Arsenal 2021/22: How They Played in Possession

In two articles, Abhishek Sharma digs deep into Mikel Arteta’s style of play with Arsenal last season. This first part analyses their approach in possession


League position8th5th
Goals scored5561
Goals conceded3948
Goal difference1613

Arsenal’s second full season with Mikel Arteta was an improvement over their first. Fielding the youngest squad in the Premier League, they went from 8th on the table with no European football to 5th (and almost finishing in the top four). They accumulated eight more points and scored six more goals than the season before. All of these are good signs. On the defensive side of things, they’ve shipped nine more goals than their previous season but their underlying xG conceded has actually improved (more on that in part two).

If we take a look at the underlying numbers from both campaigns and some other select metrics indicating their overall game plan, the key things to note are:

  • There’s a marked improvement in their overall set-piece game
  • They’re pressing better (more often and with more success)
  • Their possession game has changed significantly

We will look at the first and second of these features in part two of this piece, but let’s focus on how Arsenal changed their in-possession style of play.



General Structure, Changes, and the Knock-on Effect

Arsenal’s overall possession game has been one of the aspects that has seen the most improvement. Arteta’s Arsenal were always a possession-oriented side. Where they had trouble, for large parts of Arsenal’s 2020/21 season, particularly the first half, was consistently creating chances from the possession they had. Against mid-table or lower sides, it wasn’t rare to see them having the ball a lot but their overall attack was slow and predictable which meant once the opposition team had set up in their defensive structure, Arsenal had a difficult time breaking through. 

With the introduction of attacking midfielders like Emile Smith Rowe, Martin Ødegaard, and a change in formation mid-season from a 3-4-3 to a 4-2-3-1, some of the groundwork for an uptick in chance-creation was already laid but the entire 2021/22 season was when Arsenal really found more consistency in their attacking phases. As a result, now they are a much better sight in possession – not only managing to keep the ball for long spells in the opposition half but also ensuring that it is in the high-value areas that they want so that they are consistently threatening to score. 

So how exactly have they managed to refine their possession game? Here are some key metrics:

Passes per game468.5423.9
Pass Completion %83.6%82.6%
Possession %53.8%53.2%
Open-play xT1.211.41
Open-play xG1.291.42
Passes into Penalty Area8.959.16

For one, they’re passing the ball around a lot less now. On average, they are attempting and completing roughly 45 passes fewer per match now, despite a similar possession percentage compared to last season. The passing sequences seem much crisper and more effective. This is reflected in the fact that despite not keeping the ball as much, they’re actually creating more “threat” than before. Their overall open-play xT accumulated from passes and carries has improved from 1.21 per game to 1.4 per game. Remember those repetitive and frustrating U-shaped passing patterns which seemed to be going nowhere? Arsenal are doing less of that now. The offence is now geared towards constantly getting into high-value zones in possession and they take a less roundabout way to do that. 

This is further borne out by the fact that Arsenal’s speed of attack is virtually the same as in the previous season (1.19 m/s to 1.18 m/s) but their passes per sequence and number of sequences with 10+ passes has decreased. Taking a look at the positional breakdown of where these passes have been cut out from makes things clearer. 

There’s a couple interesting things to note here. Firstly, Arsenal are spending less time in their own half now. They’re attempting less passes inside their own half. 

Part of this is explained by the natural evolution of the team under a new manager. When Arteta came in, he had to ‘teach’ the team to play the way he wanted them to, which included building out from the back and maintaining control of the ball (especially below the halfway line). Naturally, as the team becomes better at learning those new patterns of play, they don’t have to keep doing it as much anymore. As a result, they can focus on sustaining attacks instead of building it.  

Secondly and more interestingly, the overall number of passes received in the final third has gone up (bear in mind that this is despite their overall number of passes having gone down). This increase in activity in the advanced areas is an encouraging sign which shows that Arsenal are now much better at keeping the ball in more threatening areas. In fact, Arsenal were receiving roughly 19 more passes per match compared to 20/21 in the central areas of the final third just in front of the opposition goal. 

