On the 31st of March, Arsenal announced manager Joe Montemurro would leave the club at the end of the season. The news came as a huge shock, not least because it was released just before a Champions League quarter-final second leg between Chelsea and Wolfsburg. Though murmurs of discontent for Montemurro and Arsenal’s performances, particularly in the big games, could be found, there wasn’t much to suggest there’d be a change of coach on the cards.
IIn the final few months of the season, Arsenal’s future became hard to predict. With Montemurro’s departure, little news of contract renewals for players like Leah Williamson and Vivianne Miedema, plus a disappointing season, there were worries Arsenal could lose their key players and find themselves falling behind their big-spending rivals Chelsea and Manchester City. Though, if they’re also followers of the men’s team, it’d at least be a familiar storyline for Arsenal fans.
While there have been some changes, it’s not as bad as Arsenal might have feared. Williamson’s renewal played a big part in that, as did the signings of Mana Iwabuchi and Nikita Parris. Daniëlle van de Donk and Jill Roord have left, but the Iwabuchi, Parris and 22-year-old Frida Maanum should be enough to cover their absence. The Gunners have also added Simone Boye Sørensen for some defensive reinforcement. Miedema not signing a new contract yet is probably the only question mark hanging over the squad. Otherwise, there are reasons to be optimistic for Arsenal fans.
The Big Games
You knew it was coming. At the end of Montemurro’s reign, all the criticism was about their performances in the big matches. With it being such a big talking point around Arsenal and Montemurro, it’s best to get right into it. Last season Arsenal took just one point against Chelsea and Manchester City. To make matters worse, they also lost to Manchester United and had their best non-penalty expected goal difference against teams outside the big four1.
I’m not sure if you can post HTML tables on Substack, and I don’t know how to export a better quality image from R, but the poor quality screenshot of a table shows how Arsenal’s figures have changed against teams outside the top four over the past three seasons.
Arsenal’s expected goals figures against teams outside the big four improved last season.
|Season||npxG For||npxG Against|
But their figures against Chelsea and Manchester City were the worst they’ve been in recent seasons.
|Season||npxG For||npxG Against|
With how close the WSL is at the top, these four matches can decide whether a season is great or a disaster. It makes it hard to seek solace in expected goal figures, as some variance in games between the top six can be enough to sway the title race in a different direction. Arsenal’s record against Chelsea last season speaks to that. The combined xG was 2.1 – 2.4, but Chelsea took four points with a 1-1 draw and a 3-0 win. But, back when Arsenal won the title, a combined 2.6 – 2.0 xG against Chelsea turned into a 5-0 victory and a 2-1 loss.
If it’s hard to seek solace in expected goal figures for a couple of games a season, it’s not any easier when Arsenal post the figures they did against the two Manchester sides. The early goal and late red card at home to Manchester United could have some influence, but they created almost nothing against City and United this season. Against Manchester City, they had just 0.9 xG across two games and 0.8 against Manchester United.
More interesting is why Arsenal created so little in these matches. StatsBomb no longer releases their Open Data for the WSL, so I can’t try and slice things up and approach the question from a different angle. But Jonas Eidevall wasted no time in talking about the big matches. In his article on Arsenal.com he said the following:
But in order to win more games and to beat the top teams in the league, we also need to develop our way of playing against teams that are pressing us high. We need to be much more resilient against that, to have much more effective strategies to get out of those situations.
Arsenal had some odd stats related to this quote last season. I noticed it around February or March, but it was also pointed out in January by @Odriozolite. I wanted to wait until the end of the season before looking further. It could have been a quirk of schedule rather than a long-lasting trend, considering it was after Arsenal’s back-to-back defeats against Manchester City and Chelsea when I saw it. I didn’t want to write about it only for things to even out as they had a more favourable run-in.
By the end of the season, Arsenal’s touches in their defensive third weren’t up much on the previous year. However, the last two seasons were both much higher than the title-winning season. Even accounting for the differences in the number of total touches doesn’t make much difference. The percentage of touches in their defensive third also rose. In 2018/19, ~29% of their touches were in the defensive third, compared to ~34% in 2020/21. But, does this have to be bad?
The graph below shows the percentage of touches Arsenal had in their defensive third compared to the non-penalty expected goal difference.
