Winning the duel
A deep look at how Arteta and Arsenal exploited fine margins to beat Man City — through duels, physicality, Rice, Saliba, Martinelli, Raya, pragmatism, adjustments, and much more
In contrast to the Community Shield match held two months prior, this game stood out not only for its inclusions, but for its omissions.
With Martinelli, Partey, and Stones in the process of recovering their full fitness, the early lineup graphics put them on the bench — and fully omitted the presence of Timber, Saka, De Bruyne, and Rodri. It was another reminder of the relentless demands of a football world that knows no offseason.
From Pep Guardiola’s perspective, that final exclusion may have loomed largest.
“The fact we didn’t have Rodri, I wanted to put more protection with the ball, players who are really good with the ball,” he said.
Pep’s quote nodded towards a “defend-through-possession” strategy that has underlined his career. It should have been no surprise when a top-5 on-ball player in the world, Bernardo Silva — note: I just typed that without fully vetting my thoughts on the matter, but it certainly feels a bit true — was chosen as his six, flanked by the two other best ball-players he could have chosen in Rico Lewis and Mateo Kovačić.
This was a flooded, heavily reinforced midfield that looked for control above all else. In build-up, it looked like a 4-3-2-1, sans actual wingers. Álvarez and Foden tucked inside as advanced midfielders/forwards:
That last bit — “sans actual wingers” — would prove significant throughout the proceedings, as we’ll cover later.
If Pep’s first objective was to offset the loss of Rodri with a heavily reinforced midfield, his second objective was to figure out ways to feed Haaland the ball.
This was going to prove tricky because he was set to face a player who is, ahem, a little better than Rob Holding, and is firmly placing himself at least in the conversation as the best defender in the world. His name is William Alain André Gabriel Saliba, and he’s one of the handful of players on earth who can rival Haaland for size, speed, and strength.
So if your second objective is to “get Haaland the ball,” you must figure out how to decouple Haaland from Saliba.
This is where Pep likely understood that the two main goals of his gameplan could be achieved simultaneously: by overloading the midfield with wingery midfielders, Arsenal’s full-backs would pinch inwards — and then the centre-backs (including … Saliba!) would have to cover the vacated zone behind them.
Here, right after the initial screenshot, Gvardiol makes a run in behind. Jesus could have tracked a little better, and White was occupied in the middle:
…and now you’ve got Saliba covering for White and Jesus. Haaland is free from his grips, running centrally:
This play is indicative of some recent trends at the top level: a centre-back (who often plays full-back) passed … to a centre-back playing full-back … who was … exploiting the space left by a centre-back playing full-back … who was then covered by a centre-back. But who’s counting?
This chance is snuffed out, as all of these ultimately would be. Haaland had zero shots on the day, and Saliba played a huge role in that — but so did the covering runs of his batterymate (Gabriel) and some smart rotations by the likes of Jorginho and Rice.
Pep’s gameplan proved schematically effective in two aims — Gvardiol was free, and Haaland was free of Saliba — but it was still on the players to turn that into goals.
👉 Arteta’s pragmatism
Arteta had some decisions to make of his own.
Prior to this one, we discussed possible lineup configurations:
For all our bullshit, it feels like the fans are in general agreement about the approach against Man City: if possible, a Rice/Partey double-pivot, with Havertz up front and Jesus at right-wing, assuming Saka is out. That all sounds good by me. The most interesting choice may be at left-back, where Arteta can go for a progresser (Zinchenko) or defender (Kiwior, Tomiyasu). If a double-pivot is chosen, I’d probably lean toward the latter. Things will change based on injury updates.
That’s not where we started, but it’s ultimately where we wound up.
But more interesting was how Arteta chose to deploy pressure. When it comes to “when” and “how” to shape up against the highest-tier teams, Arteta is learning his lessons.
Carrying the scars of last year, he came without the intent to onslaught his opponent with pressure from the opening whistle this time.
It was a more pragmatic approach: what was essentially a “high block,” somewhat pressurey, but ultimately intent on not opening up lanes through the middle of the pitch.
The Athletic has a great article up about the rise of midfield ‘no man’s land,’ which is one of the more interesting developments in the Premier League (and elsewhere) over the last year or two, and something we’ve covered ad nauseam in these parts — including just last week, as we talked about Lens.
The short version is that teams (like Brighton) are separating their attacking line and their build-up line by great distances, and then seeking to exploit the space in-between through carrying and creativity. This makes dribbling 10’s more free to roam; it also makes slower 6’s (like Casemiro) more exposed.
When Arsenal are out-of-possession, this space has proven difficult to defend against Man City:
…and against other teams, it’d been tough on Jorginho in particular, for obvious reasons:
The way to counteract these opportunities is to stick together. By pressing a little less aggressively, the gaps are less apparent, the middle of the pitch is more jam-packed, and Jorginho has to cover less space. This is what we saw in the Community Shield from Arteta, and what we saw again (at least to start) on Sunday.
