Are Leeds disproportionately punished for the chances they give away?

The data graphics in this piece are provided by twenty3 and the data itself is from Wyscout.

Before the COVID-19 lockdown, Leeds United were on a run of 5 games without conceding, winning all 5. Unfortunately, since returning, they’ve conceded 4 in 4 which has resulted in them losing to Cardiff and drawing to Luton.

In this time, Leeds have allowed 33 shots against, a figure of 8.25 per 90, slightly up from their season average of 7.5 but a negligible difference. However, they’ve only given up 1.7 xG in this period with none of the goals conceded coming from ‘Big Chances’ (chances that are 0.3 xG or above).

This has led many to ask the question: are Leeds disproportionately punished for the chances they give away? Why do they concede so many from such low-value chances?

As the shot map above shows, Leeds have been primarily conceding difficult chances since the return of the Championship. As you can see, none of the chances above are considered to be easy (the darker the colour of the marker, the better the chance).

It’s worth noting that the bigger chances they’ve conceded have been missed: notably, Sam Gallagher shooting wide in a one-on-one and Lewis Holtby hitting the post from a cutback in the recent 3–1 win at Ewood Park.

Has this been an issue all season?

At times this season, it does feel like Leeds have been disproportionately punished, given the quality of the chances.

Lee Tomlin’s goal for Cardiff in the 3–3 draw at Elland Road springs to mind, as does Jacob Murphy’s goal for Sheffield Wednesday as Leeds lost 2–0.

Both of those goals came from errors by Kiko Casilla. In the case of the Tomlin goal, Casilla had punched it out and was off his line, meaning that, although it was a difficult execution for Tomlin, he only had to hit the target.

Murphy’s goal came in a run of particularly poor shot-stopping for Casilla, the Spaniard letting the ball through him at the near post.

It should be noted that before the break in the season, Leeds had only conceded once from outside the penalty area. That particular goal wasn’t even a shot. It was the corner that looped over Casilla as Leeds lost 1–0 at home to Wigan.

In fact, most of the low-value shots Leeds have conceded from have been headers from set-pieces which, although they have been poorly defended by Leeds, are statistically much less likely to be scored than we tend to imagine.

This season, Leeds have only conceded marginally more than their expected goals against with 34 goals conceded to 32.41 xGA.

This would suggest that it isn’t the case that Leeds have been disproportionately punished. Although some of the goals have come from extremely difficult chances, it also means that they haven’t been punished for some of the bigger chances they’ve given away.

This is why expected goals are best used for longer-term forecasting rather than single-game and particularly single-shot analysis.

Analysing goals conceded post-lockdown

With this in mind, it’s still worth asking if there was a reason Leeds have been punished disproportionately in this mini-series of games.

To do so, we’ll analyse the goals Leeds have conceded in open-play (discounting Adam Armstrong’s excellent free-kick at Blackburn).

The first was Junior Hoilett’s shot from the edge of the box. Infogol rated this as just a 4% chance which makes it the lowest value of the 4 goals conceded.

In this case, I’d argue that the chance was a lot easier than other shots taken from a similar position.

The chance came after Kalvin Phillips had given the ball away inside the centre circle; Hoilett intercepted the loose pass and drove forwards to the edge of the box.

Phillips chased him, but ultimately ended up just running beside him, perhaps thinking that Ben White or Liam Cooper would step in to make a challenge.

However, it was apparent that White was aware of the danger on his right shoulder, as suggested by his body position. Making sure he could cut out a pass played outside of him, he allowed too much space between him and Cooper.

Cooper had eyes on the ball but was a little too far away from Hoilett. Both players were back-pedalling as well. This allowed Hoilett far too much time and space.

The camera angle above shows that, by the time Hoilett set himself to shoot, it was much too late. White and Phillips attempted to close him down and Cooper was a long way from making a block.

When Leeds themselves take shots from this kind of range, they are often surrounded by defenders especially when facing up against a team in a mid or low block. Scenarios like that are far harder to score from range due to bodies being in the way.

That being said, it still had to be an extremely good strike to beat Illan Meslier. Perhaps it would be harsh to criticise the young Frenchman, but I noticed he took a half step to his left just as he got himself into his stance to leap.

This ended up being very costly. It’s perhaps a technical aspect that he’ll need to improve on as he develops.

Next is Robert Glatzel’s goal from the same game.

Again, this goal came from giving the ball away and Leeds being poor in defensive transition. One pass is all it took after Liam Cooper played a sloppy ball to Stuart Dallas at left-back. White and Cooper couldn’t get close enough to Glatzel to close down the shot and he drove a rasping half-volley in off the near post.

It’s another occasion where the finish is superb. It has to be hit very well to beat Meslier. But a good striker of a football would be able to manage to score shots like this under little pressure more often than xG suggests, which is in this case 5%.

It would still be wrong to consider this a ‘Big Chance’, but I’m sure Glatzel would be scoring more than 5 if he took this shot a hundred times.

Finally, we get to Harry Cornick’s goal in the 1–1 draw with Luton Town. Despite being closer to the goal, I think this is a far more challenging finish than the previous two. Infogol has the xG at 0.1.

White took the blame for this but he does a lot of things right. He was clearly trying to show Cornick down the line where he could safely make the challenge.

Unfortunately, Cornick pulled off a sensational finish with no backlift. White would most like have made the challenge as Cornick set himself to shoot but the way he struck the ball took that out of the equation.

Perhaps White could have gone straight for the ball, but then he’d have looked foolish if Cornick had nipped away from him and had a one-on-one with Meslier. Sometimes you just have to hold your hands up and say it’s fantastic execution from the opposition.

It should be noted that this chance again came Leeds giving the ball away and the ball was in their net just a few seconds later.

What about Brentford?

While Leeds conceded 4 goals in the first four post-lockdown games, Brentford conceded 0 despite having a similar xGA (1.44).

They did subsequently concede a goal against Charlton from 0.74 xG but Macauley Bonne’s shot registered 0.5 and made up the majority of that total, making it a very good chance.

Of course, it could all have been very different.

Mitrovic missed a header in the six-yard box. Bobby Reid hit the post when through on goal and Zohore hit the bar in a similar situation to both Glatzel and Hoilett with a free shot from the edge of the box.

What Brentford haven’t been doing, then, is conceding lower percentage chances.

This could explain the difference between the two sides. As Leeds are so attacking, they are more susceptible to conceding goals after losing the ball in attacking transition,

There is clearly work to be done considering that all their last 3 goals conceded in open play have come immediately after they’ve lost possession. If they can fix that problem, they can be an impenetrable defence once again.

What does this mean for the final games?

It’s been very frustrating from a Leeds perspective to see the goals that have been conceded given the general lack of quality chances that they’ve given up.

However, there’s every chance that Leeds could now put together a run of clean sheets, similar to the one they were on before the lockdown.

There couldn’t be a better time to do so, particularly given the well-known issues at the other end of the field and with Brentford and West Bromwich Albion looking imperious.

You can follow Josh Hobbs on Twitter @JoshAHobbs.

The data graphics in this piece are provided by twenty3 and the data itself is from Wyscout.

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