I hope you’ve had the time of your life celebrating. In due time, I’ll share some reflections on the season writ large, as others already have so well. For the moment, there is still some minor business to handle.

Over the last week or so, I’ve been watching as much PSG as I can, poring over Newcastle, Bayern, Lens, Monaco, Chelsea, Liverpool, our tie last year, and Athletic Club, oh my. This was done with the hope of writing something interesting from an opposition analysis perspective, the kind of article I’ve done before.

But this time, I got by with a little help from my friends. I am first overjoyed to be joined by Jake, the Set Piece Sicko himself, of Pot Shot Podcast fame, who contributes a full set piece analysis below. His insights are detailed and invaluable. I have also been working alongside Bastian (@austriangooner1), one of the brightest analysts in the sphere, to try and spot vulnerabilities in PSG’s play. While Jake’s contributions are clearly demarcated in a specific section, Bastian’s and my work will blend. As a rule of thumb, if you like an insight, it was his, and if you thought Who is this moron?, it’s the moron you subscribed to.

Let’s see how we can win the fucking Champions League.

🔵 Matchup overview

Let’s get to work, starting with a basic comparison.

Some notes from that:

• Here’s a narrative violation: PSG play the slowest Champions League football (1.42 m/s direct speed), while Arsenal play the 6th-fastest (1.86). In terms of their relative passing sequences, Arsenal’s tempo is 31% faster than PSG’s.

• On that note, PSG rank second in a lot of territorial-domination, positional-play type stats, after either Bayern, Man City, or Barcelona. They log the second-highest possession (63%), the second-most build-up sequences of 10+ passes (18.88 per match), the second-most passes per sequence (5.04), the second-longest sequence time (14.19 seconds), and the second-most final-third passes per match (174.63).

• Arsenal concede 50% lower xG/match than PSG (0.84 vs 1.38).

• PSG overperform their xG by +0.78, which is the highest in the Champions League. They’ve been out-xG’d in their last three Champions League ties (Chelsea, Liverpool, Bayern). Defensively, Arsenal overperform xG against by 0.41, which is the best defensive “overperformance” in the Champions League. Both teams are operating in tails of their distributions.

• Their xG per shot (0.11), which is usually a great indicator of the overall quality of attacking play (and one of the best predictors of victory period), actually isn’t super high.

• In the Champions League, they send crosses in at below the median rate. That isn’t necessarily bad, as it’s more of a play-style choice.

• While Arsenal force more high turnovers, PSG convert high turnovers into goals at the highest rate in the Champions League (0.44 per 90).

Meanwhile, in Ligue 1, there’s what you’d expect: PSG have the highest line, they have the most possession, they face the lowest blocks, they limit opponents to the lowest build-up success in the league.

Let’s move on to the players.

🔵 The players

Here’s a look at how things may shape up.

We’ll get back into Arsenal selections later. But there are primarily two big questions for PSG, both surrounding their wonderkid (or, rather, one of them):

• Zaïre-Emery or Hakimi at right-back. This comes down to whether Hakimi is fit enough to start, and if so, how long he can go. He’s been out since Bayern with a thigh issue.

• Zaïre-Emery or Ruiz in midfield. Ruiz suffered a knee injury in January and was out for several months, while Zaïre-Emery has played the most minutes on the squad. This will be influenced by sharpness. Ruiz just played a full-90 on the 17th; if all else is held constant, then it’s a choice between the carrying and ridiculous engine of Zaïre-Emery versus some of the passing and savvy combination work of Ruiz. They look a little more vulnerable defensively with Ruiz right now, to my eyes.

This view from Transfermarkt will give us a sense of availability and selections in the league campaign:

That said, it’s only so trustworthy for our task. PSG rotate massively and hilariously. They don’t have a single attacker in the top 1,000 players in Europe for league minutes. As an example, Mohammed Kudus hasn’t played in five months … and has more league minutes than any PSG attacker. I saw Marquinhos make a European Team of the Season, which seemingly included domestic performances. He had 11 league starts.

(It is clearly the smart and advantageous thing for them to do. All I ask is that commentators mention it when they guffaw about their energy levels. And I also think it must be taken into account when giving out end-of-year awards.)

Even more, Ligue 1 rescheduled two league fixtures to give them more rest ahead of Champions League ties. One of the postponements was against their primary title rivals, who were only a point behind at the time.

My shameless hopium barges in to feebly ask: have they overdone it? They haven’t played in weeks, and there was plenty of rotation even then. In the meantime, they hosted a short friendly to try and maintain match sharpness. Could they play looser as a result?

Ah, probably not. But Arsenal have now rested, as well.

[edit: In the pre-match presser, our Starman has chimed in with some perspective.]

“A game like this is not going to be decided on minutes, it’s going to be decided on moments.”

Next, let’s look at the statistical profiles of key players (through Gradient’s advanced analytics and grades) and some notes I’ve taken on each.

I. The attack

👉 Ousmane Dembélé, forward

For each player, I’ll include some of their rankings in the advanced analytics, as well as some notes I’ve taken on approaching them. This is an opposition report, after all, so full priority should be given to nullifying opponent strengths: for each player, ask yourself, if I’m them, what kind of game do I want least?

We’ll start with the reigning Ballon d'Or winner.

Analytic strengths (via Gradient): finishing across types (pressured, off either foot, even his headed finishing rates well), through-ball passing, link play, defensive contribution from the front (dribble defending, 1v1 defending, tackle resistance), ball-carrying.

Vulnerabilities (via Gradient): more isolated dribbling, aerial duels, crossing under pressure, over-the-top passing. He likes getting touches and being involved, and can look to do it from outside the block in settled possession, in places where he isn’t nearly as dangerous.

Coaching points (our notes): don’t dive in, as his finishing beats you more than his dribble; track and communicate the runner who fills in when he vacates the nine. Track him everywhere in build-up, though you don’t necessarily have to deny him touches in wider or crowded areas in settled, advanced play.

👉 Khvicha Kvaratskhelia, winger

And here are some of his deeper on-ball grades as well:

Analytic strengths: tackle resistance, ball-carrying, on-ball attacking, open-play shooting off his right, shooting under pressure, inswinger crossing, ball distribution.

Vulnerabilities: positioning off the ball, 50/50 duels, headed shooting, left-footed crossing/delivery, weaker pressured crossing in the CL sample.

Coaching points: deny the cutback-and-shoot at all costs; show him down the line, never inside onto his right; don’t lunge, be patient in duels; double the touchline when he carries to force him wide; force him to defend, crash his post, and constantly have the full-back sprinting on his blindside.

👉 Bradley Barcola, winger

I’d currently expect Barcola to be the first name off the bench.

Analytic strengths: His numbers are a lot higher in the league. Numbers-wise, his strengths include: shooting and ball-carrying in the league sample, one-touch and pressured passing, over-the-top passing, defensive work-rate (blocking, tackling, dribble defending, 1v1 defending), ball distribution.

Vulnerabilities: tackle resistance, aerial duels, pressured shooting, headed shooting; his carrying and shooting have dropped sharply against CL-level opposition this year.

Coaching points: get tight and physical; he can be dispossessed. He is hell to face when you have a high-line, so you have to keep him as wide as possible, and engage as early as possible. He’s prone to looking caged when squeezed to the touchline and can have loose touches in those situations. He’s one of the reasons it’s worth staying low in a winning game-state.

👉 Désiré Doué, winger

Analytic strengths: he scores high in finishing under pressure off either foot, tackle resistance, defensive work-rate (dribble defending, blocking, 1v1), ball distribution. For a tekky wingfielder, he covers a lot of distance (second-most distance per 90 by wingers in the Champions League).

Vulnerabilities: aerial duels, over-the-top passing, left-footed crossing.

Coaching points: push him to the touchline, and out of central areas, however possible; deny the shot and dribble: like most good dribblers, he wants to provoke a reaction and a heavy step, so if you hang off a bit, he has fewer ideas; stay aware of his positioning at all times, as he’s an incredible finisher; you can surrender the pass if it means he doesn’t dribble you; attack the flank he vacates when PSG lose the ball.

II. The midfield

👉 Vitinha, central midfielder

For the midfielders, I thought I’d pull the wider challenge/duel stats as well:

…and finally, his deeper passing grades. 😳

I’ll throw my hand up and say I had underrated him before watching all the footage for this. I knew he was quite good, but sheepishly thought he hadn’t faced physical enough players and could be shoved off the ball. But it’s really almost impossible to get the ball off him. Agent Thomas Frank was categorical: “Vitinha is the best midfielder in the world. He will be the next Ballon d'Or winner. Wow, what a player.” Anyway. Here’s what the stats show:

Analytic strengths: press-resistant passing, ball distribution, long-range passing, through balls, ball-carrying, shooting from range. Passing, passing, passing.

Vulnerabilities: left foot, aerial duels, crossing, set-piece delivery. Not as dynamic a passer when unpressured. In possession, he constantly picks up positions that will be vulnerable on a ball loss.

Coaching points: Is it possible to mid-block one player and press the rest? I say let’s try. Lay off and deny the forward pass, then jump the receiver; beat him up in the air however possible, and bully him especially on second balls; close out his shots on top of the box; drive at him in transition, especially if he’s dropped down to the wide-CB role and they’ve lost the ball.

👉 João Neves, central midfielder

His challenges and duels:

Analytic strengths: He’s good at everything, is my scouting report. Specifically, his one-touch and pressured passing, ball control, tackle resistance, 50/50 duels, aerial threat for his size (headed shooting on set pieces). If anything, I think Enrique underuses him.

Vulnerabilities: He doesn’t grade quite as well on dribbling, ball-carrying, pressured shooting, over-the-top passing, and crossing.

Coaching points: He’s extremely busy, overhelpful, and occasionally over-amped (got a silly red in the Club World Cup final). He’s often cleaning up others’ messes in the middle, so I’d want to use dribblers and runners to drag him away from zone-14, get them to cover his spot, then attack the middle. I’d also a) want to bait his aggression with some shit-housing, and b) potentially target him aerially, which sounds wrong because he’s their best aerial midfielder. The reason is that it’d leave Vitinha+Ruiz (or Vitinha+WZE) as the midfield “second ball net,” which is much preferable to having Neves there.

👉 Fabián Ruiz, central midfielder

Challenges and duels:

Analytic strengths: left-footed passing, distribution under pressure, over-the-top and through balls, tackling, ball-carrying out of midfield. He’s their best big-space passer.

Vulnerabilities: tighter-agility dribble defending, tackle resistance, aerial duels, positioning, shooting (he’s graded out poorly this year, but we know he can hit a ball). He can feel more lethargic than others on this team.

Coaching Points: run at him as a defender; he’s usually press-resistant as a passer/shielder but not as a carrier/dribbler, so he’s less of a threat to drive into open space, and I still want to test him with shoves and tackles; once he already has the ball, I’d personally lay off him a bit.

👉 Warren Zaïre-Emery, central midfielder / right-back

He might start at right-back (if Hakimi is out), at midfield (in place of Ruiz), or arrive off the bench.

Analytic strengths: ball-carrying into space, pressured passing, defensive work-rate (tackling, blocking, dribble defending, 1v1), aerial/headed threat, plus an elite engine (8th-most distance covered per 90 in the Champions League).

Vulnerabilities: shooting, tackle resistance and isolated dribbling against top opposition, over-the-top passing, crossing, wide pacy defensive 1v1s.

Coaching points: press and harry him on the ball, he’s potentially dispossessable; block his carrying lane to force the pass; he’s fairly shot-shy; track his late forward runs, because he doesn’t get tired. If he’s right-back, play it in behind and test his long pace.

III. The backline

👉 Achraf Hakimi, right-back

Analytic strengths: shooting under pressure (especially high for a fullback, as we know too well), ball distribution, right-footed and pressured passing, ball-carrying forward, tackling, 50/50 duels.

Vulnerabilities: overall deep tackle resistance, dribble defending, crossing (especially under pressure in the CL); constantly leaves space behind.

Coaching points: attack the space in behind when he pushes up; he’s dispossessable in deeper areas, you can go hard; close out his shots hard, with sound blocking technique.

👉 Marquinhos, right CB

Analytic strengths: range of passing (including over-the-top), reading of the game, aerial and ground duels (good, not elite).

Vulnerabilities: tackle resistance, passing accuracy under pressure (big drop from unpressured), clearances, speed.

Coaching points: press him on the ball, his distribution falls off when hurried; attack him at pace, especially when Hakimi has stepped up (one of the team’s biggest vulnerabilities); deny him the unpressured first phase, push him to play long.

👉 Willian Pacho, left CB

He’s quite good. I’ve loved him since his early days at Eintracht Frankfurt.

Analytic strengths: aerial duels, dribble defending, 1v1 defending, positioning, tackling and blocking, left-footed pressured passing.

Vulnerabilities: dribbling/carrying out of pressure, ball distribution and progressive passing, over-the-top passing.

Coaching points: press him hard; cut his short passing options so he has to play long where he’s weaker; avoid lumping it onto him, as he’s an outlier for them in aerial duels, amidst plenty of weakness; like Neves, he cleans up a lot of messes, and follows the ball, so get him out of position and then attack the spot he vacates. A lot of their players can get too close to the ball and lose runners.

👉 Nuno Mendes, left-back

A joke of a player, whose only real flaw is his ability to stay fit. PSG have a solution for that: injuries and all, he’s played 40% of the minutes in Ligue 1 across 5 seasons, which magically jumps up to 75.4% in the Champions League. He’s started-and-finished (i.e. played all 90 mins) 44 total games in Ligue 1 in his career, or 8.8 times per season.

Analytic strengths: passing under pressure off both feet, over-the-top balls, dribbling and ball-carrying forward, dribble defending and positioning, inswinger crossing threat. He’s won possession more than any player in the Champions League.

Vulnerabilities: aerial duels (moderate weakness), finishing (though he can strike it, just inconsistently), raw tackling/50-50s in the CL; high attacking positioning leaves space. He can sometimes hit ill-advised switches that open PSG up to the counter.

Coaching points: his advances are when PSG are most vulnerable: they don’t usually make Hakimi drop back on the other side to compensate, so it means there is a front-6 of committed attackers, and Vitinha is covering the wide-LCB spot in build-up. Vitinha is really savvy with his positioning, but still: go, go, go. I also have stubborn faith that Saka can beat him; he turned his back on Olise once or twice.

👉 Matvei Safonov, keeper

After last season’s elimination, Arteta’s thoughts drifted to the goal.

“When you look at the two games, who has been their best player on the pitch has been the goalkeeper. He's made a difference for them in the tie.”

Donnarumma can’t hurt us anymore.

Here’s Safonov in the league, since taking the job from Lucas Chevalier:

And here it is in the Champions League, a smaller sample:

In the Champions League, he’s faced three set-piece shots and surrendered two of them.