The standout is how bad those performances against the Manchester clubs were, but with a similar percentage to the Chelsea away match, Arsenal had some dominant xG scorelines. What’s interesting is many of these matches came in the early stages of the season. The home matches against Reading and Tottenham, plus the away matches against West Ham and Brighton, came within the first five games of the season. Only the away match to Tottenham came in the second half of the season — though their figures weren’t as strong in that match.
With this in mind, it could help explain why Arsenal’s figures for touches in their defensive third were up so much around the time of the tweet by @Odriozolite. It’s interesting whether there was a conscious change in style for Arsenal midway through the season. In the opening stages, the wish to keep the ball deeper and draw the opposition in often felt deliberate.
Reading have an overload in midfield, they have an extra player because they play with a 442 diamond. At times they don’t allow you to play out because they can stay in between the lines of your build-up so they can get to both areas whether you play wide or if you play centrally. The idea was to drag as many of them as possible towards our defensive line and have different ways that we could play through and out. When you play with a diamond it can make you very narrow so we thought we would have the opportunity to play in and then go on the outside and progress quite quickly.
After the 5-0 win away to Brighton, @Stillberto asked about Arsenal inviting Brighton to press and Montemurro said:
When teams play us in a 442 we know they’re going to press us aggressively in the first phase but we want to keep making the distances longer between the first phase and the second phase where their middle four were. Once we were able to do that we were able to get through them quite easily. Our way is to keep finding the spaces to get into those areas, we knew there wouldn’t be much space between their midfield and their defensive line so we knew we had to create that space behind their midfield line.
We had a different setup against Spurs last time, we played with a lower midfield this time, we were able to combine with our back four in our build-up. So once they pressed us in that area, we were able to drag out their lines and we found space beyond and I think that’s where we hurt them a little bit more today. It was a different type of game to last time and a different approach from us too.
It all hints that Arsenal wanted to draw the opposition in to create space elsewhere. @Stillberto summed it up brilliantly when talking about what to look for from Arsenal this season by referring to it as a ‘snake charming’ style.
Arsenal often went backwards to go forwards again, they would keep the ball around the half-way line or even in their own half in order to tempt deep defensive blocks to edge away from their own area. Once Arsenal had successfully tempted the defence up the field, they would try to move quickly in behind them.
But, while a different approach could explain the change in defensive third touches for some matches, those four games against the Manchester clubs still stand out. This is where things diverge slightly, and whether it’s bad isn’t so subjective.
The below graph, showing field tilt against xG, mirrors much of the previous one and emphasises how much the Manchester teams managed to pin Arsenal back.
The matches against Reading and the home match against Tottenham are included, again showing they had strong xG figures despite a below-par field tilt, but the bottom left corner is more interesting. The United home match featured an early goal and second-half red card, which could warp the numbers somewhat, but the two away matches are worthy of further investigation.
Using WyScout figures, Arsenal had the 4th fewest losses in their defensive third last season with ~5 more per match than Chelsea and Manchester City. Removing games against the top four, they rise to 3rd and have just under three more than City and Chelsea. But against the top four, it’s a different story. Arsenal fall to having the 5th worst figures in the league. Chelsea, Manchester City and Manchester United all had ~20 losses of this type. Arsenal had ~31(!). Other teams may have a different approach, opting to go more direct so the losses will be further from goal, but it’s hard to see those Arsenal figures and think it’s anything other than an area with lots of improvement required.
To further explore these losses, I headed to some video. Manchester United away was the first match I wanted to investigate. Watching the clips, most of Arsenal’s losses or United’s recoveries came from crosses. On WyScout, Arsenal’s opponents attempted 4.5 crosses in the first half last season. United attempted 13. In the second half, things cooled down a bit, but I thought it was worth going through these crosses to see if there was anything of note. Substack isn’t modern enough to support moving pictures, so screenshots will have to do. (You can click them and open the original image in a new tab if they’re hard to see)
Manchester United (A) – Cross #1
The first cross sets the tone for lots of what’s going to come. Arsenal dealt with the first ball, but the opposition pressed to regain it afterwards. No one was around the edge of the area, with Arsenal’s central midfielders drawn to their right. The ball fell for Caitlin Foord, but she miscontrolled it. Batlle went to pressure her as the ball came down. The Spaniard was able to regain the ball and start another attack for her team.
