How the Championship was Won: Leeds United Season Review 2019/20

In this article, Josh Hobbs and Jon Mackenzie look over the season that saw Leeds promoted to the Premier League, giving eye to the underlying numbers behind Marcelo Bielsa’s team.

In the eyes of many watchers, Leeds United had been the best team in the Championship in 2018/19, despite ultimately finishing third and crashing out of the playoffs in dramatic fashion losing in the semi-final against Derby.

The summer saw them sell top scorer Kemar Roofe, highly-rated youngster Jack Clarke and, perhaps most disappointingly to the fan base, Pontus Jansson. The Leeds Twittersphere was in meltdown: surely the club would be going backwards, rather than competing for promotion again?

In the end, it was really never in doubt. Despite the loss of those key players, another horrible run of form in the new year and a global pandemic, Marcelo Bielsa’s men were not to be denied. They won the Championship title with 2 games to spare.

This article will break down the numbers behind the Champions and look at the key reasons why they were not only able to reproduce last season but better it.

Defensive Improvements

In 2018/19, they conceded 54 goals from an xGA (expected goals against) of 47.04. That was enough to rank them second in the league for expected goals against, behind only Sheffield United. Their goals conceded (discounting goals in the playoffs) saw them ranked 3rd behind Sheffield United and Tony Pulis’ Middlesbrough.

This season Leeds’ xGA has been a miserly 35.91 and they have conceded 35 goals. This makes them the best defence in the league both in terms of expected goals and actual goals conceded.

As a team, they have kept a clean sheet on 22 occasions. This accounts for 47.82% of the games Leeds have played in the Championship this season. This ability to keep clean sheets has been enormous as, although Leeds are 2nd in the league for goals scored with 77, their troubles in front of goal are well documented. With their defence being so strong, 1 goal was often enough to win a game. Subsequently, they won 9 games 1–0 on the way to the title.

These numbers are extremely impressive. It should be noted that Leeds were on target for a record-breaking season in terms of defensive performance before they went on a run of 11 games without a clean sheet, during which Kiko Casilla let in 57.57% of the shots which were on target. 54.28% of Leeds’ total goals conceded for the season were conceded in that stretch of 11 games.

The other issue in that stretch of games was that, by that stage, Leeds had lost Adam Forshaw — who had started the season so well — Jamie Shackleton and Pablo Hernandez, all to injury. That saw Stuart Dallas move to midfield. While Dallas had been one of the team’s top performers at full back, it’s fair to say that he was a little out of his depth in midfield and Leeds didn’t control games in the same manner.

This shows us well in the five-game rolling xG average that Leeds put up through the course of the season:

Each point on this rolling average gives you the average xG and xGA of five game blocks starting from games 1–5 and ending with games 42–45. As you can see, the period where Adam Forshaw was present (numbered 1–4) was the most dominant period that Leeds had in 2019/20.

Looking further along the graph, the period in which Leeds conceded 19 goals in 11 games also coincides with the only period in which Leeds averaged over 1 xGA per game. Aside from that, they have kept their xGA below 1 over a five-game rolling average. This is exceptional and shows why Leeds have kept so many clean sheets.

So, what was the reason for such an incredible defensive performance this season?

Of course, the first thing to mention should be that the team have benefited from a year of Bielsa’s methods and grown more comfortable with the system. Thus, they’ve given less away in defensive transition. At times in Bielsa’s first season, he took Kalvin Phillips off in the first half due to him struggling defensively. That wasn’t necessary at all in the season that saw Leeds crowned Champions: Phillips made 11.17 ball recoveries per 90 and covered almost the entirety of the midfield third with his defensive work.

However, the biggest change has been a personnel change at the heart of defence:

Ben White

There are plenty of pundits who considered Jansson to be the best central defender in the division. After the season White has had, where he has played every single minute of the season, there won’t be many outside of Brentford who still rate Jansson the better of the two.

One thing that White has particularly excelled in is interceptions. He ranks 2nd in the league for interceptions per 90 with 2.5. His superb anticipation has meant that he has known the perfect time to get in front of a forward and steal the ball off their toes before they can receive a pass.

Considering Bielsa is known for his man-marking system, this attribute makes White the perfect centre back for Leeds. He is able to stop so many attacks from building by taking the ball from a centre forward when they’re looking to hold the ball up.

