The Illan Meslier dilemma: should Leeds be thinking about changing their goalkeeper? - The Athletic 21/4/23

Phil Hay and Jonathan Mackenzie

The look of Illan Meslier — hands on hips, the thousand-yard stare — was that of a goalkeeper who had been here before: this week, this month, this season, this league.

The goal just scored was Liverpool’s sixth and not for the first time, Leeds United’s No 1 was counting the concessions. Six against Liverpool followed five against Crystal Palace followed four against Arsenal; three occasions in 17 days when Meslier had done little more with his hands than swipe the ball from his net.

Premier League football is ruthless like that, but very few goalkeepers in it are made to suffer like a Leeds United goalkeeper. Meslier, 23, is approaching the end of his third full season as first choice in the division and, over the course of that period, Southampton’s James Ward-Prowse is the only top-flight player with more appearances than him. Those years have been a mixture of kind and unkind, hoisting him to the forefront of the game but exposing him to an unforgiving world.

Leeds, in this run as a Premier League side, have been defensively dreadful to the extent that Meslier’s 104 games — a mere three short of being ever-present — have come at a cost of 191 goals. Below him, the next goalkeeper to suffer anything like as much from 2020-21 onwards is Jordan Pickford at Everton, playing eight matches fewer and conceding 143. Nothing at Elland Road is built on clean sheets and that is shown by the club holding the worst defensive record in the league this term.

Meslier’s campaign has not been an easy one and he came into sharper focus during Monday’s defeat to Liverpool because of one particular stat: that from Palace’s first goal in a 5-1 rout of United a week earlier to the moment of Liverpool’s sixth, he faced 11 shots on target without stopping one. Everything coming at him was finding the net and Leeds, to be frank, were getting smashed. Was this Meslier’s fault? Was he a victim of a structural mess in front of him? Was it a mixture of both? What was going on?

To this juncture in his career, the sizeable challenge of being the last line of defence at Leeds has not done Meslier much harm in terms of reputation. He is France’s under-21 pick and there was an expectation that he would be promoted to the senior French squad after this summer’s European Under-21 Championship, helped by the international retirement of Hugo Lloris. At 23, he is young for a Premier League player in his position. But Leeds are badly up against it after a punishing fortnight, their Premier League status staked on what Javi Gracia called “seven cup finals”. Fulham away is next. Changes are surely inevitable.

Whether Gracia has grounds or enough justification to replace his No 1, or whether a change there would have a positive impact, is not simply down to goals conceded. Goalkeeping is complex and in these debates, it is always necessary to try to establish cause. Is Leeds’ record poor because of Meslier? Is Meslier’s record poor because of Leeds? Or is the truth somewhere in the middle, a scenario where neither is particularly helping the other?

One way of quantifying this and fairly analysing a ’keeper’s performance is to use the Expected Goals on Target (xGOT) metric, developed by Opta to determine the likelihood of a goal being scored once an effort on target has been struck. By taking the expected goal (xG) value of a chance and looking at the location where the shot ends up, the xGOT model applies historical data to determine the probability of that chance resulting in a goal.

The visualisation below shows all the on-target non-penalty shots Meslier has faced with Leeds in the Premier League this season — 151 in all. The size of the circle represents the probability of each opportunity producing a goal — the bigger the circle, the more likely it was to finish in the net. As you would expect, strikes closer to the frame of the goal are more likely to beat a goalkeeper. Shots in the middle of the goal usually end up being easier to save because that is where ’keepers generally position themselves.

On its own, this visualisation doesn’t reveal too much about a ’keeper’s performance. What xGOT data allows, though, is the addition of the values of all the individual chances to build a picture of how good a ’keeper’s displays have been. In simple terms, it begins to show if a player like Meslier is saving the attempts which, statistically, he would expect to stop.

As the graphic above shows, the 151 on-target non-penalty shots faced by Meslier give a total xGOT calculation of 45.3. It means that at this stage of the season, with 31 games gone, he would have conceded around 45 goals, allowing for a small margin of error either way. Instead, what the league table shows is that Meslier conceded 56 times from these chances, almost 11 more than the data expects. In Premier League terms, it’s a considerable negative.

Repeating this process for every goalkeeper in the division this season makes for grim viewing. Of the 22 on the list (limited to those with 900 minutes or more so far), only Southampton’s Gavin Bazunu shows up as having conceded more goals than expected — an eye-watering 14.53 goals above his xGOT stats. The next worst performer after Meslier is Bournemouth’s Mark Travers with 6.36 over xGOT. What is clear is that Leeds are coming out on the wrong side of the numbers too often.