Manchester United (A) – Cross #2
Not much of note for what we’re looking for, as Christen Press hit her volley over the bar, but the move is immediately after the first cross. Starting with Batlle, United moved the ball from right to left and created this chance.
Manchester United (A) – Cross #3
This instance was better for Arsenal, with Mead winning a free-kick and giving Arsenal the chance to have possession. Not bad for United either, with a free-kick on the edge of Arsenal’s box probably better to defend than a counter-attack.
Manchester United (A) – Cross #4
Similar to the above instance as Hanson crossed from the right. This time Gut cleared it out for a United throw.
Manchester United (A) – Cross #5
Tobin Heath played in a cross and Press hit her volley over the bar. No need for a screenshot as it’s not what we’re looking at.
Manchester United (A) – Cross #6
Press hit her cross out for goal kick. It came from an Arsenal throw, Katie McCabe threw the ball to Daniëlle van de Donk for the below situation:
Van de Donk lost it. Press was slid in down the right but hit a poor cross.
Manchester United (A) – Cross #7
Like the first example, United dispossess an Arsenal player near the area, and were able to pressure, win the ball back and create a chance from it.
Manchester United (A) – Cross #8
Arsenal blocked the cross before it reached the box. But twice Arsenal cleared it long, only for it to come back down the right flank each time. Eventually led to a United throw-in.
Manchester United (A) – Cross #9
There were a few stages with this one. It started being quite strong before falling into a similar category as some previous examples. It’s probably easier to talk through step-by-step than post the grid of screenshots.
Battle crossed from the right.
Viktoria Schnaderbeck and Lia Wälti helped clear it away.
Ladd wasn’t as tight to DVD, who was able to feed it into Beth Mead and made it look as though Arsenal could get the ball forward.
Instead, she was pushed wide and then back to Williamson.
Williamson, under pressure from Heath (who can be seen on the right of the above screenshot), hit it long. Ladd headed it down to Groenen.
Groenen was able to turn and play it to Heath near the edge of the area. Heath took a long shot that Williamson blocked.
From looking like they could counter, Arsenal went to Heath having the ball and moving towards goal in a few moments. To make matters worse, Arsenal gave the ball away after the blocked shot, giving United a chance to attack again.
Manchester United (A) – Cross #10 and #11
The next cross is best with video, as there’s a long sequence of play on either side of the cross. It’s uploaded on Streamable here. Arsenal had possession near their box and were trying to play through the United press. Groene dispossessed Wälti, allowing United to move into the box. Press crossed from the right, but it went over everyone to Heath on the left.
Galton hit another cross from the corner of the box, which DVD headed away. Foord collected the ball near the edge of the area, under pressure from both Hanson and Batlle, and couldn’t bring it down. Batlle regained possession for her side. Hanson’s return pass to Batlle was poor and Foord was able to bring the ball forward. Miedema dropped to receive the ball to feet and spread it wide to Mead. Again, Mead ended up wider than it looked like she would be — albeit because of a miscontrol this time.
Arsenal went back and started again. They couldn’t play through the press, but Williamson hit a long ball to Miedema, who came shorter again. Katie McCabe made a great run in the half-space, making Amy Turner backtrack and creating space for Miedema to bring the ball forward. It meant Batlle had to move inside to close Miedema down, leaving Foord in space out wide. Batlle got across to block Foord’s cross but, even though it contained the same problems with losing the ball following pressure from a cross, it was a better move forward from Arsenal.
Manchester United (A) – Cross #12 and #13
The last two crosses are back to familiar territory. Arsenal cleared it away from the box, only for a United player to step in and regain the ball in midfield. Ladd’s cross led to a header from Heath, but with no real threat.
From those thirteen crosses, how many did Arsenal turn into sustained possession or an attack of their own?
From my count, four aren’t relevant for the question. Crosses #2 and #5 led to shots, #6 went out for a goal kick, and #8 didn’t reach the box — but the aftermath didn’t reflect well on Arsenal.