The image above is a perfect example of this. As Cauley Woodrow looks to put his body in the way of White and receive the ball with his back to goal, White manages to come from behind to get in front of him, knocking the ball towards Tyler Roberts, who ccomes to receive it and build an attack for Leeds.

White’s defensive actions map shows how much Leeds have been on the front foot this season. The shaded area shows the most concentrated area of his actions. The fact that it goes all the way into the centre circle and even into the opposition half sums up how he has kept Leeds on the front foot.

In contrast to this, the shaded area of Jansson’s defensive actions from 18/19 does not cross the halfway line.

It would be fair to say that White hasn’t been an upgrade to Jansson in all areas, though, as White did not dominate aerially in the way that the Swedish international did. This meant that Leeds were vulnerable from set pieces with 37.14% of their total goals conceded coming from dead ball situations. However, trading strength at set-pieces for increased control in open play is vindicated by winning the title.

As well as his defensive work, White was exceptional on the ball. His calmness in possession has made Leeds much more confident in the build-up phase and his ability to carry the ball forwards (he ranks in the 86th percentile of centre backs for progressive runs) has resulted in some dangerous attacks. The most notable example of this was his wonderful assist for Patrick Bamford as Leeds won 2–1 at Luton.

Continuing to Trust the Process

This means that they have been less clinical but they’ve always backed themselves to continue to create chances and that they would take one in the end.

As you can see in the xG difference and goal difference aggregation above, Leeds’ expected goal difference has been steadily increasing throughout the season. According to xG, they only ‘lost’ (a difference of greater than 0.3 xG in favour of the opposition) 2 games over the course of the season.

Leeds had such a brilliant attacking process but their profligacy in front of goal meant that they weren’t able to keep up with it. Combined with their poor run of conceding goals, this meant their goal difference fell right away from their expected goal difference.

However, they continued to trust the process and turned it around after their 2–0 loss to Nottingham Forest as the run of games after that and leading up to lockdown resulted in them scoring 10 goals in 6 games whilst only conceding 1.

This passing network graphic shows how the process has worked in practice. The majority of Leeds’ build-up has come on the right-hand side of the field. Pablo Hernandez and Mateusz Klich have both drifted to the right as Leeds overloaded that flank. They were then able to switch the ball to the left, which leads us on to one of the most improved players in the team:

Jack Harrison

This season he has played 4050 minutes and has made 14 goal contributions with 6 goals and 8 assists. These have come at a rate of 0.13 goals per 90 and 0.18 assists per 90 which is up from 0.13 and 0.06 last season respectively. It’s fair to say he would have hoped for more goals considering his xG of 9.27. However, his 8 assists are bettered by only 4 players in the division.

One particularly excellent one was this one against Huddersfield. His superb cross found Hernandez at the back post to head in after a superb counter-attack to kill off a resilient opposition.

Harrison’s crossing accuracy has been below average for wingers at 27.94%, but it’s harsh to judge him on this metric. As Leeds operate with only one striker, he doesn’t have an array of targets to hit with his crosses. Crossing accuracy is also a poor metric to use in isolation as it doesn’t account for the fact that some excellent crosses go down as ‘unsuccessful’ just because a striker somehow doesn’t get on the end of them. A particular incident against Luton Town comes to mind.

Harrison also made great improvements as a dribbler this season, averaging 6.05 dribbles per 90 at a success rate of 55.8% up from 5.76 at a success rate of 46.4% in 18/19.

The image below shows that he has been a constant threat high up the pitch on the left-flank with the vast majority of his actions coming there. 51.26% of all his actions attempted on the left of the final third have been successful with a success rate of 36.5% of those attempted in the penalty area.

His improvement this season, combined with the arrival of Helder Costa on the right flank, has meant that Leeds have been able to play with much more width in the forward line compared to when Pablo Hernandez predominantly played on the right in 2018/19.

Tactical Shifts

As we have already mentioned, Leeds like to build-up on the right side of the pitch, creating an overload so as to isolate the left winger and full back which offers a switch in play to force a one-vs-one situation between winger and opposition full back. Just look at the passing network again and see how the forward players drift right but the left winger stays high and wide:

Originally, this system was set up to allow Leeds the ability to use Pablo Hernandez in a right wing position, dropping deep, helping with build-up and drifting inside. With the arrival of Helder Costa this season, Marcelo Bielsa had a little selection problem: how do you fit all of your wide players — Hernandez, Costa and Harrison — on the pitch?