He is not highly involved in build-up:

…and that is especially the case in the Champions League, where Enrique has him regularly booting it out on purpose, forcing opponent throw-ins in suboptimal locations:

Through Gradient’s numbers, he fares quite well in shot positioning across the board, but is more mid-grade during the actual save. His pressured passing is also meaningfully lower than his unpressured passing.

Coaching points: Press him hard, push him to his left foot. They have outlets, but not aerial outlets.

👉 Substitutions

If Ruiz or Zaïre-Emery don’t make the eleven, they are highly likely to come on. Aside from that, the four players with 5+ substitute appearances in the Champions League are:

• Lee Kang-In: 10 appearances as a substitute, 27 minutes per substitution.

• Lucas Hernández: 10 appearances as a substitute, 24 minutes per substitution.

• Gonçalo Ramos: 10 appearances as a substitute, 18 minutes per substitute. This feels fairly likely, as Dembélé may not be fully fit.

• Senny Mayulu: 6 appearances as a substitute, 20 minutes per substitution. So impressed with his footage, by the way.

Other possible appearances are Illia Zabarnyi, Ibrahim Mbaye, and Lucas Beraldo.

Next, onto the play-style.

👉 Referee

The final will be refereed by Daniel Siebert, a 42-year-old German who has officiated two of our games this year: the first leg v Sporting, and the second leg v Atleti. He presided over one game for PSG, which I also watched (against Athletic Club). He’s been around the median with fouls (22 a game), slightly above-average with yellows (4.44 a game), and has given one penalty and two reds across nine games. Over a longer time frame, his stats (fouls per game, yellows per game, reds per game, penalties per game) all look pretty normal.

🔵 In-possession

👉 First phase

As we established, Safonov isn’t terribly involved in early build-up. Against Bayern, he purposely hit it out to generate conditions that would squeeze the Bayern attack.

In the Champions League, he has largely gone long, hitting 46 long goal-kicks, compared to 9 short ones. He hasn’t hit a single short goal-kick in the last five Champions League games. In the second leg against Bayern, he took no risks and hit it long 14 out of 18 times total. Almost all of them were hoofs that had very little chance of a PSG ball-win.

Other players pick up the slack, particularly Vitinha. Below, you’ll see a more standard look.

All three midfielders are all-purpose and can be expected to float anywhere in the central areas. While Vitinha can look to back up and disappear up to the frontline, he is primarily the deepest one, constantly getting on the ball, picking it right up from the centre-backs as he sees fit. Would you like to know how many successful passes he has in a typical PSG league game, per Opta? It’s 109.

During goal-kicks, the shape can be tremendously flat to try and generate 1v1s on the backline. I’d largely expect Arsenal to go man-to-man for this one.

And here’s what I’m talking about with Vitinha. He picks it up off the keeper, and does the job of Raya and a defensive midfielder in one fell swoop. At the end, he disguises his body to open a lane, and hits it up into the pocket.

When there are fewer numbers forward in the opponent press, that’s when things get interesting. You’ll see Mendes stay heavily involved as an auxiliary CB, while Hakimi pushes up as the de facto right-winger. This then gets the right-winger inside, swapping and changing with others.

The PSG midfielders are required to be comfortable all over the pitch. As I warned, here’s Vitinha dissolving into a striker-type position, Zaïre-Emery dropping as the #6, and Neves going out to the left-wing. Those wide duos (usually high midfielder + winger, or even winger + winger, or winger + striker) are the heart of the PSG attack.

And here’s why their mobility is so hard to defend. They test you horizontally and vertically, asking: OK, big dog, are you really ready to go man-to-man?

Below, Davies is following Doué, but he’s a left-back. Is he willing to follow him completely across the pitch? No, that’s too risky. So he hands him off to the middle and keeps his spot.

But that creates a moment when Pavlović can either leave Doué free or leave Neves free. Pavlović pushes up, Doué is free, and the press is broken.

As high-pressing-slash-bold as Arsenal will look to be, you can’t get caught in the middle: you either have to commit to going truly man-to-man in a situation like that, or, if not, give priority to the player behind the backs of the midfield. The backline and the double-pivot can’t allow Dembélé and Doué to roam in these spots. Either follow them, or tighten up your pressing distances so you’re closer.

So that’s the challenge. The nice thing is that PSG are not impervious to deep build-up mistakes. This is because they have a few players who are simply pressable. On a raw counting basis, they’ve given up the most chances and goals from errors. To my eyes, it’s often when they build up on the left, and then look to switch the angle of attack to the right.

One of the best ways to generate a ball-win is to jump these sprinting underlaps. It is just a constant move by them, as we’ll cover later, and it’s always worth anticipating.

The right (from Safonov to Marquinhos to Hakimi) can also be pressured deep.

👉 Consolidation phase

As things progress into the middle third, here is the most typical look of their build-up shape against a 4-4-2 zonal block. The front-four are completely interchangeable, and they’ll also swap with the midfielders. Because the opponent’s central CB is often left lonely, the free man is usually the far-side winger, by design.

Hakimi is almost always high in those situations.

But Mendes can join the frontline as well. In those cases (more often than you think, even against good competition), Vitinha backfills the role of wide-CB and the shape continues. Vitinha drives a lot of tempo from back here.

You’ll also see it here, in which PSG seek “double-width” outside.

That brings huge, vibrant strengths, but also vulnerabilities on a ball-loss.

👉 The wide double-moves and underlaps

For all the rotations and fluidity, I was surprised by how much the PSG attack seems to be built on a relatively small number of automatisms. So many attacks carry similar characteristics: Vitinha picks up the ball deep, and probes left and right. Two incredible attackers are in close proximity and make opposite movements: one goes behind, while one drops deep. You see it here:

But because the timing wasn’t right, they just switch. Mendes delivers it to a dropping Doué, which means it’s now Kvara’s turn to run in behind.

…and he does, starting the movement for his goal.

In more advanced areas, we see it even more.

It feels like the entire heart of the PSG attack is based on this: one wide attacker drops, one wide attacker sprints in-behind (often after a give-and-go), and the ball is delivered into space.

Ahem. One wide attacker drops, one sprints in behind, ball is delivered into space.

This feels predictable, and that’s because it is. For one, this is just about incredible quality and physical characteristics. For the other, this is about defenders who aren’t anticipating this enough, and are doing a poor job of staying goalside of their marks.

👉 Final third and chance creation

In the Champions League, PSG have had a lot of the ball, but aren’t always in a rush to push it forward. All their possession doesn’t culminate in an excess number of touches in the box: they have the fifth-most, after Liverpool, Bayern, City, and Inter.

They are often keeping it around here, then recycling the ball, and then rotating endlessly.

To my eyes, they are better at generating threat when it originates in the middle-third, because they aren’t the most dangerous (or voluminous) crossing team, they don’t have a classic aerial threat in the six-yard box, and they have such dynamic runners and dribblers. They have benefited from a lot of worldies on the edge of the box. I think the backline can step up a little bit to compress space in these situations. An aerial goal could happen, but it’s less likely than a great strike after Kvara, Doué, or Dembelé made somebody lunge.

👉 Shot profile

On that note. According to Understat, in Ligue 1:

• PSG underperform in the penalty area (50.14 xG, 46 goals)

• PSG overperform outside the penalty area (8.48 xG, 14 goals)

• …and especially underperform close to goal (-4.68 on xG in the six-yard box)

Defending this final third play will come down to patience, sound blocking technique, and closing out each attacker in a way that suits their shooting style.

🔵 Out-of-possession vulnerabilities

For this section, I haven’t set out to give an exhaustive or fair synopsis of their shape and philosophy. Instead, I went probing for vulnerabilities. They weren’t impossible to find. In the Champions League to date, Arsenal have given up 6 goals, while PSG have given up 22.

👉 Press

PSG are about as man-to-man as one could possibly fathom, and it works: they score the most goals off high turnovers in the Champions League.

This is a typical look for them, with each player doggedly following their mark wherever they go.

You will also see susceptibility in that shape. With Cucurella further down, and somebody occupying both Vitinha and Pacho, Marquinhos has to swing out to mark Pedro Neto.

What happens next is fairly expected.

Some version of this image (the 32-year-old Marquinhos pushing up, forced to track a snappy winger on the halfway line) happens in almost every game he plays.

👉 Defending in the middle and defensive thirds

Even in their own third, PSG usually turn to a high-octane style of defending, which basically resembles pressing. This is another area to exploit. They did a solid job against Bayern in the second leg, but they simply don’t look as settled or comfortable in these moments. They don’t have a lot of joint experience in that kind of defending.

João Neves is perhaps their best defensive player. And he runs around fixing a lot of things. This can also be used against him: if you’re able to gain his attention, you can create those high-effort moments where he gets dragged around, out of the middle.

Now, with play reset, the rest of the PSG shape is starting to funnel out toward their marks, and Neves is left deep. Because they all track the ball (and their man) so much, spaces can open up centrally, like this.

Especially when Neves is out of there, those spots are soft, indeed. The likes of Vitinha and Zaïre-Emery don’t make for the most imposing zone-14 defenders. Here, Olise notices and drifts behind.

With that space open, and Neves too far to disrupt the play, Olise scores.

They are also susceptible to the same attacking moves they’ve perfected. One of the best ways to get past them is through timed underlaps, dragging a midfielder (who is usually playing high anyway) all the way through the channel. If nothing else, this again clears out the middle.

Bah gawd, that’s Calafiori’s music.

👉 Central carries and duels

PSG do a good job with energy and bumps, so they aren’t always open for business in the middle. But the secret to beating any man-to-man press is to have your … man … beat their … man. This starts a domino effect where others have to cover, and somebody is inevitably found open. Rice found that last season.

…and Lewis-Skelly did, too.

👉 Rest shape, defensive transition and counterpress

But their biggest vulnerability, by far, by far, by far, is how they attack. The positioning of the midfielders is usually quite high, and Mendes and Hakimi are often out wide, unconcerned with squeezing the middle in case of a ball loss. They don’t lose the ball often, but when they do, they look wide open.

One of the most hopeful moments is when Mendes is overambitious and dispossessed. Because the other full-back, Hakimi, is always flying on the touchline, the transitional coverage is Vitinha (often wide) and the two CBs, with midfielders far ahead of the play. On a tackle, a good first glance for a pass is always the spot that the opponent just vacated.

If Mendes hasn’t pushed too high, that space is almost always at their right-back. The imperative is to counter that way and send multiple runners if you can.

After a ball-win here, we see how easy that space is to access.

Bergvall made it all the way to the box.

That’s how Saka’s goal last season started: ball-win, immediate ping to the vacated right-back position, and Trossard bullied Marquinhos for it.

👉 Restarts

That fluidity can be used against them when possession changes hands. Here, Neves lost it, and Chelsea has acres to play with on a quick throw-in. Every Arsenal player should be ready to throw.

That turned into a 3v1 in the box.

In the Club World Cup, that same trigger (a failed Nuno Mendes carry) led to this Bayern counter, reinforcing two things: the importance of a dropping striker against a man-to-man scheme, and just how open PSG can be in these situations.

👉 Wide defending

Nuno Mendes is tremendous (tremendes?). But looking back at last year’s footage against Saka, I don’t think this is an unwinnable battle.

He may be rusty, and I’ve seen a few of these this year.

👉 Trailers

Aside from Neves, PSG can lack physical ground-eaters in the middle. They can also congregate around the ball a lot. That means that these rebounds by late-arrivers can seem more “on” than usual. Look how many players are out wide near the crosser, versus blocking the “second-six-yard-box,” where most goals are scored.

👉 Aerials and duels

Finally, PSG are extremely light on aerial duellers. They can be beaten in a situation like this.

Or this.

While Pacho (good), Neves (good), and Marquinhos/Mendes (fine, not great) can hold their own, the rest of the team is filled with liabilities.

Here are some positional ranks on aerial duels in the Champions League:

• Vitinha: 153/161 (CM)

• Fabián Ruiz: 138/161 (CM)

• Dembélé: 75/110 (CF)

• Barcola: 129/149 (W)

Ironically, we may want to target one of their stronger aerial performers (Neves) for reasons we’ve covered earlier, and will cover again later.

🔵 Set pieces

For our next section, I’m again so grateful to be able to turn to the words and brain of Jake, commonly known as the Set Piece Sicko, revered for his work on the intrepid Pot Shot Podcast as well as his insights at @setpiecesicko. I’ve admired and devoured his work on set pieces for years, and I’m so happy he was willing to share his thoughts ahead of this one. Go follow him and listen to that brilliant podcast, please. This entire section is by him. Thank you.

For your common football social account, set-pieces seem to nicely encapsulate the harsh vibes contrast of these two sides.

PSG the romantics: turning their nose up at the brutish meat-walls and dead-balls of the criminally English Arsenal, who wouldn’t dream of trying to pass the ball in beautiful triangles, preferring the quarter-circle wrapped around the corner flag. Boo Arsenal! Boo corners! Long live teams who entirely exist within the boundaries of play!

After watching a tonne, I do agree that you can fundamentally distil these two sides into their differences at set-pieces, but it is primarily the way they integrate them within their ethos, rather than one just not using them. You know enough about Arsenal, so the emphasis will be on PSG, with sprinklings of “how we can get up and at ‘em”.

I. PSG’s offensive set pieces

The best encapsulation of this idea of PSG being embodied within their set-piece strategy is on offense. This will come in three key prongs: corners, free-kicks, and throw-ins.

👉 Corners

This is thankfully easiest to do due to the sheer volume PSG create (due to the sheer volume of their bank statements compared to most teams they play, but anyway).

PSG use their corners in a weird way for PL eyes, in that their emphasis is not to launch it on the keeper’s head, but instead to manufacture better situations for their star qualities to shine outside of the box.

To do this, often PSG will go short, with a midfielder joining a winger who takes. From there, we can get a few branching pathways, which I’ve tried to make feel like a choose-your-own-adventure book below:

Option A: You spy the opposition short defender is not a gifted 1v1 specialist. You choose to isolate him and run towards the goal, along the line.

Option B: You notice the opposition are being dragged towards you, leaving a large amount of room for either your burly CF or even burlier CB to dart into the space far post. You hit it first time, catching a sweet inswinger behind an advancing defence.

Option C: Your physically diminutive but technically mint midfielder is screaming for it on the edge of the area. You roll the ball towards him to hit within two touches, relying on blockers in the box and fear of Options A and B to keep the shot unblocked.

Options A and C are primarily where PSG hang their hat, making use of the snapshot ability of Vitinha alongside the individualistic brilliance of Dembélé, Kvaratskhelia, and Doué to transform a situation which fundamentally shouldn’t suit PSG into one that resembles the freedom of open play. I’ve rarely seen a team that prefers a punt from 25 yards to a chance of a close-range bundle, but it feels like a smart stylistic choice from PSG to not engage themselves in duels where lower teams may actually have the upper hand. By enabling the battle to stay wholly technical, they don’t enter this weird control-less aether that they may be used to.