The typical Arsenal set up is a 4-2-3-1. Arteta has always been reluctant to push up both fullbacks at the same time so usually we’d see one of the fullbacks (usually the right-back Takehiro Tomiyasu or Cedric Soares) tucking inside to create a back-three in possession and the other one push up forward to attack. This also seems to complement the skills of the respective fullbacks: Tomiyasu is adept at playing as a right-sided centre-back while Kieran Tierney or Nuno Tavares on the left are much better attacking threats than Cedric Soares (possessing a better change of pace, 1v1 ability, and a better final ball). This contrast in roles between the fullbacks is made clearer by the graphic below.

The redder zones represent the denser regions of touches. Notice how lopsided they are

One of the other marked changes had been in the midfield dynamics and the use of Granit Xhaka in a slightly forward role. He’s very adept at dropping into the backline and progressing the ball from a deeper position during early build-up phases and that’s the role he has been used in during most of his time at Arsenal. But often this season, he could be seen taking up advanced positions in the left half-space during build-up. The midfield ended up resembling more of a 4-3-3 when that happened (with Ødegaard in the right half-space and Thomas Partey/Mohammed Elneny as the deepest midfielder). 

It is worth noting that Xhaka had to play left-back in a few matches in 2020/21 when Tierney was out injured which might explain why some of the denser touch regions in the deeper areas but nonetheless, his change of role has still been noteworthy. 

Xhaka – advanced positions vs Everton (H) and vs Wolves (H)

Xhaka still drops deep when the opposition are pressing higher up. A notable example is the home match against Manchester United. However, against teams that sit deeper in a low block, he pushes up to give the opposition midfielders another body to worry about behind them. This also helps get in a better position to facilitate the attack. 

Manchester United (H). Manchester United are pressing high and hence Xhaka assumes his usual role in a tight double pivot with Elneny

I’ve mentioned this earlier but breaking down low-blocks had been a big headache for Arteta in his first season. Often the reason was Arsenal not having enough players showing in between lines centrally. The midfielders’ spacing would either be too disjointed (essentially rendered them inaccessible to each other in possession) or too flat (such that they could not access high-value zones). 

Giving Xhaka the licence to drift forwards in attacking phases (and worry a little less about defence) has helped in that regards as he is able to do a number of things from that position including, but not limited to,

  • early crosses to the far post,
  • support Tierney and Gabriel Martinelli/Smith Rowe with quick combinations,
  • circulate possession on the wings or switch play to other side,
  • get into pockets to receive the ball (either from deep or from the flanks) and further pin teams back.

It is also important to note that while Xhaka has, by-and-large, been reliable from the advanced role, he really isn’t a typical number eight. This is often reflected in his reluctance in making deep supporting runs beyond the last line. He also doesn’t possess the nimbleness of touch or quick feet to consistently beat players when facing them head on and then carrying the ball forwards. 

Some of these “inabilities” can just be tactical instructions and others can be worked around. Working around them is also justifiable given the other things he brings to the teams – being a natural leader to the squad, possessing a remarkably high work rate, and rarely ever being injured, and being versatile enough to play multiple positions. 

It’s also worth noting that Arsenal have signed Fabio Viera from Porto ,who is a nice addition to the midfield profile. 

On the other side of the midfield, Ødegaard runs the show. He has been terrific as part of the new-look Arsenal midfield. On paper he starts as the number ten but he often takes up positions in the right half-space like an interior midfielder.  His natural game is very dynamic: he likes to get on the ball often during build-up phases, and control possession in advanced areas. 

