What has gone wrong with Leeds United’s defence?
It was only a few months ago that I wrote an article breaking down the reasons why Leeds United could be heading for a record-breaking season in terms of clean sheets and lack of goals conceded.
At the start of December, there was nothing to suggest that this record wasn’t attainable for Leeds. They began the month picking up their 4th clean sheet in a row with a 2–0 win away to Huddersfield and came into the home game with Cardiff City having only conceded 10 goals in 21 games: a sensational 0.48 goals conceded per 90 minutes.
For 59 minutes of that game, the 5th clean sheet in a row looked to be nailed on. However, Kiko Casilla flapped at a cross that he didn’t need to come for and it went straight to Lee Tomlin, who volleyed back over his head and into the net.
That goal seemed to be the catalyst for something in Leeds defensive set up to completely crumble. They gave up a 3–0 lead to draw 3–3 and have subsequently conceded 13 goals in the 7 games since the Huddersfield win. This is a much more concerning rate of 1.85 conceded per 90 and is Leeds’ longest stretch without a clean sheet all season (the previous longest run was only 2 games). It has also meant 1 win in 7 games. In their previous 7 games, Leeds won all of them, only conceding twice.
For all the joys of Bielsaball in attack, Leeds’ promotion bid has been built on keeping the ball out of their own goal, more than it has putting it in the net of the opposition.
So, where has it all gone wrong?
Lack of control in midfield
As we’ve said many times lately, Leeds’ greatest strength this season has been their control of games. They have dominated the ball and kept opponents at an arm’s length.
During this recent run, however, they have rarely had control of the midfield. This is hardly surprising considering that every first-team player who would usually partner Mateusz Klich in their first or second most comfortable position has been injured.
This has meant Stuart Dallas — usually utilised as a wingback and previously a winger — has done the job. Marcelo Bielsa clearly trusts him as a character and he knows the system well. However, the structure of the team hasn’t looked right with him playing there and he doesn’t possess the technical attributes you’d expect of one playing in that role.
Subsequently, Leeds have given the ball away in transition more often and been punished for it. Before the Fulham game, they were averaging 33.8 losses of possession in midfield areas in 22 games. That game saw Dallas move to midfield in the second minute after an injury to Pablo Hernandez and he has started there in every league game since. The average turnover in midfield since then has risen sharply up to 37.8.
Whilst this is a small sample size, it is clear to anyone watching that Leeds are not building through the midfield nor are they retaining the ball in the same way that they were in the winning run the preceded this poor form.
The problem with losing the ball in transition from defence to attack is that is leaves a team completely out of position and vulnerable to counter attacks. Preston’s goal on Boxing Day was a prime example of that:
The clip starts just too late to showMateusz Klich losing possession of the ball after Leeds had finally broken out from the shackles of Preston’s pressing. With Leeds full backs and midfielders pushing on, Ben Pearson only needed to play one line-breaking pass through the middle to give Preston a 3 v 2 situation which they took advantage of.
Jacob Murphy’s crucial opening goal for Sheffield Wednesday in their recent 2–0 win at Elland Road also came with a loss in midfield. This wasn’t so much a transition from attack to defence as just a failure by Dallas to control a high ball. He did have pressure on him but one would still expect better from him here. His stray touch allowed Atdhe Nuhiu to release Murphy who ran through to score.
There are other errors here — from Ezgjan Alioski, who was drawn towards the ball, meaning Murphy could run between him and Liam Cooper; and then from Kiko Casilla, who allowed the shot to go straight through him at the near post. This leads us nicely onto the next problems.
It would seem that one way in which Bielsa might be able to affect this situation is through deciding on a consistent left-back.
Leeds have used three players at left back so far this season: Barry Douglas beginning the season there and playing the first four games before — as always seems to be the story with him at Leeds — he picked up an injury and was replaced by Alioski. Stuart Dallas was on a run of starts at left back before he was needed to move into midfield.