For those interested in corner lore (lore-ner?), this feels very 2017-2022 Guardiola, with corners being used to complement in-play ideology, to the extent of a team only seeing it as a method for getting back to the comfort of high technical possession.

The solution teams adopt is to simply go tight and press out quickly, removing PSG’s ability to ‘reset’ from these moments and gain dangerous territory comfortably. The natural caveat of this then brings about Option B – now perhaps you’re starting to see the dangers PSG pose here. Due to the quickness with which PSG players can change direction, teams often allow huge spaces due to the inconsistency of structures as they move outwardly – like 2000s Hulk Hogan ripping his t-shirt, slowly there are gaps, which become huge tears, revealing a weirdly soft underbelly. All it takes is one well-timed run for one of the 2/3 targets on the side, and it’s a far easier chance than any initial delivery can hope for.

What is a positive for Arsenal is that you rarely see this tactic employed vs an equal-strength side (Most of these looped crosses are aimed to Gonçalo Ramos, which says enough to this point), and Arsenal’s added height and reputation mean I would be expecting an emphasis on Options A and C, especially given PSG’s success from range in last season’s fixtures.

This isn’t the only way PSG roll their dice, even if it is the one I’d be banking on. Much like last season, their aerial targets are few and far between, but that doesn’t stop João Neves being a 5ft 9 Gabriel. If PSG ever do fancy launching it, Neves will 90% of the time be the target; he has an incredible leap for a man with his shirt tucked in, and a timing of run that would make another tucker (Frank Lampard) blush.

For some reason, the deliveries will never exceed the front post, but Neves will consistently hit the spot just out of the zonal coverage – even easier to do on PSG’s favoured outswingers. This is exactly what we see vs Bayern in the 5-4 win, and he nearly did the exact same to Arsenal in the group phase of last season’s competition.

I don’t often recommend adjusting systems for one player, but given PSG’s preference for edge-of-area plays and Neves’ desperation for CTE, I would be moving the standard outswinger zonal setup a couple of yards forward to make things difficult.

👉 Free kicks

Only a small note here after the novella on corners, and that is two core features of play I’ve noticed.

The first comes from PSG’s domination of the second phase, which is perhaps expected given their focus on intensity and pinging it from 30 yards. PSG’s set-piece strength tends to be in these ‘overspill’ moments, rather than singular routines of absolute quality, but the heart of how these moments occur is actually a clever inversion: rather than be intense, PSG players earn space by switching off.

Looking at this scuffle vs Chelsea before a Bradley Barcola Banger, what screams out to me is how PSG earn positions by ghosting behind automatic defensive movements. It’s how the first run is earned, it’s how the shooting zones are earned at the top of the box, and how movements for second balls after that are earned. Weirdly, this will require more active thinking, as the typical runs are likely to be ignored in favour of waiting until Arsenal have made the space via the expectation of those runs.

Very similar to Fabian Ruiz’s goal last term, drifting away from a backline in autopilot:

The second note here is a more ideological one. Versus Chelsea, in the second leg specifically, PSG suddenly went from the tippy-tappiest side around to absolutely launching it. Safonov got an assist. A lot of this was in reaction to the aggression Chelsea suddenly showed in their high line: rather than deal with congested space where the opposition were specifically set up, PSG went right over, isolating an uncomfortable CB playing RB against Kvaratskhelia. Glad that isn’t going to be a situation Arsenal have to deal with, eh?

This general openness to direct play, knowing when physical battles aren’t worth the hassle, is a hallmark of PSG on set-pieces this year, and it feeds perfectly into their ethos of hitting sides early then opening the space up for fast, brutal counters. If Arsenal get behind, it’ll be a key battle to ensure aggression doesn’t dominate thinking – PSG love a quick set-piece, and if it can earn them an extra 30 yards doing so, they’ll eat it up.

Aside, absolute twats from the edge are still an option here too.

👉 Throw-ins

The last point on offense maintains this line of thinking. Whilst most throw-in patterns are centred around passive-aggressive possession quality at the top end of the field (see: Doué’s goal vs Liverpool, after 27 passes), some are more overtly aggressive.

A lot of this is covered by the ever-brilliant Liam Tharme, but it’s worth repeating with a little extra sauce.

Much like in their game vs Chelsea, when faced with Liverpool’s out-of-possession-aggression, PSG utilised set-pieces to give themselves situations to isolate a stretched backline. From his own third, Nuno Mendes yeeted flat but perfectly intended deliveries to isolated players in the middle (often Vitinha) who could then drive forward in ideal 4-on-4 scenarios. Often, players aim down-line or short in these scenarios, which is why Liverpool felt so comfortable ‘boxing em in’, but PSG used their half-baked enclosure to move forward very quickly. The speed of these moments is startling, and again leads into the idea that PSG’s main quality in set-piece play is quickness of surprise, not simply the surprise itself (much like in open-play again, wow!)

This strategy of launching far and flat is even more effective towards the half-way line, where Kvara can escape backlines to generate these quasi-counters. Barcola is also a big beneficiary of these.

My two cents on throw-ins is: pick your battles. It may be worth sacrificing a final third pressure to ensure you don’t need a 50-yard recovery run. Limiting direct options quickly is a huge part of this.

II. PSG’s defensive set pieces

Now we move to the juicier bit, where you can get at PSG. The sections won’t match the previous, due to the limitation of ‘not as much tape defending when you’re good’, but a lot of the focus will be on how they’ve set up in the UCL, as this is I think where we can get most purchase in terms of squad use.

👉 Corners

A fair while ago, I wrote about the notion of goalkeepers, and how a team’s defensive structure on corners tends to betray their feelings about the goalkeeper within it.

With that in mind, here is PSG’s defensive structure on inswingers (Arsenal’s bread and butter):

This is not to say that Safonov is a De-Gean level goalline goalkeeper, but it’s more a respect of opposition strengths, and particularly opposition targeting of the Russian. Much like any PL goalkeeper, focus is often placed upon dislodging Safonov from his space and trying to secure the front post zone in particular. Part of this is due to the PSG setup favouring this space too, meaning attacking teams get some blockers for free.

Safonov is not a particularly commanding goalkeeper, hence the… everyone around him, but he is a weirdly aggressive one. He will consistently challenge crosses in the area but never claim them, opting for a punch in all situations – likely making him slightly harder to totally limit, but it means there is some optimism for weak skims provided the momentum he can build up is limited (I miss you every day, Ben White).

Chelsea and Liverpool have both had real joy in hitting the front post and probably should have had a goal for their troubles, and the fear with which PSG set up is a general limitation in how they can prevent runners. The only positive of the structure is the presence of João Neves in the Kai Havertz role, leaping high to stop this zone being rife with flick-ons a la Gianni Vio’s Spurs a few years ago. Of course, the way around this is to get even further forward, as Aston Villa focused on a year ago (this was in the first minute vs an equally uncomfortable GK in Donnarumma, admittedly with Vitinha in a front post role instead).

The aggressive-but-not-commanding problem also leads to ‘second wave’ deliveries being a nice option: mirroring the PSG ‘go short and whip it later’ method to create a cavernous space between an advancing back-line and an unsure goalkeeper (see vs Monaco below).

Another sign of this fear is a retreating Kvara at the back post, tucking in behind Safonov. This is perhaps the most important point for what I think is the way to get at PSG, because look at the image of the box as he does so:

This also works for free-kicks.

There are chasms of space at the back post, and in the zone around it for scraps. Not many teams have utilised this space, or perhaps have tried to access it through front-post flicks, but with Arsenal’s deep delivery quality from Rice and Saka, it could prove a wonderful pool for Gabriel to dive headfirst into, or Trossard to hang out at.

Baiting the keeper into overextending and throwing his arm at it also extends to long throw-ins, which Joe Gomez recently found by aiming at the front post.

I think aiming to bait the front zone and exposing the under-staffed back would be a real play here, and one Arsenal are very accustomed to (this is basically the same idea as how Kai Havertz got his free header vs Burnley).

What Arsenal should be very fearful of is the short corner. PSG, again, much like in open play, have a very snappy front-line pressure in these moments, and will force a team back very quickly given the chance.

Take, for instance, this corner Chelsea earned in the 43rd minute of the return leg:

This, however, leads us into a nice amalgamated section of set-piece analysis, which I can subtitle for you.

👉 Counter strategies

This outward pressure is consistent with PSG in basically any forward zone, with the aim to move the opposition away from their box as soon as possible and ideally get a nice counter from it, like above.

However, sometimes this is overzealous, and this presents a real strategy to beat PSG at their own game.

As the pressure builds and front-line runners move out to hit CBs hard and fast, what happens is that gaps naturally emerge – Dembélé is especially at fault for becoming a press leader with no understanding of how fast his teammates actually are. Much like with Arsenal’s second-phase issues last term, the reshuffling of parts following the re-emergence of open play presents opportunities (also known as Massive Gaps) which can be exploited. This clip specifically reminds me of something Calafiori is capable of – I know it’s Chelsea again, but there’s a reason they paid millions for Bernardo Cueva:

Recycling moments can produce quasi-transitions through some powerful running, hitting PSG at their own game, which is something they aren’t particularly suited to, frankly. Even if you want to limit how often you get into a running race, you can utilise this to earn positive wide overloads. The issue will be how much this strategy would involve Mosquera, but it enables you to vary strategy without playing into PSG’s hands.

All of these tactics are best served focusing on the far post of PSG, or at least the part of the set-up furthest away. This is typically the weakest/emptiest part of the structure (see: at corners), and it’s no surprise this is where Tottenham found joy all the way back in August. With an unsure keeper and a structure that is rarely fully tested domestically, there are weak points to poke and prod at if you can avoid their stronger areas.

The final point I’ll make here is that this will need courage. Both offensively and defensively, to avoid PSG allowing set-pieces to become another part of their game domination, you need to ensure that tactical space remains yours: not falling for baits outward, resisting the urge to defend automatically and remain active, reduce space for top-of-the-box shots, pushing up and out to limit João Neves’ running starts, pulling PSG out of shape by trusting your build-up, and giving yourself as much space as possible at the far post for a charging header. A cohesive game-plan to ensure PSG remain uncomfortable can fundamentally change the fear and pressure in these moments: some oversized cojones could go very far.

OK, thank you, Jake. Required reading, all that. Now, back to my voice.

🔵 Game-state

When I pulled game-state data across competitions, I found a few things:

• PSG look the least potent, and have lowest xG difference, when losing. In those situations, they have the most possession, shoot more from outside the penalty area, log fewer counterattacks, have the highest press, log the fewest passes per possession, and, perhaps most notably, log a lot more crosses (twice as many as when they are drawing). It’s easy to understand causes and effects there.

• They are incredible at tacking on to leads, logging the most goals and shots and goal difference in that situation. They are happy to defend a little and then play explosive counterball, which makes sense.

• In attack, they log the lowest xG when the score is even.

Now, using Understat data, we can look at how they fare per timing of the game in Ligue 1:

It’s clear they are most potent late in games, which is usually the case (there are more tired legs and the most minutes available), but it is especially the case if you have the kind of attacking quality that PSG offer. Most interestingly, their performance is lowest in the period after things settle in, between 16-30’. After watching, I’d suggest that it has something to do with PSG starting to go for it, having a sloppy moment, and being caught with a bad rest shape: in that period, the goal differential is only +3. Arsenal, often more conservative in first halves, may want to go for it early.

Also in the league, here’s what it looks like based on game-state.

They have the lowest xG/90 in even state, and the second-lowest when they are down by 1. It’s a small sample, but when down by a goal in the league, they only scored once and conceded 6 times. In other words, leads and losses can both compound for PSG.

🔵 Selections

Here’s where I’d lean.

My thoughts:

• Havertz wins out for his ability to go over-the-top and win aerially, as well as his ability to drag a CB and play total football. Gyökeres can come on as a finisher, where he’s been looking great. I’d like him to do shoulder runs on Marquinhos in the last twenty minutes.

• Timber is the obvious start if he’s free to do so. If not, I would go with Mosquera. While I suggested Zubimendi at right-back the last time we spoke, and then we saw it at Crystal Palace, I think that looks most dangerous when we’re facing a pinned opponent and need another dynamic player in the pockets. For this, I’d prefer Mosquera to face play and defend, enabling the left-back to roam at will. There is still an argument for Zubimendi, as the full-back gets such valuable free touches in this system, and he can come on as a sub, but I’d ultimately lean the other way.

• On the other side, Calafiori is a clear choice for me. His underlaps and ability to provide an option above the pivot will be huge for exploiting the spaces that PSG can leave open. Hincapié can come on at 65-70’ and do well in whatever game-state that requires (defending, or chasing a goal + running + crossing).

• Trossard is the likely choice for me (and Arteta). That move where he pulls back on the touchline and hits it into the half-space should be on, and he’s been playing great of late. Martinelli can arrive as a closer at either right- or left-wing, though I’m a little tempted to have him sprint behind Hakimi. I suppose that can wait.

• I’d go with Ødegaard in the RCM spot. Not only does he offer some control, pressing sharpness, and combination quality with Saka, but Eze offers more optionality off the bench. He can come on as RCM, LCM, or LW.

• The most difficult decision is Lewis-Skelly or Zubimendi. It’s very 55/45 to me. It ultimately comes down to a deep, personal understanding of the players involved. If Zubimendi were at his early-season finest, it’s an easy call, as he’s superior defensively and so good at dragging around man-to-man schemes like that of PSG. But right now, there is a swaggering sense of Hale End belief that Lewis-Skelly is instilling in the team, and he’s shown how many complementary characteristics he offers to Rice and Calafiori in particular.

• Regardless, I’d play Rice in the deeper role of the double-pivot. If Zubimendi starts in midfield, I’d want him to reprise the role he played in the first leg against Atleti, higher and more to the left, playing in the pockets.

The subs, then, are fun to think about: Eze, Martinelli, Merino, Gyökeres, Hincapié, Zubimendi, Madueke, Jesus, Dowman, Nørgaard.

🔵 Final thoughts, gameplan, and levers

This section, in particular, was done in coordination with Bastian (@austriangooner1).

PSG are the best attacking team in Europe. This is because of some savvy tactics by Luis Enrique, sure, but it really comes down to their outrageous attacking quality and depth, their shared identity, and their ability to rest and rotate. These factors have hidden a structural cost. Their fluidity and man-orientation leave exploitable windows, but they require a team with defensive qualities to force turnovers, and then the speed of action to punish before they can regroup.

In the Champions League, their attacks are a little slower and more foreseeable than may be commonly thought, but that doesn’t make them any less devastating. Those fuckers can strike at any time. This, then, shakes hands with their outlandish defensive intensity, which is aggressive and high-octane in every third.

The idea is to show them the cost of their rotations. We can take advantage in the moment we win possession, target the space behind the full-backs, use quick restarts to unsettle them, and transition faster than they can regroup (or interrupt). These windows are short but potentially exploitable.