It is worth noting, though, that Meslier’s form for Leeds correlates pretty well with the team’s overall performance on the pitch. The following graphic demonstrates how the difference between xGOT and actual goals conceded by Meslier has fluctuated over time. The first thing that jumps out is that during the 2020-21 campaign, Leeds’ first back in the Premier League, and up until the middle of last term, his stats remained around the average mark.

As Leeds started to decline under Marcelo Bielsa, however, so did Meslier — a decline which was seemingly only halted towards the end of the season when Jesse Marsch replaced Bielsa as head coach. But once again, when Marsch’s Leeds tailed off after a fairly bright start to the 2022-23 term, so did Meslier’s numbers. In keeping with the trend, there is even a small bump of improvement around the time Javi Gracia replaced Marsch.

With that in mind, it’s natural to wonder in which direction the causality goes here. It goes back to the question asked earlier: are Leeds struggling because Meslier is struggling or is Meslier struggling because Leeds are?

Undoubtedly, the recent decline of Leeds cannot be laid solely at the door of their young ’keeper. There are questions to be raised about how much the underlying numbers are impacted by the lack of protection he is receiving from his team-mates. But one thing is for certain: as it stands, once Leeds’ defence is broken — and broken it too often is — it seems unlikely Meslier will dig them out of trouble.

The Athletic’s goalkeeping expert, Matt Pyzdrowski, was asked to analyse the two concessions in Monday’s defeat to Liverpool which seemed to expose Meslier most — Liverpool’s second, scored by Mohamed Salah, and their third, scored by Diogo Jota.

For Salah’s effort, Leeds were badly caught out by Weston McKennie losing the ball in the centre circle, inviting a counter-attack. In no time, Liverpool worked Salah into the box, with Pascal Struijk a yard behind the Egyptian’s crossfield run, and a sweetly-hit left-footed strike flew inside Meslier’s near post, the area convention says goalkeepers should usually be able to protect. Afterwards, there was criticism of Meslier over his failure to keep the effort out.

“Meslier’s positioning was initially good, just above his six-yard line,” says Pyzdrowski. “Due to the angle and distance of Salah, he needed to then adjust and move himself from the middle of his goal towards his near post.

“Ideally, Meslier would have liked to get one more step towards his near post and set before the strike, but Salah’s quick turn and release made that much harder to do. Salah’s quick strike caught Meslier slightly out of position and opened the space he was looking to exploit. So could Meslier have done better? Yes. But it’s also very well executed by Salah.”

Jota’s goal was just as good at demonstrating the nuance of this debate — that while Meslier could have done better with it, Leeds’ organisation in front of him did him no favours: Struijk missing a tackle on Cody Gakpo, Junior Firpo missing a tackle on Curtis Jones, who then curled a pass to Jota who Rasmus Kristensen failed to mark closely enough on the other side of the field. Meslier committed to rushing out and Jota had the wit and composure to spot him advancing and curl a sharp finish over him.

“Again, Meslier’s initial position was good,” Pyzdrowski says. “Leeds were playing a high line so he needed to have a high position to sweep the ball if necessary.

“As the ball was played by Jones, he had a split second to decide: ‘Am I going to rush out and try to sweep the ball away or retreat to my line for an upcoming shot?’ In a perfect world, he would retreat to his line, but I understand why he did the former.

“The ball continuing to curl away from him significantly impacted the outcome. It looks initially as if he’ll be able to clear it, but he’d already decided to rush out, so there was no turning back. Once he was committed, he had to go.

“I think he realised his miscalculation in real time because he slowed himself down to try to spread himself at Jota’s feet, rather than slide in and risk a penalty. It was an excellent ball from Jones and it asked the right questions of Meslier, which a poorer ball wouldn’t have. Frankly, I’d probably have made the same decision in the circumstances.”

These are the considerations facing Gracia as he tries to put the pieces back together for Fulham away and the fixtures beyond it. He has Joel Robles and Kristoffer Klaesson available to him but given that Klaesson has played all of 35 first-team minutes since coming to England in 2021, the experienced Robles would appear to be the only viable alternative — and Robles is yet to play in the Premier League for Leeds.

Gracia sounded resolute in backing Meslier yesterday. “We’ve conceded 11 goals in the last two games,” he said. “It is not the fault of him.” Like so many things at Leeds, the situation is a world of questions. How much would a change there help? Is it better the devil you know? What cost if Gracia calls it wrong? And once again, who would be a football manager?