During most of the last seven games, Alioski played the position. Douglas did come on at half-time against West Bromwich Albion and started in the most recent game against Sheffield Wednesday. Whilst Douglas lacks in pace and was struggling in the early part of the second half, it is notable that Leeds didn’t concede in the second half against West Brom and the Sheffield Wednesday goals didn’t come until he had been substituted.
Alioski can bring plenty of attributes to Leeds when utilised at left-back. His energy for the press is second to none and he regularly arrives in the box to get shots away, resulting in him scoring a good amount of goals from defence.
However, his positional discipline is lacking. This average position map, taken from the 5–4 win away at Birmingham shows him (10) ahead of Jack Harrison (22) who was supposed to be playing as left winger.
One could argue that it’s strange to see Harrison positioned so deep. But this was a game in which Leeds were having to defend a lot so it seems justified to me. That said, it isn’t good to see Alioski so high up. Particularly when considering that three of Birmingham’s four goals came down Alioski’s side. Take the fourth, for example, where the Macedonian was naively rampaging forward looking for a fifth goal whilst Leeds were 4–3 up. Where he perhaps should have been looking to kill the game with a spell of possession, instead he drove forwards, lost the ball and Birmingham broke and scored.
As we mentioned above, he was also guilty of getting sucked towards the ball before the killer pass was played for the opening goal against Sheffield Wednesday, adding it to a list of recent errors he’s made playing at left back. Perhaps the time has come to use him more as an option on the right or left wing, rather than at left back. His performance at Arsenal was encouraging, showing he can press from a much higher position without his positional indiscipline being so catastrophic.
One of Barry Douglas’ best attributes is a calmness in possession. It was notable how confident Leeds were at passing through Arsenal when he started at left back. With control of the ball being more of an issue lately, a calming influence at left back might be required, rather than the unpredictable Alioski.
In the midst of Leed’s aforementioned run of clean sheets, Kiko Casilla stood as the top goalkeeper in the league according to Opta’s ‘expected goals on target’ metric.
According to this table, based on shot trajectories that Casilla had faced, his saves had prevented over 5 goals. However, his form has gone sharply downhill.
The Cardiff game was the next match after this table was published and in the time between then and the end of the Sheffield Wednesday game Casilla faced 17 shots on target and conceded 13 of those. This means a shocking 76.47% of shots on target have ended up as goals.
It is well-documented by now that Leeds’ own conversion rate is poor, with the Whites often having to work very hard for their goals, so Casilla’s current form is certainly a worry. They can’t afford to have a goalkeeper who is saving less than a quarter of the shots he is facing.
Casilla’s save percentage dropping means that he is now ranked 10th in the league for that metric, where he was previously trading with Brice Samba of Nottingham Forest for the top spot.
One might question whether this is because Leeds have recently given away such good chances and if the former Real Madrid goalkeeper could have done anything about them.
There is some truth to this. Just a few weeks ago Leeds had the lowest xGA (expected goals against) and gave away the fewest Big Chances (>0.3xG) per 90. However, in this recent run they have been so open defensively that they have been usurped by Brentford for the lowest xG & fewest Big Chances conceded per 90.
On the other hand, something that was previously consistent for Leeds this season was that they were appearing on the left hand side of the graph below, i.e below the median line for goals actually conceded from Big Chances faced. In recent weeks Casilla has saved so few that they now go from comfortably under the median to beyond it.
The fact of the matter is, in the last seven games, Leeds have conceded 13 goals from an xGA of 7.88. That is 60.62% more than expected.
At times opponents have finished clinically, such as Onomah’s winner for Fulham which gave Casilla no chance. But he was beaten far too easily by Murphy against Wednesday, stayed back on his line for Alan Browne’s goal for Preston when he could have come out to narrow the angle and has been caught out positionally on several occasions at set pieces, most notably for Jeremie Bela’s header at Birmingham, where he came rushing out to claim the cross and got nowhere near it, allowing it to be headed into an empty net.