To sum things up, here are a few levers to pull.

👉 Defensive levers (how we nullify):

1. Commit a brave, lopsided, high man-to-man press. Ferry them to their left. Our CBs can follow their central hubs to deny the connection between deeper build-up and the attack. Arsenal can take a slightly different tack with Vitinha, with more conservative positioning, cutting off his right-side access, pushing Ruiz into some back-to-play receptions, and forcing clearances onto Safonov’s left foot. This all requires full commitment. They’ll happily remind you of the cost of getting a trigger wrong even once.

2. Force ‘em long. Safonov would rather hit it out than play through. Raya’s sweeping, plus Arsenal’s aerial advantage, means a long-ball is close to a guaranteed turnover.

3. Keep the play in front. This is mostly about distances between our backline and midfield. The false-nine coverage and handoffs will require constant communication. We can’t allow the likes of Dembelé or Doué the room to carry it out. We’ll have to drag them down if we’re caught out (respectfully). And the likes of Lewis-Skelly have to be careful of early yellows.

4. Beware double moves and switches. In settled play, PSG’s best attacks have a familiar chorus. Their top chance-creation move is the wide double-action (where one drops and one gallops into the half-space). The other big move is the hit out to the opposite-side winger.

👉 Attacking levers (how we score):

1. Counter behind the fullbacks. Hakimi is seemingly always high; Mendes is often high. When we win the ball off Mendes’s side, the transition coverage is Vitinha (who is often wide) and the two CBs. The other midfielders are usually ahead of the play. The first look on every turnover should be to the spot the full-back just vacated.

2. Manipulate the man-marking. When Pacho follows Havertz into midfield, Marquinhos will then be isolated. Likewise, if the left-back (Calafiori) can drag Hakimi away, Marquinhos is on an island versus the Arsenal left-wing.

3. Calafiori underlaps. That thing where Trossard hits it from the wing into the half-space definitely feels on.

4. Go long, strategically. Havertz time. One idea is to avoid Pacho and have someone target Neves in the air on purpose, even though he’s their best aerial midfielder. The point is to leave Vitinha + Ruiz (or WZE) as the second-ball net, which is much preferable to battling it out with Neves.

5. Sustained pressure. Not for lack of effort, their entire frontline can get caught far up, as we saw against Bayern. Forcing them to defend their own fullbacks pulls them out of attacking shape and can tire them out. Let’s back-post overload at almost all times to compound it. In general (Bayern notwithstanding), they aren’t used to conceding territory for longer spells, and can look chaotic when doing so.

6. Saka vs Mendes is winnable, one hopes.

7. Late arrivers on rebounds. PSG congregate around the ball and leave the second six-yard box thin. Only Neves and Pacho can really fight there.

👉 Game-state / energy levers:

1. Du-el. Somebody asked for my ideas for beating PSG. I said, “My strategy is mostly around nerd-bashing. You find the little nerds and then you bash them.” I have not changed my stance. This involves tackling on the right triggers, Rice and Lewis-Skelly driving through on carries, and scrapping for every second ball. Throw in Havertz dropping around their midfield, and a Merino sub late.

2. Restarts as weapons. Quick free-kicks, throw-ins, and Raya pick-ups can hit the disjointed or rotated press before it sets. This is Fluid Team Kryptonite. There isn’t a better way of changing game-state and gaining momentum than throw-ins into the box. Take something out of Brentford’s book. Aerial superiority should allow us to create high-pressure moments and take advantage of the fact that you can’t clear throw-ins off first contact like you can with corners.

3. Clearances must be passes. We can’t volunteer territory back.

4. Manage the score-state. When we went through the numbers, we saw PSG’s per-90 xG climbs through the match, which is fairly normal, but especially prominent. They’re at their most efficient when up one, and most fragile when down by one. The ideal game would be to get to a one-goal lead, then make them break down a settled block, then get our finishers on.

5. Acknowledge momentum windows. Something we had been doing very well in the first half of the season, and at our peak, is knowing those moments where the momentum is there for the taking. To fall back into shape, conserve energy and retreat when the press starts to lose control. To get on the front-foot and push the issue when the opposition starts to make mistakes on the ball. To control tempo in possession and sustain possession in the final third when it’s needed instead of being overly aggressive and turning the ball over. In recent months, and past seasons, the pressure of being nearly-men has seemingly weighed heavily. We should play free of those old handcuffs.

I won’t try to TL;DR Jake’s section, which had so many good set-piece takeaways. But that battle looms large.

The broadstrokes are all true. PSG boast an incredible team. They can score at any time, even against compact shapes. Their talent is overwhelming. But they also have some clear structural vulnerabilities. And though nothing is guaranteed, I believe Arsenal fans can feel pretty good going into this. I am grateful to you, and I am grateful to this team, and after going through this exercise, I do feel more positive about our chances.

With one dragon vanquished, the team can show up with a mix of joy, drive, pragmatism, and physicality.

One more tunnel to walk through.

The tunnel was uncovered, a change Mikel Arteta, in his mania for detail, was reported to have requested ahead of the 2025/26 season. It was the same uncovered tunnel that delivered Eberechi Eze into a frenzy upon his unveiling ahead of Leeds in August, the player turning with a giddy disbelief, still coping with the idea that this was, in fact, real life. On this occasion, the first Arsenal player through was Bukayo Saka, holding a mascot’s hand and wearing the captain’s armband. At 24, he’s lived a full career’s worth of moments compressed into seven seasons, many at this ground, and through all the successes, setbacks, and battlewear, the smiling eight-year-old who joined Hale End was somewhere underneath.

The Champions League anthem blared, and on rewatch, you appreciate the choreography of the camerawork, panning through the lineups from right to left. The shot of the Arsenal lineup started, again, with Saka. Then it surveyed the remaining players, of various heights, all with their own stories. The second-from-last was Myles Lewis-Skelly, the Hale Ender who has picked up Saka’s baton and was boldly chosen against expectation. Finally, it settled on Eze, who, released by Arsenal as a 13-year-old, had taken 14 years to come home, in no small part to play in nights like these. Saka first, Lewis-Skelly and Eze last. Hale Bookends.

As the horns and choir continued to swell, the camera zoomed out from Eze, revealing the sprawling tifo behind him. A fleet of boats, the words ‘Over Land and Sea,’ the vanquished foes in the background. A decidedly good tifo, unlike last year’s lonely cannon. Things change, and they can improve, but they don’t improve on their own. Tifos or tunnels, tactics or performances, they improve because humans made it so.

There was pandemonium in the stands. All that was left was to play.

👉 The goal

The goal that sent Arsenal to the Champions League final came from a slow accumulation of factors.

Interestingly, the initial period included a lot of work by Saka to float across zones and connect play.

For an Arsenal-style attack, the spare man will swirl around the double-pivot, but with eyes on Eze, Saka would often drop in to overload the midfield and provoke pressure to squeeze in on him, hopefully opening up lanes for others. If anything, some of the responsibilities might remind you of a Jesus false-9 day.

…and another one.

While this costs you a wide outlet threat on the right, it dovetails nicely with Gyökeres staying central and pinning, giving you the ability to execute a box midfield without the striker dropping by default. This also shows you why progression looks better with Saka. Even as a right-winger, he’s a half a midfielder, helping unstick things without losing the ball.

He also gives you the presence of mind that comes from playing with a teammate who can figure shit out on his own. The first half also featured plenty of the hallmarks of a prototypical Saka Shift, in which he accepts the ball on the touchline and attracts 3-4 defenders by himself.

For the goal, he was more off-ball, and was rewarded for his travails.

In so many words, that decisive goal represented how I’d like Arsenal to attack.

Atleti finished the first half without a shot on target. And as the game progressed, Arsenal started turning up the temperature, pinning them back with more dominance, and throwing waves at the final third. The final fifteen minutes were all Arsenal.

Pressure accumulates and eventually bursts pipes.

You can see the extent of the pinning here. All 20 outfield players are in frame, save Saka, who’s a little to the right. Gabriel has joined a wide diamond to help push the opponent back. This is not without risks, as the midfield counterpressing duo is Eze and Lewis-Skelly, but Atleti are so far back that they won’t be able to push forward quickly on a ball-win.

Eventually, Atleti were able to hassle Saka during an attempted attack, and Saliba and Gabriel sloppily regroup. But Atleti are just too deep to do anything about this, providing another demonstration of how the dominance of your attack affects your opponent’s ability to send players forward.

And this is the moment where Arsenal are different this year. Regrouping in the middle third, they are less likely to push forward until a genuine advantage has been gained.

For the last couple of years, some idiot has blabbed on about how Arsenal should try and keep it in the middle third a bit more.

Arsenal would often settle for “easy progression,” advancing the ball upfield on a midspeed carry without creating a genuine advantage. Once the opponent dropped in, the rhythm slowed, and possession turned into patient but predictable circulation: side to side, waiting for a crack that didn’t necessarily come against your Fulhams and Newcastles.

In What’s the Difference, we also probed for answers, looking for the ways Liverpool was more able to generate space than Arsenal. One of the takeaways:

I do believe that Arsenal can recycle back some more, and be a little more patient about advancing to the final third. Teams basically allow those advancements to happen anyway, so Arsenal should wait for more tangible advantages to accumulate to push it forward. This will help maintain the space to shoot when the final action is nigh.

Nice in theory, but this has caused problems when the available front-three don’t then win their 1v1s, or when the long-balls look like hoofs (hooves?). In this case, Calafiori splits the CBs to bring the ball up, and adequate numbers are pushed forward.

With space behind now created, Saliba hits a ball to Gyökeres, who has timed his run for the moment Rice drops lower, gaining the attention of his opponent. These factors, plus a good matchup for Gyökeres, plus a simply improved Gyökeres, turned the odds for Arsenal in this one.

Gyökeres wins the sprint to the ball, as he often did against Atleti. But an opportunity doesn’t immediately present itself. And this is the key: by progressing dynamically from the middle third, and then immediately mounting waves of pressure from there, you are much more likely to catch your opponent in a moment of disorder.

We know what happened next: Trossard probed for a shot, then ripped it off Oblak. Because Arsenal arrived at speed, you do not see a perfectly-disciplined Simeone shape: you see a bunch of players rushing back to clog the middle, attempting to make an individual block. A bit of chaos from an opponent who gives you little.

The key was the timing of our subject. Trossard’s two-footedness gives him the space to get a shot off, and Oblak parries it to his left. But Saka is already jostling for position before the ball is struck, winning position in case of a fortunate bounce.

He didn’t seek good fortune. He was himself good fortune.

We can see it even better from the Spidercam.

Those close to me will know that this is usually a bugbear of mine. Professionals are in their own stratosphere with so many skills, so I tend to complain about the areas in which they are still too close to we mortals: specifically, blocking shots (so much turning and ducking) and rebounding (so much standing and ball-watching). I think top players are often too flat-footed on parries and should essentially be boxing out for position on every shot. They shouldn’t be waiting to see where the ball goes; they should be winning the spot first, and then reacting to the trajectory.

Here’s Steven Adams, one of the best rebounders of the era in basketball, explaining the beautiful geometry of the skill.

Saka won the physical battle, but he also won the geometric battle. In the equivalent of 23.2 90s in the Champions League, he now has 21 goal contributions (or 0.91 per 90).

“In those situations, I just try and stay alive,” said Saka after the game. “Sometimes it bounces for you, sometimes it doesn’t, but you have to be there. And I was there.”

I often reach for symbolism, shamelessly tying individual actions to the wider narrative of the season, or, when that fails, quoting a spy novel or some shit. But sometimes a quote like that does my work for me.

In those situations, he said, I just try to stay alive. Sometimes it bounces for you, sometimes it doesn’t, but you have to be there. And I was there.

“It had to be someone very special,” said Arteta. “And certainly he is very special with me and the boys and everyone attached to this club. If it had to be someone scoring that goal, it probably had to be him.”

👉 Total footballers

Luis Enrique has a dream.

“Of course. It’s my dream. My dream is to have 20 players who can play everywhere. Nuno Mendes as a full-back, Nuno Mendes as a winger, Nuno Mendes as a No 9, Nuno Mendes as a midfielder. Everyone except the goalkeeper, that would be my dream.

“You can imagine the opposition manager when he sees the line-up saying, ‘But he plays everywhere…!’ It’s a dream. I know it will be difficult to achieve but I’ll try.”

He is not alone in dreaming it. Max Eberl, Bayern Munich’s director of sport, told Nick Wright something similar in a great piece: the diagrams on TV before kickoff bear less and less resemblance to what actually happens once the whistle blows.

“The defensive midfielder and the No 10 drift wide, the centre-backs and full-backs make runs into the centre... I believe true specialists who can only do one thing will become increasingly rare, simply because the game is becoming much more dynamic and variable.”

The Arsenal/Bayern group stage game was one of the more flexible matchups you’ll ever see. Here’s Arsenal building up, with Rice as an RCB, Saliba as a #6, and the striker (Merino) as an auxiliary midfielder.

And here’s Saliba following Kane all the way across the pitch.

The complication is that none of this is free. Marcelo Bielsa, the patron saint of man-marking, expressed his frustrations with the natural tradeoffs of these movements.

“When the ball is regained, the team is set up in a defensive formation. Possession is then harder to manage as on the turnover, your players are breaking from a place of having been man-to-man marking. It’s harder to launch out into attacks from whatever formation that is formed as a result of chasing down opponents.”

It requires rare players. Players who are comfortable all over the pitch, in every possible scenario: in-possesssion, out-of-possession, pressing, counterpressing, the works.

Enrique’s evolution on this is especially instructive, especially with regard to the process of building up comfort in these rotations, and the time it takes.

Before his job with PSG, he presided over fairly rigid positional principles as the Spanish head coach. His last game in charge was the round-of-16 knockout against Morocco, in which Spain logged 1,019 passes, culminating in perfectly balanced, pinning pass maps.

The only problem? In 123 minutes, this approach resulted in 1 shot on target, a total of 0.69 open play xG, heartbreak, and an exit.

With less training, and shotgun marriages all over the pitch, a stricter positional portfolio comes with the territory of international play. The current Spanish squad keeps many of the same principles, but with a key difference: more time together, and better players. Likewise, Enrique’s transformation of PSG into the swashbuckling side we know now took time, and it should be said, downright offensive amounts of capital.

“The team started the season with principles that were already well assimilated, and the coach tried to incorporate even more mobility,” said Vitinha last year. “Today, a #6 can be an #8, an #8 can be a #10, a #10 can be a #6 and with the forwards, you never know if they’re on the left, the right or in the middle. It’s very difficult for opponents.”

Eric Löwendahl called it “Adaptive Positionism,” as Enrique increasingly sought players with the physical and cognitive qualities to succeed in such a system.