Of the remaining nine, Arsenal don’t hold up well. There are maybe only two or three outcomes that don’t feel bad. Mead getting fouled by Groenen gave them a chance to have the ball and not concede anything following pressure. Cross #9 started better but ended up with a long-ball into midfield and Manchester United attacking again. Even when Arsenal got into the final third from #11, it was because of a sloppy pass by Hanson to Foord. Foord lost possession on the second ball from the cross.
Without slicing the data up and trying to find a way to measure this, it’s hard to make a big statement. But going through the crosses, it seems incredible how Arsenal couldn’t turn one instance of regaining the ball into something positive — even if it’s just winning a throw further up the pitch. It’s also hard to say whether it’s something systemic or personnel based. Is it do with how Arsenal position themselves in these situations? Or were individual players not strong or quick enough in duels? Of course, it’s not a choice. Both can play a part.
The next step was to look at another game to see if it wasn’t just a one-off. Away to Manchester City, Arsenal conceded 26 crosses compared to the 10.1 their opponents averaged last season. Arsenal did take an early lead that could affect both these figures and Arsenal’s approach, but it still isn’t good if they can’t transition out of these situations regardless of the game state. So, buckle in, now we’re going to go through those crosses.
Manchester City (A) – Cross #1
The first cross suggested the United game wasn’t a one-off. Both Lauren Hemp and Caroline Weir tried to bring the ball down but couldn’t. Roord gathered it and looked to move forward, but Sam Mewis tracked back and won the ball before taking a shot. In the third screenshot, there also looked to be little on for Roord going forward.
Manchester City (A) – Cross #2
This cross resulted in something better for Arsenal. It didn’t start well, with Georgia Stanway able to bring the ball down on the edge of the area, but Foord waited and managed to poke the ball away as Stanway brought it down. It ran through to Miedema, who turned Steph Houghton and moved her team up the pitch. It all came to nothing, as Alex Greenwood’s challenge rebounded off Miedema and ran through to Ellie Roebuck, but at least Arsenal went from back to front.
Manchester City (A) – Cross #3
This instance is a bit of a mixed bag. Arsenal weren’t dispossessed like in previous examples but were second to the loose ball. City were able to recover the ball and apply further pressure. McCabe defended well against Kelly, but it’s not an ideal situation.
Manchester City (A) – Cross #4
From the last screenshot, Kelly hits a cross straight into McCabe and then fouls her. Not a bad outcome for Arsenal, but not really what we’re trying to look at either.
Manchester City (A) – Cross #5
Like the United game, Arsenal dealt with the first ball, but City reacted quicker to the second and had two players pressuring Roord. They won the ball back and looked to attack again.
Manchester City (A) – Cross #6, #7 and #8
These crosses all came within quick succession and followed the above sequence. City recycled the ball on the left, eventually leading to Hemp crossing again. Stanway got ahead of Williamson to claim it at the near post and came out with the ball after a duel with Kim Little. Stanway then played it to the edge of the area for Lucy Bronze, who went wide to Kelly. Kelly’s cross went over everyone to Hemp on the opposite side. Hemp volleyed a cross in which Little sliced behind for a corner.
Not a great passage for Arsenal, but we’re mainly focusing on second and loose balls, so nothing too relevant here.
Manchester City (A) – Cross #9
Arsenal did better initially on this cross, with Roord bringing it out, but couldn’t deal with the press and ended up clearing it out of play for a City throw.
Manchester City (A) – Cross #10
Kim Little intercepted a low cross from Hemp and hit it out for a throw-in. Not really enough to warrant screenshotting it.
Manchester City (A) – Cross #11
This started very similar to lots of the bad examples. Arsenal dealt with the first ball, but their player is pressured by two on the second ball. Foord lost out, letting Kelly take the ball forward. In the end, Kelly lost possession and helped create a nice redemption arc for Foord.
Arsenal managed to turn the City defence, move up the field and even get into the box.
Manchester City (A) – Cross #12
Sam Mewis played a low cross into the box. Williamson poked it straight to Chloe Kelly. McCabe went out to deal with Kelly, everyone else prepared for the second cross. Kelly moved inside and hit a left-footed shot straight at the ‘keeper.
Not great from Arsenal, but a clearance falling to a City player feels different to Arsenal not handling the opposition press.