This would also hopefully help the team settle in and avoid early mistakes or concessions. Those little mistakes and concessions almost happened, but … didn’t. Football is like that.
As you’ll see below, there was a clear pattern of ramping up the high press as the game progressed:
But the two main phases (in and out-of-possession) are never inseparable, and two other factors would cast a long shadow: first, the presence of Raya as an extra outfield player — and second, Chekhov’s Gun, those tucked-in wingers that Pep was using.
👉 What to make of Raya
We’ll cover Raya first. In my rather laudatory scouting report of the keeper written before he was signed, and prior to the talks of aggressive rotations, pure-depth reinforcements, and goalkeeper subs, I suggested that we lean more in the direction of Occam's Razor:
That’s why I’d err to the simplest explanation. I think they’re interested in Raya because they think he’s a level better than Ramsdale, as I do, and they had expected his ultimate sale price to be higher. When it actually started looking more viable, they started moving.
You may think to yourself: this still seems like a lot of money for marginal gains.
I consider him more of a genuine upgrade, but even so: marginal gains have compounding effects that reverberate throughout the pitch. At this level, marginal gains are all you have left.
But I was also raised the possibility of a phasing-in period:
Goalkeeping is communication by another name. Passing, scheme, shot-blocking, corners, free-kicks: it’s all a team sport. Any new keeper carries some risk of settling in. Brentford’s system is not always similar to Arsenal’s, and there could be some awkward moments that lead to significant questions.
We certainly saw some of that in the first half. There were a couple (uncharacteristic) actual misfires but more than that, it was about collaboration with his new teammates in a high-pressure situation. Balls were delivered a fraction late, or in a slightly different spot than expectation, in both directions.
Arteta put the blame on himself:
It’s my fault, all my fault. They can pull me up on it because I asked him to do that, especially against this team, you just have to do other things and you’ll get in big trouble. I think he was excellent, the way he dominated his box, the way he came out for crosses and set-pieces, the height that he played at. He’s got big ones, because the crowd go like this with the players. I’ve seen it, the players start to kick balls everywhere and I said to him, "you don’t do that, make sure you don’t do it, and he didn’t do it." At the end, he got rewarded because the team started the game that we wanted to play much better, so a big compliment to him.
But there is a wider picture in play. From a shot-stopping perspective, Raya has faced higher-quality shots so far: his “post-shot XG” is 4.7 and he’s given up 4 goals, whereas Ramsdale’s PSxG is 2.2 and he gave up 5. But that’s all subject to variance and sample size qualifiers.
The more significant change is how Raya fits into the rapidly-changing world of keeperdom. In the same amount of minutes, Raya has attempted 160 more passes than Ramsdale.
Ben Griffis, a data expert and champion of leagues outside of the top-5, has consistently rang the bell for all the innovations happening elsewhere. He sent me down a wormhole, and before I knew it, I was watching Hamburger SV, who currently dominate possession (~60%) and lead the 2.Bundesliga in xG.
Their keeper, Daniel Heuer Fernandes, is a card-carrying member of the build-up squad. Not in spurts, but as a rule. Lookee:
Here’s a look at Fernandes’ received passes in a recent game:
This is happening throughout the world in various pockets. There’s a simple reason behind it: pushing a keeper forward may not present as much risk as commonly thought, and it helps fling an extra attacker forward.
Raya wound up leading the team in passing progression, gaining over 5 fields of distance:
Back to Sunday.
👉 Bonus attackers
Because Arteta was using the double-pivot in Jorginho and Rice, and because Raya is like another CB, the middle was already fortified, so the full-backs (White and Zinchenko) could be used more expansively up the pitch — keeping Álvarez and Foden away from any countering opportunity, but also turning the front-5 into a front-6. Here’s a look at the play where Kovačić should have been sent off:
In truth, imbuing all of Arsenal’s moves was a lack of fear of the Man City wingers as an outlet threat. There is no love lost between Ben White and Phil Foden, and in fact, the former’s positioning throughout the game was a form of disrespect in itself.
This is probably not the heatmap you would expect from a converted-CB playing against Man City:
As we’ll cover later, he often just lined straight up as a winger next to Jesus, providing an extra man to the attack. Arsenal were highly focused on exploiting the Foden side:
Which, among other things, led to moments like this — where Nketiah, White, and Jesus all invaded the box at once, and an Ødegaard pull-back almost came off:
The second half brought a new energy to the proceedings, thanks in no small part to a half-time substitution.
👉 The contagious energy of Martinelli
Martinelli was shot out of a cannon early. It was one of the first things that Guardiola mentioned in his post-match press conference.