When I speak of adaptivity, I am referring to the player’s ability to read where in the opponent’s defensive structure there is an open space to receive the ball relative to the ball-carrier’s position and the position of local opponents. To deviate from a predefined position/area on the pitch and attack a space to create an overload as a third player or to read the game to become the fourth player in a play where a third-man play has already occurred.

None of this sounds foreign to an Arsenal ear. After the Kai Havertz signing, Amy Lawrence wrote about the Arsenal philosophy of targeting “general footballing qualities and human characteristics, [rather] than a specific positional profile.”

In the original Merino scouting report, I wrote about looking beyond positions.

I think it’s helpful to take a more expansive look at squad lists, and look less at “positional cover,” and more at “profile cover.” A player doesn’t need to slot in at the exact same position to effectively cover a starter; their qualities need to be covered, however. For instance, Jorginho and Zinchenko were essentially spelling each other last year … at different positions. If Saka misses time, Arsenal will be light on 1v1 gravity, dribbling, and playmaking — that needs to be offset somewhere, but it doesn’t necessarily have to be at RW. As such, Merino is great Rice cover, even if he never plays in the #6 … He is also, in his own way, Havertz cover. Especially down the line, it got hard to picture lineups without Havertz in them — especially when you’d like to go long or take advantage of set plays. Merino covers that, too.

April was a difficult month. The return from those lows has largely been about a) returns to fitness and b) some crowd-pleasing lineups made up of complementary profiles who are comfortable in zones across the pitch.

Declan Rice has typified this change. While he’s usually had a fairly standard range of responsibilities on the left side of the double-pivot this season, things came to a head: Zubimendi has looked fatigued, some long-ball strategies weren’t working out, and the team needed something new. This pushed Arteta into updated lineups over the last few weeks, essentially settling on a new one. At their heart, these changes have required Rice to adapt his role on a match-by-match basis. They are possible because of him.

Arteta’s comments read like a copy-and-paste of Vitinha’s sentiment earlier.

“What we need are players who can occupy any position on the field. Modern football is going in that direction. When you’re a midfielder, you’re a midfielder — not a six, eight or 10. If we want a total team, we’d better have total players to do that, to be comfortable and dominant in any area of the pitch, to dominate every phase of play. Declan is certainly evolving in that direction.”

👉 Arteta’s gut feeling

It’s easy to say that selections can be overindexed in the mind of a fan. Ultimately, it’s a player’s game, and they have to perform, and it’s not difficult to retcon narratives when results don’t come. But I’ll go the other way: there’s really nothing more important than what players are on the pitch, and how their qualities complement each other. Even more, there are opportunity costs to every decision a manager makes: by starting a runner in a closed game type, for example, you lose the ability to throw him on late against tired legs. Arteta has done a tremendous job this year to navigate the increasing demands of the league and a punishing injury list, all while shepherding a team that never lacks for effort. But there have been spells when lineups didn’t feel quite right, as I covered in my Triangles piece.

With the attack sputtering, Arteta devised a lot of lineup permutations on his iPad. He said he’s leaning on his gut more these days.

“There is something that is related to your intuition. What the game is going to require, the state of the players, the way you can imagine the game and where the players can have the most impact.

“You can get it right or wrong, you never know. But if you do what you feel, at least you have the certainty that you’ve done the preparation, you’ve done the thinking, and when it comes time to deliver, that’s something else.”

The first sign of this was not Fulham, but the first leg of the semi-final, played away at the Metropolitano. That runny, inevitable front-three (Gyökeres, Martinelli, Madueke) started again, but there weren’t many other options, and that kind of lineup can suit an away first leg in the Champions League. Calafiori and Saka were working their way back, but not ready to start, and White was gritting through.

The most interesting tactical change, then, was that Rice and Zubimendi largely switched sides, and something closer to a single-pivot was played. Rice was mostly in the deeper role, with Ødegaard and Zubimendi ahead of him.

Considering the stakes, the location, and the available personnel, this makes all the sense in the world, and is something that should be considered more next year. Rice is a unicorn as a defensive midfielder deep, and Zubimendi can do some interesting things on the edge of the box. Freed from some of his usual, grittier responsibilities, and with the clarity of mind that comes from having Declan Fucking Rice behind him, Zubimendi had one of his best games of the calendar year by combining in those higher pockets.

Some of this work helped shepherd the Gyökeres penalty:

White was fighting his body, but otherwise, it was an assured and mature performance by all.

But ultimately, something needed to change. It had been 8 games since Arsenal had scored multiple goals.

👉 The new dynamic

The attacking play can’t be understood without acknowledging the most obvious factor: Bukayo Saka was starting again. Once he started at Fulham, things started to look better. He is perhaps Arsenal’s best player when healthy.

Right from the start, Arsenal would manufacture situations like this after dead balls.

And he sent Jiménez out for some lunch.

He kept up the pressure with wide combination play.

Then he and Gyökeres continued their improved relationship with another goal.

It was 3-0 by half.

In addition to Saka’s return, there have been some welcome improvements to the overall play. There were a few reasons behind this:

• Never underestimate individual fitness. Zubimendi was looking tired. This is understandable, as he’d run more than anybody in the league, and has already appeared in 54 matches for Arsenal, logging 4,171 minutes. We’ve seen players like Havertz and White hit a similar fitness wall in the past and soldier through because of a lack of options. It was heroic shit. But while we’re grateful, we also know the cost. It should be seen as nothing but a positive that Zubimendi can rotate now if needed.

• Lewis-Skelly gels well with Rice (who can cover space for him) and Calafiori (who can combine and switch spots with him). His role hasn’t been perfectly linear or straightforward during this spell: he’s been deployed in subtly interesting ways.

• In this shape, Rice can stay closer to the CB duo. Relatedly, Arsenal haven’t given up an open-play goal since Man City.

• Calafiori is vital to Arsenal playing at their vibrant best. Furthermore, the duo of White and Calafiori offer good decisions and assertive, crisp play as total footballing full-backs. We’ll cover all that later.

• Trossard is looking closer to his sharpest self.

• Thanks in part to lineups that play to his strengths, Gyökeres has improved his overall play. But also, it’s simpler than that: he’s improved as an individual player.

The buildup against Fulham had Rice and Lewis-Skelly as a central pivot, Eze roaming above them, and White and Calafiori offering options.

It’s easy to see why this just works. These are players doing what they do best.

As it progressed, we saw the box midfield. This is where everything feels devastatingly simple: three CBs who are good at CB things, two double-pivot midfielders who are good at double-pivot midfielding, two wingers who are good at wingering, two pocket players who are good at receiving in the half-space, and a striker who likes to strike.

If you draw the lines differently, you can almost make out the old, famous WM shape of Herbert Chapman.

Then, it was time for the second leg of the semifinal, and the previously-mentioned iPad.

“It was probably the hardest line-up that I had to do,” said Arteta. “I don’t know how many lineups I did on my iPad. I deleted them and started again all over. And I really felt that leaving the players out tonight was going to be hurting for some of them. At the end, I went for my gut feeling. I spoke to the ones that were not starting and their reaction was incredible. And you saw when they came in the game at a difficult stage, they were incredible. I think they deserved it [to hear from me], especially the deserve [to hear] how I feel about them.”

He ran it back. According to The Athletic, it was only the second time all season that Arteta has been able to repeat a lineup.

👉 The daisy-chain

But this time, it was a little different. And, contrary to what you may hear on talkSPORT, it was quite risk-tolerant. There were ample rotations, but generally speaking, Lewis-Skelly was often the lone-6 as play moved into the middle third, with Rice pushing forward on the right. This makes a lot of sense in-possession, but on a ball loss, you have to consider the impact of Lewis-Skelly being left all alone, with Rice likely ahead of the play. Arteta knew the risk and accepted it. Rice kept pushing.

Outside of the out-of-possession “risks,” there are a lot of benefits: Rice gets to run and peel players off Saka, and Eze gets to push more central. Once White pushes forward, seven different players can join the frontline.

In deeper build-up, it was often a truer double-pivot, with the open RCM slot being open for business: this is where we saw Saka float and playmake.

But the real gift of the Rice/Lewis-Skelly partnership is that they’re good at different things, and that difference makes them more dangerous together.

Think of central progression as a daisy-chain: every link needs a player in a threatening position. But instead of calling it “breaking lines,” look at them as gates. Picture a training drill where the ball has to pass through a sequence of checkpoints on its way to the goal, and each gate only opens if someone is positioned to receive, turn, and threaten. Now, see this:

Rice is best when facing play. Lewis-Skelly is superlative at sitting behind the first gate. And Eze is superlative at sitting behind the second gate. Simple as.

You’ll see the daisy-chain effect here. Rice actually plays a switch, but there are multiple coherent arguments for central progression here.

…and again, you can see the gates: Rice outside the shape, Lewis-Skelly tucked behind the first gate, and three players ready behind the second gate.

This play, early in the Fulham match, shows the setup humming. Rice drops as a libero, Lewis-Skelly receives behind the first gate, and Gyökeres finds the perpetually free far-side full-back.

But how different are things, really?

👉 Zubimendi & Lewis-Skelly

In the original Zubimendi scouting report, I covered his unique sensibilities as a Spanish midfielder.

He is constantly moving, scanning, and setting up angles, but as a lone-6, he’s not over-eager to get on the ball or looking to rack up 100 touches. He’s more situational and flexible; he’s talked in interviews about how sometimes he thinks he should get on the ball more, before remembering his responsibilities in his setup.

Because he’s usually outnumbered by pressers, he’s often dragging them away from the 8s, creating lanes, and freeing them up to turn and attack. If the press has gaps or is more mindful of the 8s, he pulls all the strings himself and is happy to drop all the way down and split the centre-backs if needed. If there are numbers high, he pulls the pressers and they go long as he pounces for second balls.

So much of a player’s touch quantity is informed by the system and the player’s responsibilities. But you don’t want to neatly explain everything away, and there is definitely an archetype of a lone-6 that says, usually with ample reason, “just give me the fucking ball. I’ll figure this out.” Zubimendi appears humbler than that.

This has persisted for Arsenal, as Zubimendi occupies every zone on the pitch in buildup.

Plus, he has the Saliba trait of defending through anticipation, which makes some of his eventual duels look easier than they should be.

I’m not sure people fully grasp what kind of defensive shift Zubimendi has put in, but according to Gradient Sports, he is the…

• #1 ranked midfielder at duels in the league

• #1 ranked midfielder at aerial duels in the league

• #5 ranked midfielder at 50/50 balls in the league

We’ve covered how Arteta sees the tactical world shifting, and how it has resulted in both a swirling double-pivot and a wider, freer full-back role.

“Before, when you used to do a game plan and you just invert a full-back and bring an extra player in midfield, or a false-nine, opponents are finished, big overload, 4v3 inside, 2v1 inside, time on the ball, so dominant, 70-80% of possession, the other opponent, two counter-attacks, set-pieces, the game is done. Now, teams are adapting.”

It sounds like he believes a fluid midfield double-pivot is the best response to the rise of man-marking, which “un-inverts” the full-back and keeps them wider, which also gives them longer transitional runs. That tilts his equation toward Hincapié.

I’m not 100% sold on that in all situations, and yes, I think Lewis-Skelly’s skillset is well-suited for the floating “double-6” thing that Rice and Zubimendi are doing. It may just take some time to get there.

In August, I wrote this:

• “Calafiori role” (LB), not LCM, is floating around the pivot, going wide, box-crashing, unlocking new “Rice role” at LCM: deeper, shielding.

• “Cala role” (LB) is less suited for MLS; the LCM role now fits him like a glove.

It’s taken some time to get here. While I’ve been impatient to see it more, we can also acknowledge that 17 starts and 33 appearances across competitions for a 19-year-old isn’t exactly nothing, especially considering how many of his ilk get run into the ground by overeager managers. There’s a 2,452-minute difference between Zubimendi and Lewis-Skelly, and the former has shown signs of fatigue, so when you combine that with a development curve behind the scenes, the timing aligned nicely.

Arteta was asked why it took so long.

“Because probably I don’t have a clue and maybe I should have done it earlier, I don’t know. But I have to do things when I believe that the player is ready, the team is ready and the opponent is the right one to play with him in that position. We've done it today, it’s the first time. It was a big risk because I knew what was going to happen, if he wasn't this great, we would have lost the game. How do you play a kid at this age, in this scenario, in a position that he hasn't played all season? I knew that but I had the feeling that it was the right game for him.”

If I had to hazard a guess as to what they’ve been working on, it’d be how Lewis-Skelly defends transitions through the middle. He’s a gambler by nature, and if that style persists in the fastest league in the world, he risks too many pull-downs and awkward fouls, or else being played through. By being a little safer with his angles, he’ll keep the play in front of him. Arteta deemed him ready. Rice helps.

“It’s hugely important for me to set that tone,” said Lewis-Skelly. “When you’re in midfield, there’s a certain rate you have to attain, which is to set the tone, get the engine of the team going. I thought I did a good job of that today.”

So, how has the introduction of his style actually impacted Arsenal’s play in the middle? I pulled the numbers to compare.

Interesting:

• What you’ll see in the passing numbers is that they are remarkably similar. They receive and complete about the same, though Lewis-Skelly is marginally more accurate.

• As a narrative violation, Zubimendi passes forward more, and Lewis-Skelly passes backwards more. Now, all forward passes are not created equal, and some of the best passers in the league are those who pass backwards the most (Rodri being a shining example). Still, it’s an interesting context.

• Zubimendi is more likely to go long, at a lower success rate. I’d chalk this up to what he does on regains: opportunistic first-touch slaps forward into space.

• Zubimendi has been much more active around the box. For all Lewis-Skelly’s rambunctiousness in the middle third, Zubimendi is more likely to offer chaos in the box. Not mentioned here are his 6 goals and 3 assists across comps.

• So, it’s all quite similar. But there is one area in which Lewis-Skelly and Zubimendi are dramatically different players: Lewis-Skelly is ~4 times as likely as Zubimendi to engage in an attacking duel.

In a situation like the one below, Zubimendi would probably tap it forward on the first touch.

Lewis-Skelly treasures moments where he can pull a Mousa Dembélé and shove them off. When these moves work, they’re especially powerful for Arsenal, because they create a total collapse of the opponent’s shape, which is a situation that is hard to generate otherwise.

We saw one of those moments recently when he set up a Nwaneri goal for the England U21s. You get the sense that teammates almost delight in passing him into difficult situations so they can see him twist out of them.

He’s one of the best foul-getters in the game.

But, it should be said, these duels are riskier in the middle of the pitch, which can make him more partner-dependent as a midfielder. You want Lewis-Skelly to duel and be himself, but you also want somebody who can quickly sweep up a ball like this (which Madueke does here, naturally).

I was pithy on social media and said, ‘“Zubimendi ran more than anyone in the league and now looks a bit tired, so thank God Lewis-Skelly is here to ball out” can be your entire take on the situation, you know.’ Lewis-Skelly has been as impactful as you could have reasonably imagined, as a Chekhov’s gun pulled out in the third act. At a certain point, it felt almost too late in the season to be having this conversation. And yet, here we are.