Manchester City (A) – Cross #13
Kim Little blocked Sam Mewis’ cross before the ball bounced to Foord. Foord tried the first time ball into Miedema, but Houghton got a toe to it. Roord collected the loose ball and went wide to Foord. Foord found herself in this position:
She went back to McCabe (just about pictured on the left), who hit it up the line off Kelly and won her side a throw-in.
Probably a bit better from Arsenal. They got to some second balls, didn’t get dispossessed or caught in transition and won a throw-in. Not bad for City either. They stopped a potential counter and could look to squeeze Arsenal in from the throw.
Manchester City (A) – Cross #14
Lydia Williams claimed Mewis’ low cross.
Manchester City (A) – Cross #15
Mewis’ cross flies past everyone.
Manchester City (A) – Cross #16
Hemp’s cross blocked for a corner.
Manchester City (A) – Cross #17
Hemp’s cross blocked for a corner.
Manchester City (A) – Cross #18
Kelly crossed to Ellen White, but she was offside.
Manchester City (A) – Cross #19
Houghton’s cross went out for a goalkick.
Manchester City (A) – Cross #20
Kelly’s cross went out for a goalkick.
Manchester City (A) – Cross #21
Kelly’s cross went out for a goalkick.
Manchester City (A) – Cross #22
Mewis’ cross was put behind for a corner.
Manchester City (A) – Cross #23
Cross following a short corner. Better from Arsenal, they looked to push out and won a throw-in around a third of the way up the pitch.
Manchester City (A) – Cross #24
Weir’s cross was straight at Williams.
Manchester City (A) – Cross #25
Beckie’s cross was blocked and went out for a corner.
Manchester City (A) – Cross #26
Beattie cut out Beckie’s cross and smashed it out for a throw-in. Eventually led to City’s winner.
The second half saw City’s crosses falter more often, but similarities to the United match stand. With how many weren’t relevant in the second half and how some came in quick succession in the first, it’s harder to gauge how Arsenal stack up with the United match. But there were only two occasions where Arsenal recovered the ball and moved into the opposition half.
Consulting my little spreadsheet, 15 crosses didn’t go to the ‘keeper or out of play. Out of those 15, McCabe won a free-kick and a throw-in, Miedema went on a solo run, and Foord played Miedema in behind. Otherwise, I think the other 11 all finished with City keeping possession. Not all of those were winning second balls, as some were following an overhit cross, but it doesn’t reflect well on Arsenal.
It’s hard to know how to judge without any reference point, but you imagine if Arsenal want to improve in their record against the top clubs, they need to find a way to better transition out of these situations. It’s almost as though when the tide turns, and they find themselves having to batten down the hatches, they struggle to turn things back in their favour. The good news for Arsenal fans is Eidevall addressed it so early. How he approaches teams with a high press will be worth keeping an eye on throughout the season.
Attack and the return of pressing
Further up the field, there’s less concern for Arsenal. Yes, they created almost nothing in some of the big games just mentioned, but it stems from being unable to get the ball forward. Once they get the ball forward, they’re a potent attacking outfit. A lot of this is thanks to Vivianne Miedema.
Miedema’s feats are known. She’s one of the best forwards in the world. There likely wouldn’t be too much argument if you said she’s the best. But it’s her supporting cast that I want to talk about.
The above graph, showing how Arsenal’s attackers non-penalty xG has changed since 2018/19, sums up the main point. Miedema had a dip in 2019/20 and, to make matters worse, no one stepped up to offer much help. In the title-winning season, both Nobbs and Van de Donk offered a goal threat — although Nobbs didn’t play many minutes as she got injured not long into the season. Mead’s xG went up in 2019/20, but not enough to replace what was lost by the dip from others.
In 2020/21 though, things were better for Arsenal, thanks to Caitlin Foord’s arrival. She arrived in January 2020 but didn’t get to make an impact in the league with the suspended season. She added a goal threat from out wide and filled a much-needed gap for Arsenal. Jordan Nobbs and Jill Roord also offered more in terms of goal threat — though it’s questionable whether it got the best out of Nobbs individually out wide, the added xG was beneficial.
These numbers can also be due to the roles played, particularly for the midfielders. Each of Roord, Nobbs and Van de Donk has spent time deeper and in more advanced positions.