I mentioned earlier Arsenal’s new-found ability to get into high-value zones. A big part of the reason for that is Ødegaard. The positions he takes during attacking phases allows him to receive in space in between the lines. Under pressure in advanced areas, his first intention isn’t always to release the ball but instead, he uses his touch and ball manipulation to hold on to the ball and get out tricky situations, before reassessing the options available to him. This allows Arsenal to retain possession in the final third and build sustained attacks. 

He’s also formed a very exciting connection with Bukayo Saka on the right. In fact, the two of them have combined with each other to create a shot within the next action 34 times the past season! This is the highest in the league by some distance. For comparison, the next best combination in the league is Mohammed Salah and Trent Alexander-Arnold at 26. 

This brings us to the forwards. Arsenal have an exciting cohort of young forwards but the pick of the litter is Saka. The 20-year old has been incredible this past season, finishing with 2978 minutes played, 11 goals and seven assists from the right wing position. His importance to Arsenal is such that often the entire mechanics of the offence, particularly during transitions, are designed to allow him to receive the ball in the most optimal condition for him to take on a player. This involves the right-back inverting to pull in the opposition wide forward to create more space for Saka to stay wide and get into a 1v1 situation with his fullback. 

Manchester United (H). Cedric’s narrow position pulls Sancho infield which opens up the pass to Saka and gets him isolated with Telles

At his best, Saka can almost single-handedly bring the team into the final third using his progressive carrying. This is extremely useful during counter-attacking situations. He’s good at taking on a man once he’s facing him but also equally good at turning them using his body superbly if they get too tight. He prefers to cut inside on his left foot to either shoot or combine with his teammates, usually Ødegaard. And did I mention he can score as well? His tally of 11 goals in the league made him Arsenal’s top goalscorer last season.

He’s good at drawing fouls (fouled 2.3 times per 90), and has great decision-making in the final third. He is also very reliable defensively, tracking back responsibly to help out his fullback. This is crucial because Arteta’s system demands the wingers to defend. The defensive setup relies on being compact but having only two midfielders means they can’t always move over to cover the wide areas effectively. Hence the wingers have to get in line with the midfielders and plug the holes. 

Saka is also an able presser off-the-ball, although his pressing intelligence (when to curve his runs, which lanes to block, being aware of movement behind him) could use a bit of fine-tuning. 

The left-wing dynamic is slightly more straight-forward. It is between Gabriel Martinelli and Emile Smith Rowe to play as the left winger with Martinelli having started more games of the two. Neither of them are true wingers but both of them are effective in their own ways. Smith Rowe likes to drift into the halfspace and have the left-back provide the width while Martinelli likes to receive wide. 

Manchester United (H): ESR drifting infield and receiving from Elneny pulls Diogo Dalot inside and creates an enormous amount of space for Tavares to run into and get a shot off

Martinelli, while not as good a dribbling threat as Saka, can certainly hold his own against the average fullback in the league. He doesn’t have Saka’s silky touch and instead favours the byline more, choosing to use his acceleration and quick change of direction to run past the fullback and pull in a cut-back. 

Smith Rowe is a slightly more tricky customer: a lot of his game is based on subtle off-ball movements. In possession, he likes to occupy spaces in the pockets between the midfield and defensive lines (especially when Xhaka is deeper). From here he can either receive laterally from the wide fullback and turn to face the field or choose to make in-to-out runs off the shoulder of the fullback in the space behind. One of his main strengths is to time runs into the box during transitions to run on to lay-offs and get a shot off.  

Notice how Saka and Martinelli’s heatmaps are almost mirror images of each other as this reflects both of their tendencies to receive wide. Smith Rowe’s however more resembles an attacking midfielder

The position Arsenal struggled most to get right the past season was definitely striker. Pierre-Emerick Aubameyang, Alexandre Lacazette, and Eddie Nketiah were the three strikers in the squad at the beginning of the season. Aubameyang ended up leaving in the winter for Barcelona, and Nketiah was inexplicably ignored for almost the entire season leaving Lacazette to shoulder most of the strike burden for the entire season. The French international has not been without his own issues either and all-in-all, Arteta never really settled on the one consistent name at the top of his team sheet. 

