Top six slowdown?

The January malaise that has struck towards the top of the table has taken subtle but different form for each club. What? Malaise for the top clubs? Well, perhaps except Tottenham and it might not even matter for Chelsea. Let me explain.

Earlier in the season the league had a very clear structure. The big six were annihilating the rest of the league while the bottom sides were posting truly historically dreadful shot numbers (in a variety of forms) and suffering for it (Hull, Sunderland, Swansea) or unexpectedly prospering (Burnley).

More recently, results have slowed up significantly for Liverpool and slightly for Arsenal, while Manchester City have been erratic and Manchester United regularly frustrated. Chelsea have come down from their ridiculous peak that involved winning every game, while Tottenham have been pretty consistent. At the other end of the table, there have been improvements. The emblem for this change is obviously Hull 2 -0 Liverpool, but that feels like the first crocus of spring; finally we can see the evidence but the bulbs have been in the ground for a while. To describe, in the first twelve games of the season Hull, Sunderland and Burnley were all taking fewer than 33% of the shots in their games; really poor. The second batch of 12 games has seen the worst three teams via this measure taking between 37 and 41% of the shots and Stoke and Middlesbrough have replaced Burnley and Hull in the bottom tier. The bad teams are no longer bottoming out.

Now the top teams are still chugging along at a decent clip and winning a fair share of their games against the rest of the league but changes are afoot. Liverpool, Manchester City and Chelsea were the top three shot teams in the league in games one to twelve. All were taking 16-plus per game with Liverpool at nearly 19 per game. Since then Manchester City’s numbers have held up fine but Liverpool have dropped to 15.4 per game in games 13 to 24 and Chelsea are right down at twelve, with five games against direct rivals having some impact.

All the while, the three teams’ defences have remained stoutly resistant towards opposition shot volume. None of them have conceded an average of more than an extremely strong eight shots per game through this recent period, but the detail here is where they differ: Liverpool were able to ride defensive frailty in the autumn while their attack was crushingly strong, but the shots they concede are rated as the third easiest in the league; they allow few but good chances. When the attack dropped off a little, the lack of defensive resolve has severely hampered their effectiveness in securing results.

Manchester City have suffered similarly but their porousness has looked more a combination of a dismal run of variance compounded by their invisible goalkeeper, although the shots they concede are very average, a contrast to those that Guardiola’s former teams have allowed; Bayern regularly allowed few and comparatively difficult chances.

Chelsea are the one team of the three that appear to have the defensive end locked up; low shots against, low quality of shots against and a small positive variance. They have also been scoring their own reduced shots at a league leading rate, which sure helps. Indeed their whole set-up is incredibly reminiscent of 2014-15, where they rode the numbers somewhat through an extremely hot autumn. Fuelled then by Diego Costa and Cesc Fabregas they were latterly shepherded home by Eden Hazard leading some cute but tight games as they failed to hit the same attacking heights beyond Christmas as prior. Meanwhile all the other contenders fell away and with a smart lead secured, they came home relatively unchallenged. This time, Chelsea have the best defence in the league and it is that which has provided the bedrock that their organised and talented attackers have been able to eke further points from. Antonio Conte certainly deserves his credit. He has tuned his team in his image and brought the exact defensive resilience that built his run of titles at Juventus while creating a scenario that gives Chelsea a ton of breathing space from which they should be able to coast home. In a way, the attack, though quieter, hasn’t needed to be as fired up as earlier in the season.

Along the way, Arsenal have moved slightly in the opposite direction. The early warning signs this year were a lack of shots and their average expected goals per shot were down from last year’s significant peak. Slowly, they’ve found a bit of extra volume (16 shots per game in the middle 12 matches) and slightly raised their total expected goal rate and per shot expectation. Of course this has coincided with a run of 7-1-4 which feels all too familiar, has landed them on the edge of the top four and is partly caused by them matching an expectation of goals against that only ranks 8th during this middle period. Also factors during games 13-24: while they are conceding just ten shots per game, the 41% of them that are landing on target is league high and on average they are from closer than any other team in the league, very much at odds with last season in which their shots against distance was league best.