“We started really well in the first minutes, they started really well in the second half,” said Guardiola. “With Martinelli they increased the rhythm.”
His presence led to chances immediately, with Rice kicking off this snap attack with his trademark knee-tackle. One of the nice surprises of watching all that Rice tape for the scouting reports was learning how handy he is at pumping in through-balls with his left foot:
And that was the real difference for me in the second half: not just the introduction of Martinelli and his infectiously pent-up energy, but a clear decision by Arteta to dial up the pressure.
When pulling the stats to compare the halves, two things stuck out: Arsenal doubled their rate of high pressure, and also doubled their number of defensive duels from 19 to 40:
(By the way, in case this is news to you: data providers calculate possession differently. Some do it by a pure ratio of number of passes, some by actual time spent with the ball. This is the second kind.)
That didn’t require a rewatch to see. Arsenal were going all the fuck out on the press, and all the fuck out in the duel.
It wasn’t just Rice, Gabriel, and Saliba. The entire team would just swarm, and swarm, and swarm. More than anything else, I think it was this relentless pressure that ultimately picked the lock.
👉 Rice’s second-half role
But Rice relished in the team’s newfound espirit. Just as he was at his commanding best in the waning minutes at Lens, trying to will the team to victory, he was again expansive — not just in his play, but in his movement. After the break, he went 20/20 passing (100%), 4/4 in defensive duels (100%), and 9/11 in duels overall (82%).
Now, this was a double-pivot role that is certainly not a like-for-like comparison with Xhaka or Havertz, aside from place in the block and initial starting position, but nonetheless — you can see how freewheeling his second-half play was through this map of all of his actions:
He could often be found within the pivot:
But he’d also swing up to the 8, and have Zinchenko drop in:
…but his most interesting (and encouraging) sequence was around 58’, when he carried up to the right-hand side, and just stayed over there for a while — fully knowing his place in the puzzle, and confident that his teammates would adjust. Others, like Jorginho and Zinchenko, seamlessly filled in.
It led to an overloaded right-hand side. With White again pushing all the way up to the wing, Jorginho could drop into the right-back spot, and dictate play from here without any pressure. A lot of chances, including the decisive one, came from this area:
And you can see the degree to which Arsenal were willing to commit players forward during this period. After another “White as Winger” carry forward through this overloaded right, the play moved to the left, and all the gang was there to receive a cross.
It was a six-man front, plus Zinchenko high up, against the reigning treble winners:
A couple more chances followed.
👉 The response of Man City
There were a few points of genuine Man City threat, as there always will be. Now that Arsenal had ramped up the press, the aforementioned “No Man’s Land” risked returning. Here you can see Rico Lewis clearing a lane by guiding Martinelli out of the middle, and pointing out Álvarez in space:
(As an aside, Lewis is one of those players with an overhead tactical camera implanted in his brain; that is an impressive 18-year-old.)
Alas, everything was covered, and the tide was turning, so Pep needed fresh ideas. What transpired was a fascinating back-and-forth between mentor and mentee.
Matheus Nunes, John Stones, and Jeremy Doku came on — which struck me as sensible. But Pep’s decision to leave Foden on — Foden, who had been targeted and overloaded for the first 70 minutes — would prove damaging.
At first, Doku came on as a typical right-winger, and looked to run straight at Zinchenko, who hasn’t looked at full-90 fitness much this year. Zinchenko, however, had his number, and dispossessed him in clear fashion. This led Doku to seek a switch with Foden.
During the build-up phase, at 72’, Doku ran across the formation and asked the latter to scram:
White went back to his usual role, and held his own.
👉 Arteta’s countermeasures
It was time for Arteta’s response. At this point, he called upon Partey, Havertz, and Tomiyasu, and the XI became the one we were hoping for at the outset. It is a big, imposing, athletic lineup, and it lived up to its billing.
Over on Twitter — no, I’m not saying X — we shared that in the Community Shield, Timber had beta-tested the Tomiyasu run that would pay such dividends on this day.
In so many words, this was completely part of the plan.
Arteta motioned to Tomiyasu to stick closely with Havertz, and when the time was right, the team pounced. A few factors coalesced into one: White, once again, made a run up the wing, which pushed the defenders back and gave the 6 (in this case, Partey) room to operate.
Havertz occupied CB’s and kept composure, and Foden was again targeted:
Goal. Jubilation. Limbs.
I’ve always said that we should just shoot it off their faces.
🔥 So, what did we learn?
For a neutral, it may be tempting to limit our takeaways from this one. Key players were out, the affair was cagey and close, and one deflected ball was the difference. An early Man City goal was certainly possible; meanwhile, the void of Rodri highlighted, just as Saliba’s did last year, that some players are almost impossible to replace.
