Arsenal answer critics but their problems run deep
By Daniel Tallos
Bournemouth, Swansea, Brighton & Hove Albion. Throw in Östersunds FK and you have the most unlikely quartet of football clubs that have swept Arsenal FC aside in the past two months. A far cry from Arsène Wenger’s ‘Invincibles’ of the early 2000’s, Arsenal’s recent loss to Brighton was their fourth in a row, capping a miserable week in which they fell twice to Manchester City in consecutive 3–0 defeats.
A spirited Europa League performance against AC Milan and a routine home win against Watford has momentarily stopped the rot, but with Arsenal now 30 points behind league leaders City, and 12 points behind their top four rivals, there is, yet again, substantial pressure on Wenger to resign.
As another season of disappointment draws to a close for Arsenal, let’s explore some of the key drivers contributing to their poor results, and how the club can improve moving forward.
A sharp decline
After delivering three Premier League titles and four FA Cup trophies in his first eight years at the helm, Wenger established Arsenal as a perennial threat in the EPL. This was followed by a period of relative stability, where the club continually mounted a challenge for the league title, while performing well enough in the early stages of the Champions League.
However, despite another three FA Cup trophies and a second place finish in 2015–16, Arsenal were considered to be under-performing, especially against the backdrop of their move to the Emirates, purposed to ‘help keep the club at the top’.
After an alarming 5th place finish in 2016–17, Arsenal missed out on the Champions League tournament for the first time in 21 years. A lacklustre summer transfer window failed to address the major issues within the team, and as it stands Arsenal are on track to finish 6th, likely with their lowest points tally since the 1994–95 season. They have also only won one of nine matches against the ‘top six’ Premier League clubs — by comparison, Manchester City has won seven of its eight matches against the top six.
A brief look at Arsenal’s team statistics is all that is required to confirm the popular anecdote regarding Arsenal’s worsening defence. Consistently conceding less than 0.9 goals per match pre-2009, a sharp decline in recent seasons has seen this increase to almost 1.5 goals in 2017–18. Although the teams attack has not declined in the same manner, at the current rate of scoring and conceding, Arsenal will record their lowest goal difference since Wenger took over in 1996.
It’s easy to highlight the downward trend in Arsenal’s performance and results over the last decade, but in what areas are this once powerhouse club deficient and what’s does the future hold for the proud Gunners?
After more than 20 years deploying four defenders week in and week out, most recently with a 4–2–3–1 formation, Wenger famously switched to a back three in April 2017, which he has mostly continued to utilise this season.
It is generally agreed however, that this change was not out of a desire to be tactically innovative (unlike Conte and his title-winning Chelsea side) but rather as a response to the wing-back ‘trend’ and an attempt to arrest another mid-season form slump, which had become a permanent fixture on the Arsenal calendar.
The underlying issue behind this change of formation provides an appropriate starting point for dissecting Arsenal’s failures.
At first glance, the major issue plaguing Arsenal is clear — they are simply conceding too many goals.
In the past two seasons, the average number of goals conceded per match has increased such that they now possess a below average defensive record, shipping 1.37 goals per match compared to the league average of 1.35. If just away matches are considered, where Arsenal have won only three from 15 this season, this statistic blows out to 1.6 goals per match conceded.
Compounding the matter is the consistently improving defensive record of the top four clubs (in each respective season), currently conceding 40% less goals than Arsenal.
The graph below plots the actual goals conceded per season against the expected number of goals against (xGC), based on the Understat.com expected goals model.
As previously identified, Arsenal has progressively conceded more goals each season since 2015–16. What has changed this season is that the team is conceding more goals than expected. Opposition teams are also scoring goals at an improved rate, with 13% of shots resulting in goals conceded this season — compared to a more efficient 7% in the 2015–16 season.
Given that, on average, the top four teams are improving defensively and conceding less goals, it can be inferred that Arsenal’s problems in this area are a result of the teams overall defensive performance.
To chronicle this trend, we have looked at the total defensive contributions per match over the past four seasons. This statistic combines the total number of blocked shots, interceptions and clearances per match, and the data shows a clear decline since the 2014–15 season. This trend is also reflected in the number of aerial duels won per match, with a sharp decline between 2014–15 and 2015–16, perhaps mirroring Per Mertesacker’s transition out of the team.
However, perhaps the most damning statistic is the significant increase in individual errors leading to goals per match in the current season, almost equal to the last three seasons combined. This goes partway to explaining why the team has conceded more goals than expected this season, and sheds light on the key issue at Arsenal, the individual performances of a host of first team players.
To further explore the individual performances of some key players, the defensive contributions (per match) for Arsenal’s most-selected four defenders, in addition to the goalkeeper, was mapped over the past four seasons.
The first observation is the drop-off in Mertesacker’s contribution in 2015–16, and his justified replacement by Mustafi in the subsequent seasons. Of more concern is the slow decline in Koscielny’s effectiveness defensively, given he is the teams first-choice defender. While both Monreal and Bellerin had reduced contributions in 2016–17 and this season, this can partly be explained by Wenger’s tactical changes which have seen Bellerin deployed as a wing back, and Monreal more advanced on the left (and contributing more in attack).