Zubimendi absolutely still has his place. And with some rest before the Champions League final, I’d still be strongly considering playing him in that left-8 role that he played in the first leg of the semifinal.

👉 The all-important full-back question

Tomiyasu showed us how it’s done.

The role of the full-back is all-important these days. One of my feelings is that top coaches have essentially hacked the ideal role as the all-action, perpetually free man. But the physical demands are unsustainable, so you need a superhuman athlete to defy gravity, or ridiculous rotation.

Injuries and all, in five seasons at PSG, Nuno Mendes has played 40.2% of league minutes; he’s started-and-finished fewer league games than Ben White did in one year at Leeds. His playing time magically jumps up to 75.4% of available minutes in the Champions League. And remember the example from Enrique’s dream: “Nuno Mendes as a full-back, Nuno Mendes as a winger, Nuno Mendes as a No 9, Nuno Mendes as a midfielder.”

The FIFA Training Centre recently wrote Innovation: The emerging positioning of the inverted full-back, in which they tracked some of the evolving responsibilities. There are definitely some sample biases here, but you can also see how full-backs are getting more interior.

What they found:

• Full-backs received the ball 10 meters inside the sideline in the final third, which was 1.6m narrower than at WC2022.

• Inverted full-backs had 23.1% higher reception rate in inside and central channels than WC2022.

And we’re seeing typical “inversion,” that idea of Zinchenko folding into the double-pivot, a little less. What is more often the case is that full-backs are floating around as a third man, looking to exploit jumps by the block, and create overloads. On the far side, they are often left unmarked as hybrid pressing schemes try to squeeze one side.

But you can see, just in those arrows, how far the distances are to travel, how many zones they might have to be comfortable in, and what kind of tactical and physical load they are under. As instructive as his stint was, Tomiyasu was not known for his health.

For today, the point is this: in such a shape, some of the most decisive touches are going to come from the full-back. They will find themselves in dangerous territory, with space that teams haven’t planned for. If things are looking at their fluid best, you may have a full-back to thank. If not, same.

Here is a pristine example. With Atletico sending four players forward in a passive pressing front, the free man is over the press to White. This is the kind of ball that would have been nice to have in the League Cup final. But then, play squeezes onto the Arsenal right, so the point of attack changes, and who is the free man after that? The other full-back: Calafiori.

Less of a “total football” skillset from those two, and that shot never comes.

In more straightforward moments, you can still see how easily the full-back can be found unmarked, particularly if an opponent is caught between a zonal and pressing moment.

No, this isn’t a replay.

These players also play vital roles in the balance of wide triangles.

As I wrote in, erm, Triangles:

It’s oversimplified, but in any outside trio, I’d like to see a few adjacent profiles:

• (Below) An orchestrator: somebody who can dictate play and pick apart the opponent shape with passes.

• (Beside) A dribbler: somebody who can drive at opponents and collapse the shape, starting a cascading effect.

• (Beyond) A runner: somebody who can interpret space and pull it apart: “decompacting” the shape horizontally, diagonally, or vertically.

Ben White is so valuable and difficult to replace because he is many things in one: he has the controlling instincts of a midfielder, and the tireless runs of a wing-back, all in the body of a former CB. This makes almost anything flow better because of the natural balance he provides to other profiles. Remember, too, that his “late return to form” also coincided with Saka’s return.

The same is true on the other side. Lewis-Skelly and Calafiori work together in a way that is easy to understand. They can swap positions, they can both play midfield, they can both play left-back, and most importantly of late: Calafiori offers a credible option behind the “gate” for Lewis-Skelly to punch it through to.

We look for triangles with an orchestrator, dribbler, and runner. Like Nico O’Reilly and Nuno Mendes, Calafiori is all that in a single player, which balances out the profiles around him. Physically or temperamentally, these players have more in common with, say, Jude Bellingham than they do with a more prototypical full-back like Kieran Tierney.

Most of the early passage of play against West Ham looked like this.

Mosquera arrived at the Metropolitano and immediately made the argument for how dangerous these full-backing touches are.

My points, to close this thing up:

• Full-backs are indescribably important right now. Defending 1v1, contributing to build-up, making long recovery sprints, operating all over the pitch, and most critically, arriving at the decisive moment with the ball at their feet. The ask may indeed be unsustainable over 50+ games. Going into next season, I’d like to add a hybrid midfielder capable of covering at right-back, the way Lewis-Skelly currently does on the left.

• It is no surprise that Lewis-Skelly is thriving specifically with Calafiori beside him, floating and swapping but always providing an option in front, and with Rice alongside, making the big recovery runs and cleaning up, giving him the confidence to turn and carry. I would personally argue for Calafiori to start every game he is fit. Him for 60’ and Hincapié for 30’ just works.

• White’s loss looms large. I cannot adequately describe how grateful I am for everything he has done and sacrificed for this club.

• I would be hesitant to pull Rice out of the middle again. Keep the partnership with Lewis-Skelly rolling in the league. The only question is who starts the final, and depending on fitness, I remain open to Rice at right-six and Zubimendi at left-eight, as we saw at the Metropolitano.

• Timber’s return is vital for that. In the meantime, the cleanest answer is Mosquera at right-back. He is doing fairly well; his spatial awareness and confidence in manipulating the first line are particularly interesting. But as we’ve established, the full-back is often the one receiving the free-man touch, and their read of that moment is frequently decisive. He is a talented young centre-half. In forward areas, he usually looks like a talented young centre-half. If we are going to use him up there, it might be fun to have him as a target man, or back-post headerer. Use his strengths.

• For the two remaining league games, I would be, at very least, interested in starting Zubimendi at right-back. This is not a new idea. It was something I wanted to explore as an option when he was signed. But if his superpowers are tactical intelligence, work-rate, spatial interpretation, quick combo play, and the ability to duel and generate chaos in final actions, I would greatly prefer to keep Calafiori in his best position (with Trossard and Lewis-Skelly to play with), the Lewis-Skelly and Rice partnership intact, and try something new on the right. Zubimendi seems almost agnostic to the zone where he plays, and his two-footedness helps everywhere. The obvious risk is the physical ask and the recovery runs down the line: it is not lost on me that we’ve just been talking about Zubimendi’s physical fatigue, and now I’m considering putting him into a position that can be super demanding. It’s marginal, but my hope is that Saliba and a deeper Rice can cover most of that against Burnley and Crystal Palace; and my real hope is that we have so much of the ball that there isn’t much running in the first place. If it were up to me, I think it’s the route I’d take. If we believe that the position now belongs to total footballers, Zubimendi is certainly one.

👉 Issues at West Ham, and what to learn

The initial lineup against West Ham was to run it back, again. But there were tweaks: Lewis-Skelly was essentially a lone-6, with a diamond formed around him, all designed to overload the back line.

This represents some real commitment from Arteta. It provides between 3 and 6 options for Lewis-Skelly to distribute to, but it also can leave him alone with the CB’s to defend transitions, with Rice sprinting back. Arteta decided it was worth the risk, and the team looked good in the opening period.

This kind of play happened early, and kept happening.

I feel fairly confident that the pressure would have kept squeezing West Ham, and a goal would have come before too long.

Once White went off, though, things lost their thrust. Rice was inserted at RB, and Zubimendi came on to form a double-pivot with Lewis-Skelly. I understood this move better after rewatch: Rice was playing a high-pocket RCM role anyway, so the thinking was just to keep that going with Zubimendi, keep Lewis-Skelly doing what he was doing, and Rice could continue the good work he did at Brighton at RB.

I’d offer two reasons for this not working out:

• The vibes were off on the right flank. Rice just never looked comfortable or particularly assertive over there this time. Yes, the vibes were off is analysis, thank you.

• Lewis-Skelly and the current version of Zubimendi didn’t complement each other in spatial defending. With Lewis-Skelly deeper, there were distance issues.

• From there, Calafiori went off at half, and Lewis-Skelly went to left-back. This was understandable, but it put Lewis-Skelly at a less advantageous role, kept him outside the block, and separated him from Rice.

Here’s a perfect example. Summerville receives the ball, and instead of getting goal-side, Lewis-Skelly gambles for a tackle and leaves Summerville open to carry forward. From the other midfield slot, Zubimendi tries to mop it up, but is a step late. Normal, but this is the kind of run that the £105m alien closes down.

In the second half, we saw a shift to a new setup.

Eze is free to receive. But this is giving Lewis-Skelly those fewer, higher Calafiori touches; it’s giving Rice, Mosquera, and Eze the touches of a floating #10, where only Eze is at his best there. Basically, the receivers in the “first gate” (behind the front-two pressers) and the “second gate” (behind the midfield) aren’t as dangerous as the starters were.

Things changed later, as Arteta sought a lineup that could pin and overwhelm. But much of the game stayed cagey for too long. My lessons were simple: Zubimendi and Lewis-Skelly may work as a duo under peak fitness, but not right now, and especially not with their distances too far apart. And we’d better make damn sure there’s some liveliness in those pockets. Full-backs are highly important for that.

The decision at full-back has been weighing on Arteta.

“Declan Rice, when he’s played as a full-back, against Brighton, was exceptional. But then the other night, it was something else. It depends. The sample is still so little, it will remain so little from here to the end of the season to judge if a player is good enough who can perform in the position, and what is the cost of moving that player from one position to the other, and that’s the balance that we need to find.”

👉 Ødegaard the lockpicker

As the scoreless draw continued, the spaces seemed to close instead of open. Look at the vertical and horizontal spacing for West Ham.

Last time, we talked about the difficulties of some of our lineups to unlock a situation like this.

“When facing a lower block in the Premier League, they are not the ideal candidates for pinning. Pinning requires a mix of retention, patience, and, in the decisive moment, the ability to execute a refined tight-space action: a snap shot, quick dribble, or slipped pass. As a striker, you’d want to dominate in both reaction time and aerial physicality. As a winger, you need to win a crowded 1v1.”

This meant that Arteta wanted personnel to match the moment, and turned to Havertz and Ødegaard. We can see how much prodding the situation required.

This is where Ødegaard is just special.

We’ve had years of Ødegaard getting the ball into the box more than anyone in Europe, and many assuming it was because of his role or positioning or whatever. Really, it’s just hard to work the ball into a packed 10-man box, and he’s great at it.

Even in a year such as this, here’s Ødegaard in the Premier League this year, amongst all players, via Opta and Gradient:

• #1 in through-balls/90 (1.14)

• #1 in passes breaking final line (1.42)

• #2 in open play assists/90 after Cherki (0.35)

• #4 in open play chances created/90 (2.04)

• #9 in xA/90 (0.25)

The recent setup (Calafiori at LB, Trossard at LW, and a Rice, Lewis-Skelly and Eze midfield) has worked well, and I’m in no rush to change it, especially with lingering questions about Ødegaard’s fitness. Ødegaard has also been closing out games well, and if you need to change the dynamic, you can just put him on quite early. That kind of luxury is why you buy an Eze in the first place. But with low blocks a-coming, it’s tempting to start Ødegaard again.

I’ll call out two other aspects of that play, as well.

The first is the entry pass from Rice. I don’t have anything more interesting to say than I think he’s looked good combining on the right. He has more experience on the left, but I seem to prefer his combination play and crossing angles over here.

Another. (Marginally offside, I think)

Weeee.

…and the final point there is Trossard, who has notched an assist to Gyökeres, the rebound that led to Saka’s goal, and the winner at West Ham, right when we needed him most.

He’s so interesting because of his ability to recreate himself every year. He went from 1 goal and 10 assists in his debut campaign to 12 goals and 1 assist the next. This year, he’s been more of a wide facilitator, adding some interesting tempo things out on the touchline, and seems to have a natural understanding with Calafiori and Gyökeres down that left side. He had a down period, but what he’s really been exploiting of late is his ability to slow things down and keep defenders off-balance with his two-footedness. His old teammate at Brighton, Pascal Groß, has made a career out of chopping and crossing, but Trossard has the advantage of being even more genuinely two-footed with his chops, which creates the kind of anxiety in defenders that Son Heung-min used to generate before a shot.

It was nice to see him lurking at the penalty spot before the goal.

👉 Gyökeres

I’ve been asked whether Gyökeres has been improving or whether he has just faced some good matchups. The Atleti performances could feasibly be kept in a compartment, as he can beat them physically. But Joachim Andersen and Calvin Bassey are no slouches.

Looking through the footage, I see two reasons for his recent uptick:

1. Surrounding him with a cocoon of passers high (Trossard, Eze, Saka) and even low (Calafiori, White) suits him far better than playing him with looser, running types.

2. Yes, he has just … gotten better at playing football in Arsenal-type game-states.

The clearest example is the timing of his darts in the box. In the past, he’s struggled with reactive movement, waiting for the ball to get hit to make his run, often arriving a step late. The problem, as Gary Lineker tells us, is that you’re then confining yourself to the same information that the defender is working with. Ideally, you have a feel for your wide attackers and can better predict when the ball is coming in and gamble as a result.

For both goals against Fulham, he started his move before the ball was hit.

Here, he drove into the space instead of tracking the ball.

And yes, your eyes don’t deceive you: his dropping to link up play has gotten better. It’s still not enough to be considered a genuine strength, and it’s probably not his touch that is improving, but his level of certainty: if you’re clearer about what you want to do with it, you’ll touch the ball with more confidence.

I’ve been satisfied with my preseason read on Gyökeres. It laid out my reservations, but also:

I also don’t think anyone should discount the possibility that he is a stubborn motherfucker … If he’s fast, elite with his intensity, high-level tenacity and game mentality, and has good ball-striking, it’s hard to be too blasé about that skillset. He’s [also] durable and ready for a Premier League calendar.

Almost every professional footballer is headstrong in their own way. Gyökeres seems to be in the 99.9999th percentile of that.

👉 Declan Rice, he play nice

You asked me for my votes for Arsenal Player of the Season? Wait, you didn’t? Ah, you’re getting it anyway. Along with some grades from Gradient.

1. Declan Rice (Gradient: #1 rated overall player in the Premier League)

2. William Saliba (Gradient: #1 player at positioning in the Premier League)

3. David Raya (Gradient: #1 rated keeper in the Premier League)

On Saliba: I watch a lot of football, and am certainly not free from bias. But I do think Saliba’s the best defender in the world, and is absolutely vital to everything we do. Van Dijk is now a step off, so if you asked me which world defender I’d want to start a team with, I’d pick Saliba every time. He should not be undercelebrated for making everything look so easy.

My dumb post on bluesky:

remember how arsenal weren't challenging for titles and then saliba got in the team they immediately started challenging for titles and then he got injured and they lost the title and then he got fit and they started challenging for everything again

👉 Looking ahead

For Burnley and Crystal Palace, I don’t want to disrupt the chemistry of Trossard and Calafiori, or Calafiori and Lewis-Skelly, or Lewis-Skelly and Rice. Mosquera is the most likely candidate at RB, but I’d prefer a little more on-ball menace. And I assume that Havertz and Ødegaard are still being managed physically.