For a stretch of games, some of Arsenal’s best performances came with Lacazette leading the line. However, the trouble with Lacazette is that he is not a very prolific goalscorer. He ended the season with only 2 open-play goals. His underlying numbers aren’t any better as he only managed to generate 0.28 non-penalty xG per 90. His main strengths were dropping deep to offer passing options inside the opposition block, linking play, and progressing the ball out wide. But not having a striker who lacked the necessary box presence or off-ball movement on the last line severely limited several avenues of chance creations for Arsenal. It was clear that Arsenal needed a striker with a much more well-rounded profile.

Obviously, Arsenal have since signed Gabriel Jesus and tactically, this makes a lot of sense as he can do a lot, if not all, of the above. The signs from pre-season, with the usual sensible caveats around pre-season, are extremely promising. Saka, in particular, seems to have formed an excellent understanding with Gabriel.

One of the summer signings was Aaron Ramsdale, who’s had a significant impact since coming on. He’s contributed majorly in improving Arsenal’s game in a number of departments such as the pressing as well as defensive set-pieces but let’s talk start with his passing game here. 

Ramsdale has been an upgrade on Bernd Leno with respect to on-ball ability. I’ve mentioned how Arsenal have become a little more direct in their approach – keeping the ball for shorter durations in their own third. Ramsdale’s superior distribution allows Arsenal to go long from goal kicks and still retain the ball in the middle third. Arsenal are attempting more long goal kicks than in 20/21 but with a similar success rate. His wide passing range means he can come up with more solutions to get the team out of pressure, whether it be chipped passes into the fullbacks or hard punched balls into the midfield, and turn defence into attack.

Indeed, the percentage proportion of Arsenal’s possession sequences reaching the final third with Ramsdale involved is 11%, while Leno’s was 8%.  

Chance Creation

All of the improvements in possession have had a direct carryover onto the way Arsenal have been creating chances. They have become less reliant on crosses this season. In 20/21, Arsenal were attempting 13.7 crosses per match, which was 3rd highest in the league. This season it is down to 11.4, which is the 10th in the league. 

It is also important to note that while Arsenal did cross more in the previous season, they weren’t very good at it. It was often a symptom of not having any alternate ways to create chances rather than a system built around high-quality crosses as a means of chance-creation. This is better explained by this chart from Ashwin Raman’s excellent piece on Arsenal’s toothless attack.

And this is what the picture looks like using 2021/22’s numbers –

In 2021/22, Arsenal were actually last in the league at the proportion of shots they generated from crosses. However, they’re also not as dependent on crosses as before so it’s not entirely a bad thing.

Part of the reason behind this shift in the crossing tendency is because Arsenal lacked a striker who can be a focal point for crosses in the box. None of Aubameyang, Lacazette, or Nketiahs’ aerial games in the box was/is anything to write home about. I suspect, if Arsenal were to get a striker who could fit that profile, we’d naturally see a more concerted effort to create from crosses. 

The plot above shows the drastic change in chance creation trends for Arsenal in the two seasons. They’ve moved from a chance creation mechanism which was based on wide attacks (particularly from the left hand side) to creating shots from quick combinations in the central areas. In fact, Arsenal are the highest in the league at creating chances from the central third of the pitch with 46% of their chances coming from there. In 2020/21, they were second-last in the league from that exact same third, marking a massive shift. So what’s changed?

If you had to describe it in one word, it would be Ødegaard. The Norwegian has created over 22% of Arsenal’s chances by himself! His game in the final third is based on finding pockets of spaces in the central regions to slip through balls, set up layoffs or drop deep in the right halfspace to curl crosses to the far post for someone to run onto. As can be seen in the map above, the areas he likes to set up shots from (or take them himself) are exactly the areas Arsenal were struggling to get into in their previous season. 