In the end, injury solved this problem. With Adam Forshaw out, Hernandez was required to play in the middle much more. Leeds were now playing with two traditional wingers in Costa and Harrison, although both are left-footed meaning Harrison went outside and Costa inside.

On the face of it, you might expect this to mean that Leeds’ attacking phases changed somewhat. However, the data shows a remarkable similarity in terms of where Leeds have attacked. As this flank attack diagram for the first and second half of the season shows, not much changed from the beginning and the end of the season in terms of location and volume of attack:

As you can see, the volume of attacks stayed relatively stable, although the colour of the arrows shows that Leeds became marginally more dangerous down the right-hand side in the second half of the season.

However, when you compare average positions for the same periods, a different picture emerges:

The average player positions reveals a tactical shift that took place at some point midway through the season.

As you can see, the first half of the season saw Leeds playing a fairly standard 4–1–4–1, albeit with a slightly inverted right back. By way of explanation, Luke Ayling missed the first 9 games of the season due to injury and so Stuart Dallas filled in at right back for those games. Marcelo Bielsa likes Dallas because he can use him as an inverted full back, pushing him into the middle during build-up phases to help out Kalvin Phillips.

In the second half of the season, however, we started seeing Leeds playing something of a 3–3–1–3 hybrid with Dallas playing as the right wing back but essentially playing as an inverted full back alongside Phillips. This allowed him to offer protection for Ayling who got forward much more frequently which in turn allowed Costa to use him overlapping on the wing. Dallas gave that extra cover in defence and also the extra help overloading on the right-hand side.

As you can see on the average position map from the second half of the season, Costa was able to get forward more often as a result of this and in a wider more traditional role.

Pablo Hernandez

Hernandez himself said that last season was perhaps the finest of his career. He racked up 12 goals and 10 assists at a rate of 0.28 and 0.23 per 90 respectively. It was an almost superhuman effort from the Spaniard and it was cruel to see him miss out on promotion after such an incredible season.

Due to the crazy standards he set in 18/19, there were times over the course of the season where it wasn’t uncommon to see the phrase ‘Pablo is broken’ on Twitter. Problems with his hamstrings meant he missed 10 games of the season and, in the end, he only played half of the available minutes. He was regularly substituted and played primarily as an impact substitute in the 9 games after lockdown as Leeds carefully managed him through the run-in due to the fact that he’d picked up another hamstring injury just before the return of the Championship.

In the end, Hernandez played like a man on a mission in the post-lockdown mini-season and finished the season with 9 goals (0.3 per 90) and 10 assists (0.37 per 90). His underlying numbers actually increased as he ended the season with 0.23 xG per 90 up from 0.21 in 18/19 and had an xA per 90 of 0.3, up from 0.27 in 18/19.

Amongst those assists were some of the most spectacular passes the Spaniard has made across all his seasons at Elland Road. Between his assists for Dallas against Stoke, Costa against Cardiff, Harrison against Fulham and Dallas against Charlton, he had himself a ‘pass of the season competition’.

One key aspect of Hernandez’ season was his shift from the right wing where he played the majority of 2018/19, to midfield. This didn’t change his average position hugely. When he played on the wing, he tended to drift inside to the right half-space and when he played centrally he tended to drift right towards the half-space. This is shown by the two passes received networks below. The first is from 18/19 and the second is from this season.

The main benefit of Hernandez moving centrally was that it meant that Leeds could play Helder Costa on the right, thus being able to stretch the play with pace on both wings.

Whilst Costa’s price tag of £15m meant that his total goal contribution of 9 (4 goals, 5 assists) left a little to be desired, his ball carrying from the right gave Leeds’ attacking process another string to its bow and meant they replaced some of what Samu Saiz brought to the side in attacking transition.

The other key aspect of Hernandez moving centrally is that it helped Leeds get him on the ball as often as possible. He received the ball on average 47.09 times per 90 in 19/20, up from 42.63 last season.

Once he receives the ball, his first thought is to progress it towards the opposition goal. As Leeds’ playmaker in chief, he hasn’t simply created chances but has made the team tick going forwards.

The graph above shows all midfielders in the Championship who have played at least 1000 minutes across the season. On it, passes per 90 are plotted along the x-axis against the percentage of those passes which are progressive on the y-axis. As you can see, Hernandez stands alone. No midfielder in the league is as involved as he is whilst keeping their side constantly moving forwards. Only Harry Arter has made more passes per 90 than Hernandez. But he sits in the bottom right quadrant of the graph as he progresses the ball less than 12% of the time.