As such, here’s where I’d lean at the moment.

Zubimendi can offer a lot of the combination and half-space play that White does, and he also has a chaotic, unexpected streak that can create moments against overmatched teams: we’ll need to send numbers forward, creating diamonds instead of triangles.

To me, this feels like a good balance between “keep what’s working” and “injecting something new.” The team retains some great finishers, as well. But as always, so much depends on a clear understanding of physical levels, which is something the team has a better view of.

For my next piece, I’ll work on a full PSG preview and opposition report. Really need Timber for that one. Getting him fifteen minutes against Palace would mean the world, and Lord knows he deserves it.

🔥 Wrapping up

Going through this all has been my form of stress-eating before the three to come. I am confident, yes, but also … all the other things. Perhaps reading it was a form of stress-eating for you, too. It is, after all, a BBQ.

You’ll notice I made it an entire newsletter without mentioning VAR once. That is because my friend and writing confidante, Tom, already covered the narrative of the last week so perfectly. Fuck ‘em.

Last month, I wrote about what it means to grow up and to take on the weight of what you’ve entrusted yourself with. I closed it like this:

While we get nostalgic about childhood and wax poetic about simpler joys in simpler times, those are simple joys. Simple joys are found at the beach, and there’s nothing wrong with them. But the deepest, most resonant joys are hard-won, and they come from a sense of responsibility, to our community, to our family, to our children, and maybe even to strangers on the internet. (Hi, I love you.)

On a grand scale, the same feeling applies, I suspect, for the players. Whatever resonant joy they find, if they find it this season, will come from the deeper place, from the responsibility they have entrusted in themselves, that the club has entrusted in them, and that we’ve all entrusted in them to carry on our behalf for another month or so.

I don’t know what will happen. So much depends on the lineups and the performance levels; from there, happenstance and variance will do their work. But I agree with Tom Hiddleston, and I agree with Amy Lawrence: what a fortunate place to be in, to feel the weight of such expectations. You can’t do something consequential without consequences.

The decision to start Lewis-Skelly in midfield, in a Champions League semi-final no less, was a tactical one, sure, and one that would carry consequences. Arteta meditated on it at the presser: “I knew what was going to happen, if he wasn’t this great, we would have lost the game. How do you play a kid at this age, in this scenario, in a position that he hasn’t played all season? I knew that, but I had the feeling that it was the right game for him.”

As much as it was tactical, it was also an emotional bet. He was betting on the idea that the right jolt at the right time can shift things beyond the midfield. And that was Arteta’s responsibility; Lewis-Skelly’s responsibility, then, was to deliver. You make the call, you play the game, and you live with what happens next.

I keep returning to that Saka quote about the rebound. In those situations, I just try to stay alive. Sometimes it bounces for you, sometimes it doesn’t, but you have to be there. And I was there.

Five and a half years of work to shake markers and get into the right spots. And now, untold anticipation, two league games, and a plane to Budapest, where a tunnel is waiting and the anthem will again blare.

I watched The Big Chill a couple of years ago. If I remember correctly, it’s an examination of what happens when the promise of youth collides with the conventional realities of adulthood. A shared loss reunites a group of friends, and they are equal parts sad and funny and confused and melancholic about their lot in life. I seem to recall a young, sardonic Jeff Goldblum giving a memorable monologue about the importance of self-serving justifications, how they’re more important than sex, because he couldn’t get through the day without two or three juicy ones. Some members of the group have adjusted fairly well to the demands of their thirties, all considered. Some have not.

“Idealistic young people learn harsh truths about a cold world” is not a novel subject. Wasn’t that what The Godfather was partly about? The way Al Pacino’s eyes grew baggier and more distant, as his view of the world grew too, um, accurate, and the weight of his own choices piled up? Wasn’t that what F. Scott Fitzgerald explored in This Side of Paradise, talking about a new generation, who, “shouting the old cries, learning the old creeds,” and through “a revery of long days and nights; destined finally to go out into that dirty gray turmoil to follow love and pride,” only to grow up to find “all Gods dead, all wars fought, all faiths in man shaken.” I’m near-certain I have that one right; I looked it up on Wikipedia.

I’d usually describe these scenes a little better. Then I’d weave it into an analogy or something. But I’m tired. Adulthood has been catching up.

One mental model of this Arsenal project is the meticulous work of taking a starting eleven, piece by piece, and handing it to the young, under a manager who had never managed before. Saka, the face of the franchise, whose old nickname literally ends with “-boy,” and his captain-in-crime, Ødegaard, were joined by the likes of Martinelli and Saliba. Granit Xhaka made way for Declan Rice, or Kai Havertz, I can’t remember which, and even signings like Calafiori and Timber were younger than they seemed.

Along the way, veteran arrivals started to balance things out, and the likes of Smith Rowe and Nketiah bid adieu. Ødegaard grew a goatee to show the passage of time.

Last summer, the die was fully cast.

This evolution, this next stage of the process, was most apparent in five acquisitions. The 22-year-old Benjamin Šeško was deemed too expensive and developmental for the moment; the 27-year-old Viktor Gyökeres was considered a better investment for a team desperate for goals and anxious to score them right now. Martín Zubimendi was signed at a prime 26 years of age. The Premier League-proven 27-year-old Eberechi Eze made his romantic homecoming, and “Premier League-proven” can be a pricy phrase. The chosen backup keeper wasn’t a developmental prospect, but an affordable 30-year-old in Kepa. The rotational midfielder was Brentford captain Christian Nørgaard.

These signings were offset with some younger players, to be sure, including another intriguing Premier Leaguer in Madueke, a battle-hardened scrapper in Hincapié, and a readyish defensive prospect in Mosquera. But the overall age makeup of the squad tilted up. In his first summer at the helm, Andrea Berta sent an unmistakable message: the time is now. Hale Enders like Lewis-Skelly and Nwaneri found minutes harder to come by, and the squad’s average age, weighted by minutes played, is now 26.8 years old. That’s 11th in the league, right in the middle of the age table, a table I know you monitor with your morning coffee. Expectations were well and truly here, and those old justifications, those of “phases” and “inexperience,” all “juicy” in the words of Jeff Goldblum, would receive duller comfort.

And so The Big Chill refers to that cold walk into real life and its consequences. It’s fitting that Mikel Arteta, in his pre-Sporting press conference, kept returning to an antonym.

“Fire. I’m on fire. That’s it, nothing else. I’m dreaming so much. I’ve done so much to be in this position because I know how this club was ... That’s what has driven me every single day - zero fear. Fear I had when: ‘oh, if we don't get this done, this club, I don’t know what is going to happen.’ That was fear. Now there’s no fear. It’s just purpose, fire, direction and conviction that we’re going to do it.”

The words had their reasons, and were more than a metaphor: apparently, that madman lit a real fire at Colney. After two consecutive cup exits, only the Champions League and Premier League remained. A couple of days prior, a shambolic performance against Bournemouth resulted in all of 0.18 total open-play xG, confirming the worst fears of the fanbase and the greatest hopes of our rivals.

Things got dark. At the time, OptaJoe told us:

After only suffering three defeats in their opening 49 matches of this season in all competitions (W37 D9), Arsenal have since lost three of their last four (W1).

The second leg against Sporting was put-up-or-shut-up time. When the sharp end of the game arrived, the moment called for experience and technical security, and dare I say, a little of that fire. It may have been subtle, but we got it. The line below denotes the substitutions of Trossard and Jesus.

In football, fire sometimes looks like that: the steely determination of communicating well, not getting pulled out of shape, not doing anything stupid, and keeping the fucking ball close to your foot. As much as we all can get excited about the Next Big Thing, you can ask Chelsea fans whether they’d trade some of their Football Manager talent for a few steadier hands.

After what we’ve been through, sequences like this count as an uproarious highlight.

That is the feeling of arriving at adulthood, seeing the situation for what it is, and instead of sheepishly wishing it away, facing your responsibilities with the experience you’ve earned over the years.

There is another form of fire. Nothing is more fatal to The Big Chill than the sight of a young person, brimming with talent and ideas, their life in front of them, bursting through to remind us of what experience alone cannot conjure.

Manchester City, that most villainous of villains, now lurks ahead, and everything is possible. Professional determination will be required. Hopefully, some youthful exuberance makes a showing, too.

At this stage, it would be hard to overstate how little I care about which style of play gets us to where we want to go. Hard things are meant to be hard. And even after Bournemouth, even doomscrolling the most rabid of replies, even considering some of them myself, I never thought it was the beginning of the end. But I do reckon it is the end of the beginning.

👉 Responsibility, revisited

Early in the season, I covered the moment when Gabriel was characteristically barking instructions on a short corner, making sure the back post was covered, only to uncharacteristically lose touch-distance with his own man (Woltemade) for the goal. He later made amends with a headed winner at 90+6’.

The wider conversation was one about responsibility. I suggested that, if Gabriel was only concerned about his own mark, he would have been near-certain to prevail.

But this is what makes him an alternate captain. And in an ideal world, every manager would want every player to feel maximally responsible for everything, especially in the man-to-man era. Players should feel comfortable floating all over the pitch, they should communicate, they should point, they should cover, they should dominate set pieces, and they should defend as if their lives depend on it, every play.

With the fixture list as full as it is, the question becomes whether there are practical limits to what you can ask. I’ve always been interested in decision fatigue as a footballing concept.

“The more choices you have to make, the more it can wear on your brain, and it may cause your brain to look for short cuts,” said Dr. Lisa MacLean. “There are four main symptoms: procrastination, impulsivity, avoidance and indecision.”

[…]

“If your brain is worn down, it may cause you to become more reckless with your decision making or not think things through,” she added. It can also “cause you to simply do nothing, which can cause even more problems.”

The world clatters in to remind us of our limits. Knowing my readers, or at least liking to think I do, I don’t doubt there are a few ‘over-responsible personality types’ among you. The result is So Many Fucking Decisions: good intentions, even some good deeds, and the reliable hum of overwhelm. Clinical psychology studies link it to anxiety and perfectionism, the natural byproduct of people feeling personally responsible for every outcome they might be loosely connected to, even when they aren’t the main character.

I think about this when I remember how Ben White described the role of a full-back under Arteta.

“To play full-back for him, you’ve got to be a centre mid, a centre-back, a winger, a No 10. So, it’s been about developing the whole of my game, rather than just as a full-back or a centre-back.”

I think about it when watching Zubimendi dart across every zone of the pitch in build-up.

During his Man of the Match performance against Sporting, more than once, he helped manipulate the shape, only to find himself free on the left wing.

(And they say we didn’t sign a LW.)

I think about it, too, in the team’s lovable and frenetic runs to regain position after a ball loss.

PSV coach Peter Bosz called this superpower out after doing some soul-searching after a drubbing.

“Me and my coaches studied them: ‘What is it that they do differently to us?’. The answer is that they are outstanding in the opposition box but also their own. They get a lot of players behind the ball as soon as possible. They do it with 10 or 11 but we only did it with six or seven, and then the distances are bigger. It’s the transition.”

So, too, did Arteta after the Sporting second leg.

“When I see them tracking back when we lose the ball, the habits that they have are just amazing. That’s why I know what they are doing, these players. That is the reason why we are the only English team in the competition, because this league and this schedule takes the hell out of you, and it’s very difficult to do what we’ve done. We are not perfect, we need to improve things, that’s for sure, we recognise that, but there’s value in what these players have done because they deserve it.”

This is undoubtedly true, and a massive reason for Arsenal’s success. Questions over an unrelenting season remain, however. How sustainable is this? Would some players benefit from narrower portfolios?

👉 Physical demands

You can’t properly understand the season without looking at the players on the pitch. This graphic from Transfermarkt tells us a thousand little stories.

It has become abundantly clear that, no, Arsenal were not faking it when players like Saka pulled out of international duty. He also missed Bournemouth and Sporting, and, sadly, is now due to miss the Man City clash. According to a Miguel Delaney piece:

Some sources are adamant Arteta should have played even weaker teams in both domestic cups, so that as much energy as possible could go into the league.

The Basque can’t really help himself, though, and there are many who have expressed similar concerns about the constant intensity of training.

That’s also why the much-discussed international withdrawals weren’t some Sir Alex Ferguson tactic. National team staff could see that so many Arsenal players were close to the red zone.

I don’t really share the critiques of Arsenal’s deployment of the squad in the cups this season. I’m old enough to remember us turning up to Mansfield with an untested back-three formation and a lineup featuring two sixteen-year-olds and Madueke at wing-back. (I am about five weeks old.)

The Champions League lineups have also looked nice to me.

There are questions about the carryover effect of previous seasons. We must begin by acknowledging that good players play. I think every manager in the world places stalwarts like Declan Rice and Bukayo Saka in every important game possible, and the same is true for the first choices at CB and GK. But in particular, the physical toll on Ben White and Kai Havertz did not feel unavoidable; their inevitable injuries did not feel surprising if you were watching. I’ll always remember Havertz collapsing, game after game, even in dead cup ties, the January window open … and still playing full-90s until he finally broke.

And I do have persistent questions about selection decisions in the league: first, how often we’ve seen dynamics that simply do not work. Then, Timber and the midfield. If this iteration of the full-backing position persists, my inkling is that it just has to be rotated a bit. Meanwhile, last I checked, Zubimendi had run more than any player in the league, and is now showing the effects; most of that is unavoidable and part of the job, and it’s easy to say in hindsight, but also … there are degrees to these things, and it’s been pretty extreme. Eberechi Eze has played more league minutes for Crystal Palace this season than Christian Nørgaard has for Arsenal.

Rice was powering through illness against Sporting, per Arteta:

“Declan yesterday was shattered, he had no chance to play today, he wasn’t feeling good at all. Today, he played 94 minutes at the level that he’s done. Every single player is putting everything on the line to contribute, to give his best, and I really value that from the team.”

These players are giving it their physical all. And a graphic like that one from Transfermarkt is still incomplete, because it merely denotes pure availabilities. But with injuries come easing-in periods, lower sharpness, ducked duels, and the like. We have little idea if a player is at 45% or 95% fitness until the whistle blows, and even then, we’re guessing.

👉 A seven-game window

So in many ways, the analysis of this period is simple. After the cup exits, I offered that “a lot of everything is simply explained by which players have been on the pitch.” Of course, the opponents have their say as well, and many of them have played tremendously. But let’s take a quick tour. Of note:

• Everton (2-0): Arsenal started Calafiori, Saka, Eze, Madueke, and Havertz, and looked good against a lower block, accumulating 25 shots in all. The closers (Martinelli, Gyökeres, Hincapié, and most notably, Dowman) came on against tired legs and played their roles to perfection.