Two related points pop up from the graph. The first is that Nikita Parris looks to be a good signing and, like Foord, should add more goal threat into the side. Miedema can be free to roam, knowing there are players like Foord and Parris who will thrive off the space she creates. But the other point concerns Beth Mead.
Despite coming to Arsenal as a #9, Mead hasn’t generated much xG in recent seasons. Last season her role looked to change to being an orthodox winger, staying wide on the right and not getting into goalscoring positions. Her actual numbers may not have been as strong as 2018/19, but her xA was better than the previous two seasons. From 0.39 per 90 in 2018/19 to 0.22 in 2019/20, her xA hit a high of 0.43 last season. Making the change more interesting is how so much of her attacking figures have dwindled.
Mead’s shot-creating actions for set-pieces fell from 1.48 in 2018/19 to 0.38 last season. Shot creating actions with live passes fell from 2.78 to 2.50. Her key passes, passes into the box and crosses into the box were all down last season from the title-winning season, yet her xA was up. She was creating less but creating better quality chances.
With the competition of Parris and Foord, it’ll be interesting to see how Eidevall uses Mead. She’s a player you feel is capable of generating more xG. She used to play as a forward and, without getting into the finishing skill debate, looks to have strong shooting technique. Mead also offers a good segue into talking about pressing.
Mead offers a lot out of possession. She made the 2nd most pressures per 90 for Arsenal last season, behind only Van de Donk. Only Wälti, Van de Donk and Schnaderbeck attempted more tackles than the winger too. Supporting the figures, when writing about three things to look for at Arsenal next season, @Stillberto says the following about Mead:
Beth Mead’s superpower is pressing opposing full-backs and regaining the ball high up the pitch. It’s one of the reasons that she remained a stock selection under Joe Montemurro, because she offered that ability to counterpress and create chances in a much different way to the other attackers.
With Eidevall talking about how he wants to press higher, Mead should continue to be a valuable player for Arsenal. Seeing if she becomes more of a goal threat of her own will be interesting to see. Her most productive season, albeit with lots of overperformance, was the title-winning season, which was also when Arsenal’s press looked to be higher up the pitch.
2019/20 was when Arsenal’s press looked most aggressive, at least looking at the percentage of touches that were pressured in the middle and attacking third. But the title-winning season saw the highest percentage of pressures coming in the final third. On WyScout, their PPDA (which isn’t great) has climbed every year. In Opta data, thanks to The Analyst website, their PPDA and high turnovers fell behind the other title challengers last season, with Chelsea making ~4 more high turnovers per game.
Last season also saw Arsenal’s opponents take the fewest touches per final third entry in the past three seasons. But, like with the attacking figures, Arsenal’s npxG conceded was better than the title-winning season for games outside the top three/four, so it’s hard to be too critical if what they changed led to the opposition not creating as much.
How a higher, more aggressive press is utilised to create more chances and catch opposition in transition will be more interesting than how it affects Arsenal’s defence. Under Montemurro, Arsenal used a high press and were fine at the back, but if Eidevall wants to introduce an aggressive press as a means to create more, Arsenal’s attackers could thrive.
I thought writing on Substack would encourage me to write little and often. Instead, I started rambling during this and went on much longer than intended. But, forgetting all of that, it’s a fun time to follow this Arsenal team. Eidevall looks to be coming in with a specific idea of what he wants and needs to do. They’ve lost some big players, but Mana Iwabuchi, Frida Maanum and Nikita Parris have the potential to add both more danger and excitement to the Arsenal attack.
Considering how strong Arsenal have been against the sides outside the top four, it also means Eidevall only needs to improve the side in six games for the team to make the progress required. In some ways, it’s much worse, as talked about with the variance earlier, but if Eidevall’s methods can improve Arsenal in those matches, it’ll look like a big step up.
Arsenal fans will get their answers pretty quick too. They face Chelsea on the opening day and Manchester City in their third match. You can already picture all the narratives should Arsenal lose those matches, talking about how it’s more of the same and no progress has been made. Whatever happens, it’ll be interesting to see how Arsenal change under Eidevall.
Like the last post, I originally posted this on Substack and copied it over to here, hence the moaning about Substack and lack of GIFs/videos. I’m not sure what I’m going to post on for the new season, I’m just experimenting with Substack at the minute.