With over a quarter of the season gone, it may seem odd that, when Arsenal and Liverpool clash on Saturday evening, there still remains uncertainty about how good either of the sides are. But neither of the opponents from Matchday 11’s most intriguing tie have lived up to the sort of numbers they have been posting in the Premier League so far this season.
In many respects, Arsenal are to this season what Manchester United were to last season. Despite finishing second in the league last year, Jose Mourinho’s United scored nine goals more than their expected goals (xG) values suggested they would, conceding 15 goals fewer than the same metric projected they should have conceded. As any statistician worth their salt will tell you, over- or under-performance against xG values is unsustainable in the long run and, lo and behold, this season, Manchester United find themselves regressing towards the norm with wild abandon.
By the same metrics, Arsenal’s goal-scoring has already outstripped Manchester United’s, Unai Emery’s side enjoying an eye-watering extra goal scored per xG in every game they have played so far this season. Defensively, they are much closer to what you might expect, conceding a goal fewer than their xG values project.
However, that isn’t to say that there are no worrying factors in this regard. In 10 games, Arsenal have an xG Against (xGA) value of 14.14 which is enough to push them down to 10th in the Premier League table below teams like Everton, Leicester and Watford. That they are fourth in the league table has everything to do with the fact that their attack is over-performing at a rate greater than their defence is performing.
Liverpool are suffering from the opposite problem: their attack is running too cold and is being saved by an impressive defence. That said, Jurgen Klopp will be feeling more sanguine than his counterpart because the bulk of his side’s goal drought can be traced to one individual: Mohamed Salah.
As Andrew Beasley argued in his latest stats column for the Liverpool Echo, ‘[p]rior to the Huddersfield match, Salah had taken shots worth a total of around six expected goals in the league, but had only scored three times.’ Picking up a goal against the Terriers and another at the weekend against Cardiff, the Egyptian is now under-performing to the tune of 1.78 xG, far closer to his projected figure. Once again, regression to the norm strikes back.
At the other end of the pitch, the marked improvement of Liverpool’s defence continues apace. Yes, Klopp’s back line is over-performing: conceding four goals against a predicted xGA of 7.52. However, that figure is way below Arsenal’s own xGA value of 14.14.
On the face of it, then, Liverpool come into this match the favourites. But how will the game play out on the field?
Although they are fabled for playing a [quote/unquote] ‘high octane brand of gegenpressing football’, Liverpool have evolved under Jürgen Klopp to the extent that they now no longer press aggressively from the front, but invite the opposition to transition into the midfield areas before springing pressing traps.
This shows up nicely in Twelve’s data. In an aggressive press, you might expect to see a lot of pressing actions from the forwards as well as the midfielders. As you can see from Twelve’s Off the Ball metric, the majority of Liverpool’s pressing is being done by their central midfielders:
Against an Arsenal team whose manager is trying to transition into a more competent possession-based side, Liverpool’s mid-block pressing approach should be in its element. Although Klopp could be tempted to implement a somewhat more aggressive press in light of Arsenal’s woes playing out from the back in the early part of the season, there is no reason why the Reds shouldn’t keep up their more usual approach to teams, funnelling them into specific areas in the midfield before closing the net.
One way or another, by the time Saturday’s Premier League fixtures draw to a close, we should have a much clearer sense of just how much Arsenal have improved and just how good Liverpool really are.