Debunking the Liverpool ‘defensive crisis’ myth

Since day one of the season for Liverpool, a narrative has been pushed that defensively they’re not as secure this time around. Virgil van Dijk had a lapse in judgement against Leeds United to gift Patrick Bamford a goal and Marcelo Bielsa’s team carved out multiple openings against the champions. In reality, a number of those would have been offside, but the new rule, which allows the game to continue until a shot is taken, skews perception.

For some, that is irrelevant. Jürgen Klopp’s side had a few forgettable moments after securing the title and when all of these moments are grouped together, it doesn’t paint the prettiest of pictures. 

Then that Aston Villa game happened and the narrative was further strengthened. 

Liverpool went from having one of the meanest defences, not just in England but across Europe, to having one of the most porous backlines, and it happened in the blink of an eye. Only Leeds and West Bromwich Albion (both 17) have conceded more goals than the Reds this term. 

The headline figures don’t look great for the Merseysiders. If the season played out using their current average, they’d finish having conceded 76 goals (vs 33 last season). That definitely isn’t going to happen but it does put into perspective why so many people believe their defensive unit is struggling this time around. 

If you write off the Villa game as a freak result, Liverpool have conceded 11 in seven. That is still a relatively high number for a Klopp side, but it does bring their goals against per game average down to 1.57. For context, table-toppers Leicester City’s goals conceded per 90 average is 1.13. 

When you look at the underlying numbers, the two sides are even closer. For this, we’re looking at non-dead ball expected goals against (NDBxGA); this strips away all efforts from set-pieces to determine how good teams are defensively in open-play. Liverpool’s average is 0.96 per 90, while Leicester’s is 0.94. For this particular metric, the Reds rank sixth in the Premier League.

The shot map above details all of the open-play efforts Liverpool have faced this season. There are only a handful of chances that fall into the ‘easy’ category. These are considered high-quality opportunities, with an expected goals rating of close to one. The shot map is the total for the eight matches the Reds have played in the Premier League during the 2020/21 campaign so far. With the exception of the Villa game, Klopp’s side aren’t allowing low-quality efforts to be scored. 

Liverpool have been rather unfortunate, too. Non-dead ball post-shot expected goals against (we’ll just call this PSxGA to avoid the world’s longest acronym) takes into account where the effort goes on target. For example, a player may have a shot with a NDBxG total of 0.6 but it’s a poor effort and the PSxG total is 0.3. They’ve devalued their effort with their finishing. The Reds, however, have seen the opposition add value to their efforts. 

For the PSxGA metric, they rank 13th in the Premier League with an average of 1.24. On a per 90 basis, the difference between their non-dead ball xGA and PSxGA has been 0.28. That will be heavily influenced by what happened at Villa Park, though, with a number of goals on the night hitting the corners. You’d expect this to settle down a little as the season progresses. You’d also expect the Alisson factor to come into play at some point. 

He’s one of the best in the world at shot-stopping and has a knack for saving those in-the-corner efforts. Of the 16 goals Liverpool have conceded in the Premier League this season, nine arrived in the two matches he missed.

You also have to consider the fact the Reds have been without Van Dijk, Fabinho, Joel Matip, Jordan Henderson, Naby Keïta and Thiago for periods of the season already, all of whom influence the defensive side of their game. Injuries galore and bad luck yet Liverpool still haven’t been giving up high-quality efforts on a regular basis. 

Klopp’s side also have the highest NDBxG average in the league and the largest NDBxG to NDBxGA difference. It all points towards another title challenge and goes a long way to busting the myth that they’re all of a sudden a defensive disaster.

All the graphics and visualisations in this article use Wyscout data and were produced in the Twenty3 Content Toolbox.

If you’d like to learn more about our products or services, and how they might be able to help you, don’t hesitate to get in touch.