But I’d suggest that it’s OK to infer a little more than that. Arsenal have now been Man City’s equal in consecutive contests, and there are tangible reasons for that shift. Arteta has gotten more pragmatic, his adjustments have proven successful, and the overall feel is more composed and mature. Arsenal have injuries of their own, and won this without perhaps their best player, all while starting their third-choice 6.
But really, it’s the new players, dummy. Arsenal’s additions — Havertz and Rice, plus Raya for this one — have proven decisive in both games. Arsenal are now able to field a giant, athletic, tireless lineup and one of the best keepers in the world at kicking it up to them. The duels changed the course of the game, but Raya’s long-passing to Havertz and Jesus sealed it, as well as increased sense of calmness throughout. Young players are getting a little older.
So, it’s close. And Arsenal still have plenty to iron out in terms of regular chance creation against the lower blocks. But they now appear to have a reliable mechanism for bringing the fight to top foes in the Champions League and Premier League alike, and that is no small thing.
Final thoughts
Not sure I did enough of a job of celebrating Martinelli there. He was incredible. Here’s what I wrote afterwards: “Martinelli's career is so exciting because you see how good he is already, but can still find ~5 areas where you just *know* he’s gonna get a lot better.”
It’s hard to see a double-pivot lineup like this scoring 4 or 5 goals against parked buses, but it’s also hard to see it conceding … like, ever. With Partey coming back to health, I’ll be curious to see how much the double-pivot comes into play versus middle-tier opponents; Ødegaard/Rice/Partey is just proper Big Boy stuff. For as good as Rice already looks at the 6, he may look even better when playing more free like this.
Jorginho was exactly what was needed here. It is a rare luxury to have him as depth. Aside from the Foden stuff, Pep’s big error was not forcing Jorginho to cover more ground; the overloaded midfield played to Jorginho’s communicative, brainy strengths. He’s a good defender in a game like this.
I am not dogmatic about Havertz. I think it’s fairly clear he should earn regular looks up-front against the Man Citys of the world. As a left-8, I remain patient; my argument would be that there’s no reason he can’t rotate in and out for now, and also do a lot of what we saw Sunday from that spot — have Jesus drop, go long to Havertz, get Martinelli and Co. to grab the loose balls, prosper. Oh, and pump some far-post crosses into him, please. But Sunday showed why his £65m transfer wasn’t as risky as it may seem: his floor scenario is as a 24-year-old rotational striker who can do that against Man City. There have been worse investments.
I think Jesus has earned the spot as the second-choice right-wing; I think others should perhaps get the look earlier on the left, though Jesus should start somewhere regardless. Without Jesus’ 1v1 abilities and smarts, things could have gotten really stagnant here, so he was huge — and especially so at coming down with the ball late. He could have had Rice on two inside cuts; Rice is looking good in those poacher situations.
Ødegaard got a lot of coverage from Pep and was a big reason that Kovačić felt like he was always behind in his reactions. All told, he did a good job of finding his place, but yeah, one of those shot selections was maddening. I like players who shoot with bravado, because that swaggery confidence can seep into all aspects of their game, but he maybe needs to turn it down a dial.
Nketiah is still quite good, and I still believe in our sneaky-good #9 depth for this season (Jesus, Havertz, Nketiah, Trossard, Martinelli). But like any player at this level, he relies on having a physical advantage over somebody, or otherwise being savvier than them. Against the very top sides, he can just face players who are as clever as him, but possess more physical gifts. He’ll clean up goals in more back-and-forth affairs against other sides, but it can be frustrating to not have a false-9 or a target man for Raya to hit in one like this. I’m still pro-Nketiah, but would have sided with a Havertz start here. Going conservative early paid dividends, however.
Trossard performed… about as you’d expect if he’s deployed as a left-winger against Kyle Walker. Those aren’t optimal conditions for him to succeed, and I’d look forward to Reiss grabbing more starts there, with Trossard nabbing more stuff inside. He still managed some good stuff on switches.
Two weeks, and another call-out for Gabriel’s ability in blocking shots. Saliba rightly deserves all the credit in the world, but man, Gabriel sure does too.
Despite all his exploits, White has a very specific thing he needs to work through: biting forward on through-balls behind him. It caught him against Aston Villa, and almost did here.
At the end, Saliba blasted a ball to the moon. I was able to count five-one-thousand before it hit the ground again.
That new keeper of ours? He kept a clean sheet without having to make a save against one of the more imposing clubs in history. These guys are #big and Arteta is #good at coaching out-of-possession. La Masia Moyes strikes again.
Will be back next week with something with a wider scope. Looking forward to that.
For now, I hope you enjoyed.
Send along to everyone you know.
…and happy grilling.
excellent as always BC, and as always, UTA
Raya: he’s Busquets with gloves on!