Tottenham have risen to be the highest volume shooters in this middle third of the season but are slightly hampered by the majority of them being relatively low quality. They take under 40% of their shots from inside 18 metres which is the lowest total of any team in the big five European leagues by more than 5%, but that they take so many shots means that they are powering an acceptable expected goal rate. It’s not a study in efficiency but for now it’s enabling them to keep well in the mix. A prevailing narrative that has built up justifiably from the heavy skew towards long range shots is that Tottenham are bad at them. When a team is crushing long range efforts and netting them with some regularity, there aren’t many complaints about players attempting them, like, say last season when Tottenham scored ten, at a rate that matched expectation. When a team has scored just one goal all year from long range despite taking a ton of shots–enough say for a reasonable expectation to be seven or eight–the fans get a little riled, despite it largely being one of those things. That’s Tottenham this year.

Defensively, they ran hot early on but have suppressed shots in recent weeks impressively, all the way down to fewer than eight per game. The big warning for them here is exemplified by the 0-0 draw at Sunderland, the type of result that they have minimised this season but that can be liable when their attack lacks decisive incision and they will also surely start conceding more regularly from open play at some point, seven all year is remarkably low.

Manchester United have also managed to maintain extremely solid metrics. Their issue is results, as they are have bounced through a series of headscratching draws against weaker opposition. Across the whole season, they have drawn six matches (five at home, one away including both games against Stoke) that they had significant and clear shooting advantages in with regard volume and quality. That’s another vote for “shit happens” but has meant that although their unbeaten run has been noted, Mourinho’s genuine progress and solidity with his team is currently looking like it might be for naught, as they sit on the fringe of the Champions League places. Arsenal and Liverpool should probably be most cautious of the strength of their old rivals, though a couple of damaging results for any of Chelsea’s chasers will quickly turn the top four probabilities upside down.

The bigger picture around this is that it is quite possible that the extremely strong early form shown by the big six, form that was reflected both in metrics and point gathering–coupled with the early ineptitude of the league’s worst teams for both measures–was a kind of bubble that has slightly eased over time. The lack of upset victories from smaller teams was unusual, as was the lack of a team or two from the middle classes skewing positively. Slowly, as attrition affects the league and our old friend random variance meanders along, we’re seeing something that looks a little more normal. The top six don’t usually crush the rest of the league quite as convincingly as they did through the first half of 2016-17, and maybe the second half of the season will see more of these less predicted results.

Leicester

In this week last year I wrote about Leicester getting a few breaks. At that point, Robert Huth had just scored a brace in a 3-1 victory against Manchester City to add to his earlier late winner at Tottenham. We may now be living in strange times, but the seeds were sown some way back. A year on, Huth’s most recent event was less pleasant for him and involved being planted 40 yards up the pitch by Henrikh Mkhitaryan en route to the Armenian’s opening goal as Manchester United routinely brushed aside the Champions. I would not be surprised if Leicester’s groundsman has been instructed to tend Huth in his new location, to add water, a little plant food, in the hope that he will flower come the spring and bring back beauty in place of the ugliness that currently surrounds his team.

huth flower

At this point last season they had converted over 14% of their shots, enough to stick them in the top five teams this decade. By season’s end it had dropped back to 13% and their defence had secured the dough but that understandable slight drop off was followed by a further move in the season’s opening 12 games down to 11%–still above average–while games 13 to 24 have seen the wheels fall right off and their all shot conversion is down to 7% and since the Manchester City win, just 3.1% over nine games. A small irony can be noted here considering a comparable weird freak of variance allowed them to concede just 4.2% of their opponents shots in the second half of last season as they waltzed their way to the title. Two huge non-sustainable skews, one defensive and positive, one attacking and negative and each pointing Leicester at the opposing edges of the league table.

Last year they ended up being a 50% shots team, this year it’s more like 42% but the skews that are killing them right now are in finishing with their own drying up and their opposition’s cantering along happily. That recipe will damage a team because in many circumstances, they won’t be doing all that much different to before, when it was working out and it’s hard to remedy a problem when it’s concealed to some degree in uncontrollable and dare I say it, luck related factors. Jamie Vardy and Riyad Mahrez combined to score 32 non penalty goals from an expectation of around 24 last season and added in nine penalties between them. This season they have seven between them at a rate that matches expectation and just two penalties. The Eden Hazard/Mahrez body swap has expired.