“…Martin is an unbelievable player. For this goal, I know Martin is always there in that position in training. He always says, ‘When you go wide, I’ll be at the centre of the box’…”

Gabriel Martinelli, My Game in My Words

The following viz illustrates some common patterns in the Gunners’ attacks and how it has evolved over the past two campaigns. 

I basically took all of their sequences resulting in a shot and with a minimum of eight passes, and clustered them based on their order leading up to the shot. We can see how Arsenal have begun regularly to get into areas to combine, which were pretty barren in the first season. Note that while K-Means Clustering is a great way to visually check out passing patterns in football, it does have quite a few limitations (lack of context, noise within clusters, geometric limitations, etc.). To work around some of them, I took clusters ending in zone 14 (at least 80% of passes ending in zone 14) and ranked teams according to their respective z-scores in only those clusters. Arsenal were second in the league, behind only Manchester City (to learn more about advanced applications of K-Means to football passes, read this piece). 

Another mechanism for chance-creation is cutbacks. A lot of Arsenal’s attacks are down the right channel . The focus in these situations is often to combine quickly and get in behind defences rather than cross from deep. This is also reflected in the numbers. In 21/22, Arsenal were third in the league at successful cutbacks at 0.57 per match, behind only Manchester City and Liverpool. The players who ended the season with the highest number of cutbacks from the squad were Martinelli and Saka with eight and six respectively. 

Finally, Arsenal’s attacking transitions have also looked much better and have become a genuine threat. Here’s a breakdown of teams’ game speed tendencies as well as effectiveness.

Although, as seen from the two graphs above, Arsenal are attempting a lower proportion of their shots from counter-attacks, the actual number is the same due to the overall shot count per match being higher. The more significant difference however, is in their efficiency. In 2020/21, only 11% of their final third entries from counter-attacks resulted in a shot but that number has improved to 27% in the 2021/22 season (third highest in the league). This is also reflected in the final output. In 2020/21, Arsenal took 10 shots from counters all season of which only one resulted in a goal. In 2021/22, the number of shots from counters almost doubled to 19 and a total of six goals (three from Smith-Rowe, two from Saka, and one from Martinelli). 

If the opposition have overcommitted, Arsenal like to take advantage of the spaces and try to break quickly. Ødegaard takes up positions behind the opposition midfielders and tries to feed one of the runners ahead of him, typically Saka or Martinelli. Both of them try to stay wide for as long as possible so that they can either time a run behind their fullback as the ball is released or receive the ball and square up to them in an advantageous situation, such as where they are gathering momentum and the defender is back pedalling towards his own goal. 

Arsenal’s second goal in the away match against Chelsea was a good example of what happens when Arsenal’s transitions go right. Some key things to note: 

  • Odegaard’s position in space behind Kante (in still #1). He is in an ideal position to receive on the half-turn and face either side of the field. 
  • Saka staying wide, and drawing Malang Sarr out, and stretching the Chelsea backline as much as possible to create space infield.
  • Saka’s movement allows Odegaard to move infield after playing the pass and occupy a very dangerous position (in still 3).
  • ESR timing his run to meet Odegaard’s lay-off perfectly and get his shot away.

It’s important to note that while Arsenal have improved with their overall chance creation, it has been in spite of their striker woes not because of it. Arsenal’s overall shot volume has gone up but their actual shot quality has dropped significantly. There’s a strong argument that getting a striker with a good box presence and more striker-y instincts would perhaps be the most straightforward way to fix this. It seems like Gabriel Jesus could be the answer to this, and make Arsenal even more dangerous.

In part two, we will look at how Arsenal have changed off-the-ball, and consider a few other elements of how Mikel Arteta’s side are evolving.

At Analytics FC, we provide software and data services to entities within football looking to realise the gains possible from analytical thinking.

Find out more about us, or get in touch if you have a question!

News, straight to your inbox

Provide your email address to subscribe and get email updates