    19th Stoke
    18th Southampton
    17th Crystal Palace (remember completely off the chart lol)
    16th Huddersfield
    13th Brighton
    And 6th Arsenal. lol. In our sad, sad, season at least I can still laugh at Arsenal.


I took the xG (Goal), xA (Assist) and actual numbers for the players in the Premier League, and selected the Top 50 based on xG+xA values, and plotted them against their actuals. Additionally, I color coded the player names for each of the Top 6 clubs, and added two lines. One diagonal showing a flat ratio, meaning actual G+A matched xG+xA exactly, and a perpendicular line that separated the best of the Top 6 from the rest (well, except for our 4th best, Cesc Fabregas, who is too low to be included in the group).

My goal was several-fold:

  • See how the Top 50 players stacked up to each other
  • See how each of them measured up to their expected values
  • See how collectively each of the Top 6 clubs clustered compared to each other
  • Look for any bright spots from the rest of the clubs.

Without further ado...

Players on the bottom half of the positively increasing line are doing better than expected. For example, Mo Salah, his actual is 38, and his expected is 28.96. That’s why he’s so far off the chart to the right. I trimmed it at 30 on the vertical because no one was higher, and I hate wasted white space like that.

Things of note:

  • Of the Top Six, Manchester City easily have the most included, with 6 players. They are in bright light blue. Liverpool, with the darker red, have 4 players. United now have 6, with the addition of Sanchez. Spurs have their four, Arsenal now only have three, and Chelsea also only have three (barely, Willian was the cutoff). There are a lot of lead lines, and I apologize for that, some of these players are quite clustered, and the image gets huge otherwise. Blame Salah. He’s the outlier.
  • Arnautovic and Lingard are both in the positive, those two are hard to follow the lines on. Their data points are right in that bit with Wayne Rooney and Willian and Paul Pogba.
  • Notice how all four Spurs players are right in line with their expected values, although all four are very slightly under.
  • Similarly notice that five of the six City players are over, except for Jesus. I say give the guy a break, we did just have Easter.
  • Did anyone before looking at this chart really expect Firmino to be where he is? I know I didn’t.
  • I also didn’t expect Sterling to be higher than Lukaku.
  • Eden Hazard is closer to Wayne Rooney than David Silva (or Dele Alli, and about the same distance behind Morata).
  • For those that kept talking about how good of a season Mahrez is having, yeah, he’s performing above expectations, so it’s not as obvious as all of you keep making it out to be.
  • The combination of Chelsea only having three people above that threshold line I drew (sorry, Cesc) and not having anyone out in front, where the four clubs in front of us each have multiple people, or in United’s case have six in front of the line, speaks volumes to our attacking issues. Realize that the trio of Morata, Willian, and Hazard are all outperforming their expected values. There’s simply not enough hands on deck, and we are lacking a proper ship captain in the attack, to continue the metaphor.

Thoughts, Questions or Feedback? Feel free to share in the comments section.