The final observation from this data set is the player rating (developed by WhoScored.com), which although arbitrary, is a consistent and independent measure of historical individual player performance. The trend in average player rating of Arsenal’s top four defenders (and goalkeeper) is correlated with the teams worsening defensive performances.
Moving to Arsenal’s attack, the issues with the team are less existent. In terms of goals scored per match, Arsenal is well above the league average and has only slightly dropped off the pace relative to their top four rivals.
They are also scoring goals near or about equal to expected goals (xG), with no real discernible negative trend over the past four seasons with respect to goalscoring.
Despite this part of the team seemingly performing consistently, Wenger has brought in both Alexandre Lacazette and Pierre-Emerick Aubameyang to offset the recent departures of players such as Alexis Sanchez and Olivier Giroud. These acquisitions have further strengthened the Gunners frontline and based on current performances, scoring goals is not a primary concern for Wenger.
Team play in disarray?
So often the scapegoat for poor results, Arsenal’s midfield has seemingly been in transition for a number of years, with only Ozil and Alexis Sanchez (until his departure) arguably putting in consistent performances on a seasonal basis.
A quick look at some of the major ‘team play’ metrics illustrates the emphasis of passing within Wenger’s tactical approach. Arsenal is currently recording over 600 passes per match, second only to league front runners Manchester City. However, the percentage of key passes has bottomed out at 4.1% in the past two seasons, implying that despite the increased number of passes, they are not necessarily improving their output. Add to this a declining tackle count (and tackle success percentage) and it is easy to see why fans are often critical of Arsenals team play in general.
Given the role of a teams midfield is primarily connect and coordinate defensive and attacking phases, it is this part of the pitch which essentially governs a teams overall cohesiveness.
As the issues identified thus far are generally related to Arsenal’s defensive phases, we have taken a closer look at the players deployed in the defensive/holding midfield and central midfield roles. This is a position where Premier League titles have arguably been won or lost in recent times, with stellar performances by Kante (Leicester and Chelsea), Matic (Chelsea), and Fernandinho (Manchester City) proving crucial to their teams titles.
Over the past four seasons, Wenger has typically employed a double-pivot, or two holding/defensive midfielders, across both his 3–4–2–1 and 4–2–3–1 formations. Based on WhoScored.com lineup data, the following players have appeared in more than 10 matches as a holding midfielder:
To compare these players on an even playing field, Understat.com data from the past four seasons was utilised as it presents a number of key team play statistics on a ‘per 90 minutes’ basis. Of the wide array of statistics available, the following were selected to make comparisons:
Key passes per 90 minutes (KP90)— the key passes made by the player on a 90 minute basis
xGChain90 — defined as the xG of every possession the player is involved in over a 90 minute period.
xGBuildup90 metric, which is the total xG of every possession a player is involved in, excluding key passes and shots, per 90 minutes. As such, this metric isolates the buildup aspect from the final pass or goal scored.
Two things are immediately clear from this analysis — the current importance of Aaron Ramsey to the overall team cohesiveness (as evidenced in the recent win against AC Milan), and the impact of Santi Cazorla’s absence.
These inferences are clearly seen in the passmaps from the current season and the 2015–16 season. In the 2015–16 matches against Hull City and Dinamo Zagreb, Cazorla consistently has the most touches (dot size), and his key distribution channel is to either Ozil or Sanchez. Contrast this with some typical passmaps from this season, where the distribution to the attackers (Welbeck/ Mkhitaryan/Iwobi) has primarily come from the wing backs, and the holding midfielders are passing a higher proportion of passes backward or laterally.
So what should Arsenal do to arrest the decline which will likely see them finish below 5th on the Premier League table for the first time in over 20 years?
First and foremost, the club needs to provide a firm stance on Wenger, be it (a) an end to his term and progress to a new manager, or (b) the see out of his contract through to 2019.
In terms of the team’s performance on the pitch, the area in most need of improvement is Arsenal’s defence. Whether it is a change of coaching (i.e. the removal of Steve Bould) or player acquisitions, it is evident that change is required. Attention needs to be given to succession planning for Koscielny, with his defensive contributions and player ratings slowly in decline.
The second key issue identified was the number of individual errors leading to goals, particularly in the current season. Whether this is a flow-on effect of reduced team cohesiveness, low confidence, or simply poor performances, it needs to be addressed. The formation of solid partnerships, both in defence and midfield, has the potential to reduce the likelihood of errors. Looking back at past seasons, there was more certainty in the lineup, particularly in the centre of defence.
Finally, while Wenger has spent big on attacking signings like Lacazette and Aubameyang, they are unable to perform at their peak unless the buildup play behind them provides adequate service. In light of this, the holding midfield role is a key area of concern for Arsenal, particularly given their inability to effectively replace the injured Santi Cazorla. In order to compete with the ‘top six’ Premier League clubs for the title, a solid holding midfielder player/partnership is almost a certain requirement in 2019. The rumour mill will be full of names come transfer season, with Doucouré, Meyer, and Jankto all in the mix.
Whatever happens, it is clear that Arsenal are not performing at the level the fans of the club are accustomed to, and while they will likely string together a few wins to finish the 2017–18 season, clear change is needed if they wish to return to the upper echelon of the English Premier League.
N.B All data is from the sources mentioned, with additional chart input data from WhoScored.com and the official Premier League website.