There is so much more to say about Hernandez, such is his importance to the team. He has the highest ‘+/- per 90’ (a metric created by FBref.com which is essentially goal difference during the time an individual is on the pitch) with +1.37. Leeds averaged 2.02 points per game over the course of the season. But with Hernandez in the team the figure rose to 2.22.

In the interest of this piece not going on for an eternity, though, we need to move on though and focus on some of Hernandez’ supporting cast:

Leadership Group

Three players in particular stand out in this regard. Club captain Liam Cooper has had an almost spotless season at centre back with a poor pass to give away the second goal in 2–0 loss at Cardiff the only blemish on an otherwise superb season. In the end, the season saw him ranked 3rd in the league for defensive duel success rate.

Stuart Dallas had a fairly indifferent 2018/19 as he struggled to pin down a starting position and fought off several injuries. However, he ended the season brilliantly, filling in at left-back as Barry Douglas and Ezgjan Alioski were both injured. In 2019/20, he started every game except one which he missed through injury as Bielsa grew to trust him so much that he simply always found a way to get him into the team.

He also weighed in with some crucial goal contributions, scoring 5 and making 2 assists.

Another of Leeds’ leaders has been Vice-Captain Luke Ayling, who played like a man possessed after Leeds lost 2–0 away at Nottingham Forest. That particular loss saw Leeds’ hold on an automatic promotion place cut to only a single point.

Ayling scored 3 in the next 6 games as Leeds went unbeaten, winning 5 in a row before lockdown. That run culminated in what many believe is the goal of the season, crashing a flying volley in off the crossbar against Huddersfield.

He also played a huge part in arguably the defining moment of the season, as Leeds grabbed a late winner at Swansea.

Leeds had struggled to break down Swansea for much of the game, one Patrick Bamford header aside. As time ticked down, Leeds worked the ball forwards from the back and Ayling charged forward from right-back, overlapping Helder Costa, who played the ball into his path.

Driving to the byline, Ayling pulled the ball back towards Pablo Hernandez, who collected the ball — via a slight deflection off Marc Guehi — before slotting into the far corner and sending Leeds into raptures.

Ayling had been a threat in the opposition penalty area all season. He ranks 8th in the division for fullbacks when it comes to touches in the box. This is particularly impressive when considering the fact that he has played a significant amount of minutes at centre back as well.

Unwavering self-confidence

The former Boro man’s struggles in front of goal were extremely well documented. Press and fans alike asked the question, ‘can Leeds go up with Bamford leading the line?’

Bamford went 10 consecutive games without a goal and ended up finishing the season almost 11 goals behind his xG for the season. Even during this period, Bamford kept himself in the team with his incredible work leading the press, which Bielsa didn’t see in Eddie Nketiah, despite the Arsenal loanee’s obvious prowess in front of goal. When Bielsa finally decided to put Nketiah in from the start, he picked up an injury and it was at that point that Bamford began to find his shooting boots again.

Even throughout his dry spells, Bamford kept putting himself in positions to score, taking the most shots per 90 in the league and scoring crucial goals. Leeds ultimately didn’t just go up with Bamford as their striker but did it extremely comfortably in the end.

Bamford finished the season as Leeds’ only player to break double figures for goals. When having a season like Bamford was having in terms of conversion rates, many strikers would hide away from getting chances. Throughout, he maintained his self-confidence and he ended up with the reward of being top scorer for the Champions.

And in the end, Bamford came back from the lockdown period much stronger in front of goal than he had been before it:

As you can see, Bamford’s pre-lockdown goals figure of 0.39 per 90 was slightly bettered after the restart with the striker creeping up to 0.41. However, comparing the expected goals figures that these goals came from, he went from underperforming his xG by 0.26 per 90 to overperforming by 0.04.

So, what next?

These things will need to happen quickly, particularly tying down Bielsa. For now, though, this achievement needs celebrating. The underlying numbers have been spectacular all season and it might have taken a while for the table to show Leeds’ dominance but in the end it did. The largest margin between first and second in the Championship for a decade. Champions by 10 points.

All the data visualisations in this article except for the two xG plots and the scatter graph come courtest of Twenty3. All data is from Wyscout.

You can follow Josh Hobbs on Twitter @JoshAHobbs.

You can follow Jon Mackenzie on Twitter @Jon_Mackenzie.

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.

A Leeds United blog which focuses on the tactical and statistical aspects of the game