Finally, Leeds’ Achilles heel all season has been defending set pieces. Even in the previous article where I spoke in glowing terms about Leeds’ defensive structure, I had mentioned that this was where they had been vulnerable.
In recent weeks it has felt at times like Leeds might concede from every set piece they faced. The truth is that only 28.57% of corners given away in the last seven games have resulted in an effort on Leeds’ goal, which is actually less than Leeds’ average of 29%.
However, Leeds have conceded 3 times from corners and free-kicks during that time. Semi Ajayi’s opener at the Hawthorns in Leeds 1–1 draw with West Bromwich Albion felt like a classic example of Leeds’ issues as Casilla failed to hold the initial cross before two poor headed clearances resulted in Ajayi poking home from close range. The fact that it all happened in the first corner of the match in the second minute of the game compounded the feeling that Leeds were seconds away from disaster at every set piece.
Leeds could almost afford to be poor at defending set pieces earlier on in the season as they were so tight in open play that they were the only real chances they were ever giving away. On top of this, if they were controlling the ball as much as they wanted to, they would naturally limit how many set pieces the opposition were winning in dangerous areas. As long as Leeds were passing the ball in front of the opposition, it was impossible for them to win corners or free kicks in the final third.
Can they fix this?
Whilst it has been an atrocious run for Leeds in defensive aspects, I think the answer to this is: yes.
Most importantly, midfield options are returning: Pablo Hernandez came on as a substitute in the most recent game, Jamie Shackleton returned to the matchday squad and Adam Forshaw is edging ever-closer to being included again. One would expect that Hernandez will start in midfield for Leeds’ visit to QPR, which should mean Leeds dominate the ball again and pin QPR back in the way they are more accustomed to playing.
It may be a cliche, but having been missing since the first week of September, Adam Forshaw will feel like a new signing when he finally makes his return and he is the missing link in Leeds’ midfield currently. Leeds’ structure and control of games has not been more impressive at any point this season than it was in the opening month of the season.
As for the left back issue, I’ve made my suggestion that Alioski needs taking out of the firing line and Barry Douglas given a run of games but Marcelo Bielsa’s lineup choices can often be surprising. He can go from picking the same lineup week after week before suddenly throwing a curveball selection in there, so it will be interesting to see what he chooses to do. One can’t help but feel that he can’t be happy with Alioski’s recent performances, though, since he substituted him at half-time against West Bromwich Albion and dropped him for the next game.
I’m afraid that whilst I believe defending from open play can be fixed, I’m not so sure about set pieces. Perhaps Leeds could attempt to fix it in a similar way to how Pep Clotet fixed Leeds’ troubles with set pieces in his time serving as Garry Monk’s assistant manager. At that time, however, Leeds possessed players with great height and aerial ability such as Pontus Jansson and Chris Wood who were the only outfield players given no men to mark at corners and were simply assigned the task to attack the ball whether it went to the front or back post. This current side does not have such strong aerial presences as those two.
As has been discussed often, one of those players was in the side until the summer. But he was sold and replaced by a different type of defender, Ben White, who has improved Leeds in open play. This is the trade-off of recruiting technical players who are comfortable with the ball in open play: they aren’t so physical and good at defending set pieces. Pep Guardiola knew this and once said of his Manchester City side, full of technical players and with the best defensive record in the league, but extremely vulnerable from set pieces, ‘when it’s a set piece, you go to church to pray.’
If that’s how Pep felt in a team winning the Premier League, Leeds fans might need to get a bit familiar with their local church before the end of the season.
You can follow Josh Hobbs on Twitter @JoshAHobbs.
If you enjoy this content and want to help us create more and better pieces, why not consider signing up to our Patreon page to get bonus material?
If you have any interest in contributing to this blog, get in touch with us @AllStatsArentWe.