The N’Golo Kante factor has been repeatedly cited as a driver behind the decline of Leicester and return of Chelsea to form. That’s too simplistic for me. Though Kante is clearly excellent, no one player has the ability to drive the extenuating circumstances that have contributed to each team’s trajectory. Chelsea reorganised an already very strong squad that massively underachieved last year and always had the potential to regroup into a contending side, while Leicester were inevitably going to find themselves less successful. Talk of regression as a single factor in Leicester’s decline misses that they have managed to drift past what might be considered a norm for the team. Usually you would look at this team and say they have enough about them to scrabble away from trouble and this was a blip but 14 games remain and that is a short run. An unfathomable bad run could easily stretch out til season’s end and if they don’t start finding the net soon, the resemblance to what they contrived to do to the opposition at the back end of last season will become ever more pronounced.

Could they get relegated? Now that would be unlucky.

West Ham

West Ham’s dramas this season have been twofold. Latterly the Dimitri Payet situation after initially walking into an odd set of shooting numbers across the opening section of the season. They were taking 53% of the shots at week twelve but just 35% of the shots on target. That’s weird. They were also scoring just 8% of their own shots and conceding 14% of the oppositions. These ingredients meant that they didn’t win many games (three) and found themselves in a lowly position in the table (17th). This was a lull akin to that which Leicester are now languishing within.

The run of games 13 to 24 has found the team taking just 46% of the shots but now 45% of the shots on target, a far more normal distribution. Due to the skew at the top, that seemingly sub par figure is enough to rank 10th and their shot conversions have flipped, they are now converting at a rich 13% and conceding at a still-not-great 11%, but it’s been enough to right the ship and helpfully offers a neat counterpoint to the often erroneous “one player can affect results” narrative. Payet was their attack–no team relied more on a single player to create chances than West Ham–but his departure hasn’t hampered their progress. The likelihood was that some of these numbers would shake out and they would find a touch of “form” and climb the table, and indeed they have.

Burnley

Mystery team Burnley have repeatedly defied terrible metrics and a non-existent away record to land themselves safely in mid-table. Expected goals looks at their defence and scratches its head. They’ve conceded about three quarters of the goals expected and theories abound. Does their deep defense stop the opposition? (maybe?) Are they blocking heavily and above expectation? (Yes, to a degree.) Is this enough to account for ten to twelve goals? (Probably not.)

What about their attack? After all they have had to score goals to win nine from thirteen at home. Straight away one stat leaps off the page: Burnley have scored nine goals from over 18 metres or more in 24 games, including seven classed as outside the box. In itself that’s not remarkable, but if we recall that Burnley don’t shoot much and started the season with terrible shot numbers it starts to raise an eyebrow. If we look at a goal expectation for these shots it comes out around four and on a per shot basis, they are scoring from range at a rate over expectation higher than any team in the Premier League this decade. If I tell you Liverpool and Manchester City 2013-14 are ranked two and three in this list, it becomes clear what kind of attacking overperformance this is: spectacular and unsustainable.

This matters not for this season but with eight of those nine goals coming at home and contributing to one goal victories over Crystal Palace, Bournemouth and Southampton as well as the 2-0 win over Liverpool, their points value becomes clear. It won’t be something they can rely on next season.

 

_______________________________________

 

Thanks for reading

Not done one of these for a while so it ended up long!

@jair1970

  • kidmugsy

    I’m disappointed by Leicester’s decline. Some part of the explanation might be the (mild) clampdown on wrestling in the box, and part is the absence of the remarkable Mr Kante. But there is a general air of sluggishness, imprecision and disorganisation. Can any of these be quantified, apart from the two centre-backs becoming old?

    • James Yorke

      Dustin took a pretty solid look at Leicester pre-Christmas, and a lot of it looked like teams had adapted how to play against them (finally!)

      I’m sure you could identify things, but you’d most likely need very detailed data and time etc

  • allanderek

    Very nice article. Quick question with regards to Burnley. The most obvious statistic is the difference in their home and away form, so obvious that the mainstream media are reporting on this loud and often. Do we think this is simply random variation? Perhaps caused by the fact that their wins are at least to some extent caused by some random variation.

    • James Yorke

      It feels like it is. They’ve taken the lead in 11 of 13 games at home which has obviously fed into the results significantly and there’s no way they are a team that is of sufficient quality to mean that’s a signifier towards quality and it’s 1 from 11 away meaning they are about par by that measure.

      I’d suggest the home/away split is a red herring to at least some degree (beyond that you’d expect their home record to be better regardless)

      • allanderek

        Yeah, it feels that way for me too. Thanks.

  • IanSeed

    “The Eden Hazard/Mahrez body swap has expired.”

    That’s gold. Great column all round.