• Leverkusen, second leg (2-0): Arsenal again had the services of a balanced midfield (Zubimendi, Rice, Eze), as well as Saka. Gyökeres played with Trossard, and there was a mixture of retention and quality elsewhere. Arsenal logged 21 shots, and Hincapié helped shut down the threat of his old team. Leverkusen threatened very little.

• Man City, Carabao (0-2): No Raya, no balanced midfield (with Eze, Merino, and Ødegaard unavailable, the starter at RCM was Havertz), and Gyökeres up front. Pep created a bespoke gameplan to flummox the Kepa-led Arsenal build-up, and Arsenal never responded. With no free between-the-lines threat, and City’s superior second-ball setup, the plan worked. Arsenal trailed by multiple goals for the first time in any competition.

• Southampton, FA Cup (1-2): Kepa again starts behind a YOLO team: a lone-6 (Nørgaard) behind two 10s that played in the box (Havertz, Ødegaard). There is actually a lot of risk-taking that, with cold connections, doesn’t fully pan out, and turns into counters the other way. Southampton played incredibly well.

• Sporting, away (1-0): Arsenal turned to the conservative game-plan they often employ in away games in the league. Raya, Calafiori, and White start. Ødegaard turns down the risk dial, looking to avoid costly losses. Zubimendi looks beat, and Gyökeres ends with one shot, and again the finishers (this time Martinelli and Havertz) save the day.

• Bournemouth (1-2): A dreadfully unbalanced front-four (Madueke, Havertz, Martinelli, Gyökeres) plays like a dreadfully unbalanced front-four. It puts pressure on build-up, which is also imperfect. A low point.

• Sporting, home (0-0): Eze returns and provides the midfield some balance. It allows Zubimendi to return to his best. Gyökeres plays poorly up top. There’s little real progression provided from the full-back pairing (Mosquera and Hincapié), but there is some physical robustness. Some mature retainers join late (Havertz, Jesus, Trossard, and yes, I’ll include Dowman in that) and see out a professional victory in the tie.

The strangest part of this period is how nerve-racking the lineup drops are. So much of the stretch has been determined by who is available, who is available late, and also by whether Arteta is succumbing to some of his more (in my view) questionable instincts. I met the Everton lineup with something approaching glee, even if final actions wound up being a little sloppy until they weren’t. On the other hand, I fully processed a loss to Bournemouth before the game even started.

👉 Changes and their children

There have been changes afoot at Arsenal. Some of them have been borne out of necessity. Some are clearly intentional, both in line with a recruitment strategy and a reaction to the Bournemouthification of the league.

Let’s flatten a complex conversation into three categories, and look at some of the tradeoffs we’ve experienced as a result.

➡️ Change #1: A swirling midfield double-pivot

Since the introduction of Zinchenko, we have most usually seen a double-pivot of a “lone” #6 joined by an inverted left-back. Even Tomiyasu and Tierney would invert when played there. After Rice joined and mostly played on the left, this setup often pushed him up into the half-space, next to the striker.

This year, with the league getting more man-to-man, with Rice not perfectly-suited to the off-ball, pinning role of a high left-8, and with the addition of Martin Zubimendi, it has been closer to a true double-pivot. The two play parallel roles that fall between the responsibilities of a #6 and an #8, floating and moving around.

Sometimes that looks like this:

Often, Rice is the deeper one.

When Ødegaard plays, and the matchup is ripe for it, all three midfielders swap at will.

Pulling Rice down from the frontline was the right call, and one I had been hoping for, as it gives him the chance to do two things: face play (where he’s best), and hold fort below the play on a ball loss, which helps him disrupt counters with more ferocity. Both have been improvements in isolation.

But there are other, second-order effects here.

With two deeper, defensive-minded midfielders and an off-ball #9 (which we’ll cover soon), they need someone to pass it to. This increases the pressure on the receivers ahead of the ball, as Rice is no longer up there, and they are often the free player once a presser steps up.

Look at that pre-season shot again. It’s Calafiori and Ødegaard trying to troublemake in the blindspots.

That looks good on the merits. Both are adept at finding pockets of space, dropping through the middle or wide, doing unexpected things, and then turning on the ball. It gives the opposing CBs a tough choice: follow the player into midfield (and leave the striker in a 1v1), or keep the backline fortified.

I looked through some early games to get a sense of the initial strategy. Against West Ham, Ødegaard left the game early, but the deep players still had the likes of Calafiori, Eze, Saka, and Timber to aim for in the pockets. Chances would start with balls like this:

I had forgotten how much of the playmaking was directed toward Timber in particular. Nowadays, we’re likely to bemoan the lack of dangerous touches more generally. Back then, we wondered why so many dangerous touches were going to Timber, even when his final action could come off messy. (It was because he’s good at receiving them.)

This collapsing central play was often triggered by Zubimendi and a pocket player. In this case, it was Timber again.

One may log on to say, or read, something like this: “a top team doesn’t get its full-backs into those kinds of attacking positions.” Then, one will remember who scored both goals to knock us out of the League Cup.

…or Hakimi’s strike late last year.

These roles are often the free man, picked up late when defenders transition from zonal to man-to-man and back again. Above all, they require an incredible amount of tactical know-how and physical robustness.

Whatever the starting position, those players are often the domino. Timber started losing some verve and sharpness to his actions before officially going out with an injury. When White has played, he has still been pretty effective on the attacking end, just less dynamic positionally, and losing runners defensively. Then, with Ødegaard missing so much time, Calafiori often unavailable, and Saka out, the pockets thinned out, and a team that prefers to play wide anyway gave their double-pivot few options to line-break in the middle.

Here, Zubimendi lets it run centrally, but there’s no one ahead to drive it to.

It’s no surprise that the team looked best once Eze got comfortable, and the midfield rediscovered a semblance of balance.

Oh, and this one.

This balance was lost again when he, too, got injured.

Havertz has a few nice qualities as a #10, but shouldn’t be played there behind three runners, which overburdens him as a passer. After all, you can’t have two off-ball central strikers. We saw those limits in an off-performance against Bournemouth.

Said Arteta:

“I don’t know how many actions there were the other week when we broke the opposition press. In a lot of situations, we have the advantage of so much space to run, and we give the ball away. It turns out to be a counter-attack. These are the moments that shift momentum and shifts the emotional state of the game as well for both teams. That’s what we have to do better as well.”

A lot of his appeal in that role is his ability to swap and change with another forward. But if Gyökeres is there, Havertz can’t play the furthest role of a target-man #9; someone, after all, has to drop and link play. That means, even when Havertz is on the pitch, the high balls often go to Gyökeres.

The other effect of the double-pivot is that progression is increasingly a collective responsibility. For Arsenal, it can sometimes feel like it’s just about prodding around for numerical advantage, regardless of who it is to. In other words, it can feel too agnostic to the individual qualities of the personnel. For instance, if there is a 1v1 on the highest line, we’ll hit it to that player high, regardless of whether they are good aerially or not.

The strategy is egalitarian, humble, and team-based, and lacks the “give me the fucking ball” vibe of a Rodri or Jorginho. Ødegaard is really the only player who tries to take more than his equal share of touches. I’d like to see Rice, Zubimendi, and, especially, Saliba, demand touches of the ball, regardless of the opponent’s setup. I have no problem with Zubimendi picking it up off Raya.

Analytics confirm that Arsenal are rarely challenged at the first line, are great at breaking the last line, and rarely initiate line-breakers through the middle.

The final secondary effect to mention: this double-pivot has changed the role of the left-back to something wider and ahead of the ball, which suits Lewis-Skelly a little less.

➡️ Change #2: Maintaining space to attack in the final third

Both last year and earlier this year, some idiot kept insisting that Arsenal should retain the ball in the middle third more. That idiot was me:

Arsenal would often settle for “easy progression,” advancing the ball upfield on a midspeed carry without creating a genuine advantage. Once the opponent dropped in, the rhythm slowed, and possession turned into patient but predictable circulation: side to side, waiting for a crack that didn’t necessarily come against your Fulhams and Newcastles.

In What’s the Difference, we also probed for answers, looking for the ways Liverpool was more able to generate space than Arsenal. One of the takeaways:

I do believe that Arsenal can recycle back some more, and be a little more patient about advancing to the final third. Teams basically allow those advancements to happen anyway, so Arsenal should wait for more tangible advantages to accumulate to push it forward. This will help maintain the space to shoot when the final action is nigh.

That change has occurred, but both the injuries and the selections by Arteta have made it so that it hasn’t gone exactly as I envisaged.

My February newsletter, Triangles, was many things, but it was largely a critique of certain configurations of this squad.

Perhaps the worst offender was this lineup against Nottingham Forest.

As I wrote then, Lewis-Skelly and others were options:

Arteta was not short on resources for that one. Aside from the XI, Arteta had Saka, Trossard, Merino, Jesus, Eze, and Havertz available at the very least. Still, we could see a series of issues from the lineup drop.

I then listed off my grievances:

• Without Merino or Havertz, there was no aerial outlet.

• The middle had no coherent method for generating threat on its own.

• The left didn’t either, and notably so. Facing Dycheball, that grouping consisted of Rice as a pinning player in the left half-space, Martinelli running at a fairly settled block, and Timber inverting into midfield, onto his wrong foot as a left-back. Who is dribbling to unsettle this block? Who is providing wide help? Who is providing the killer ball?

• With Timber staying low as a pivot, and Rice high on the left, this also constrains Zubimendi’s movement. He’s not going to join the high attack as much because he has less rotational coverage.

• The front-three had two off-ball runners and Madueke.

I shared an interesting study that found the following in 3v3 matches:

[T]eammate combinations explained more variance in team success (20–23 %) than individual players (11–12 %), though substantial residual variance (64–69 %) indicates performance depends on multiple factors beyond these measured effects.

Continuing:

We found that while individual players contributed meaningfully to a team’s goal differential (11–12 %), their impact was consistently outweighed by the effect of teammate composition (20–23 %).

This doesn’t feel like the path to glory to me.

Nor this.

Then, I outlined the qualities I’d like to see in wide trios.

• (Below) An orchestrator: somebody who can dictate play and pick apart the opponent shape with passes.

• (Beside) A dribbler: somebody who can drive at opponents and collapse the shape, starting a cascading effect.

• (Beyond) A runner: somebody who can interpret space and pull it apart: “decompacting” the shape horizontally, diagonally, or vertically.

It’s safe to surmise that Arteta has some different ideas. He’s gone back to lineups with wonky left sides and a Martinelli, Gyökeres, and Madueke frontline enough times to suggest it is a legitimate representation of his preferences, or at least his curiosities. Their collective shortcomings are then compounded when joined by a player like Havertz at #10. (To Arteta’s credit, there weren’t many options against Bournemouth.)

If you are looking to play a trio like Martinelli, Gyökeres, and Madueke, you must ensure there is space for them to exploit. When facing a lower block in the Premier League, they are not the ideal candidates for pinning. Pinning requires a mix of retention, patience, and, in the decisive moment, the ability to execute a refined tight-space action: a snap shot, quick dribble, or slipped pass. As a striker, you’d want to dominate in both reaction time and aerial physicality. As a winger, you need to win a crowded 1v1.

These three are better at running, and runners need space to run. To preserve that space, you’d want to keep possession in the lower and middle thirds, bait the press forward, and hit them over the top. The last City goal showed the outlines of that.

There have been a few issues in practice, however.

1. Martinelli and Gyökeres have not been winning their 1v1s

2. Gyökeres cannot bring down high aerial balls

3. Gyökeres hasn’t been able to link play or beat CBs (i.e., get in front of them) in the channel

4. All three aren’t our best threats at receiving centrally or in the pockets, which puts added responsibility on the #10 to do so alone

5. The lack of an aerial threat puts additional pressure on the buildup

6. When you lose the ball deep, or hit an unsuccessful channel ball, it is harder to spin up your press and counterpress

7. This strategy generates fewer corners than pinning

You can see the outlines of the strategy in a play like this.

But that has been few and far between. It has been more likely to look like this.

Against Bournemouth, those who played in the front three were a combined 1/9 on dribbles.

If you lose the ball on a speedy power-carry like that, all you can do is fall back.

Alternatively, when your rest shape is set up, a high loss can look like this, and turn into compounding opportunities:

Whereas failed long-balls look like this, and turn into extended spells of possession for the opponent, and less aggressive pressing for Arsenal, more waves of pressure, and fewer corners.

In a great James Benge piece, he included this graphic, which shows how Arsenal’s field tilt has been on a downward trend. Almost everything contributes to this, from the change in play-style, to the updates to how the league plays, to the recruits in attack, to the absences of Ødegaard, Saka, and Merino.

➡️ Change #3: A pinning, off-ball striker instead of an all-action #9

In spurts, we’ve seen what the team still looks like with a false-9. But Havertz has only started once as a striker in the league.

Whatever their flaws, the likes of Merino and Havertz are linkers. Gyökeres, on the other hand, has been much more static and central.

There have been three advantages to his play. For one, I think Eze likes playing directly behind a pinning striker.

For another, Gyökeres has been physically robust and available, making 46 appearances across competitions so far.

And finally, his penalties are clinical and satisfying.

In general, Gradient Sports ranks Gyökeres as the 7th-best shooter among Premier League central forwards.

But we can be honest about his limitations elsewhere. In the league, he’s still searching for his first open-play, first-half goal. His ranks for passing, duels, on-ball attacking, and carrying are all near the bottom of the league. Most strangely, in the league, he has been asked to participate in 209 aerial duels that were generated from a teammate’s pass, which is 142 more than the next-highest player at Arsenal (Merino with 67). He has won 30.1% of those.

I like his stubbornness and his shot, and I support him completely. But it feels clear that he should be largely used as a “finisher” moving forward.

👉 What the numbers say

These differences show up in the stats. Whatever our critiques, we must remember that these are the tactics that have brought Arsenal to the top of the Premier League and into the Champions League semi-final.

First, we must ground ourselves in the context of what’s happening league-wide, via this great graphic from Scott Willis.

Then, we can see some of the biggest differences year-over-year for Arsenal.

Some notes on that:

• In a year that the likes of Havertz and Merino have been hurt, we’re competing in ~9 more aerial duels per 90. Opponents have something to do with this.

• There have been more long passes, more clearances, and more losses in the middle third.

• The team is backpassing less and hitting it forward with less accuracy.

As we covered last time, when you hit it long or clear it, you have to fall back into your block, and defending is more tiring than attacking. All minutes are not created equal.

Ball possession markedly affected physical intensity, with the rates of distance covered running (+31%), at high-speed (+30%) and in total (+7%) greater out than in possession.

I would also like us to do quick backpasses from within the block more frequently. That’s how you start to pry open lanes.

Here are a few updated numbers from Gradient to sharpen the picture of how Arsenal are actually playing:

• Arsenal’s pass completion under pressure sits at 80.8%, good for fifth in the league, behind City, Chelsea, Villa, and Liverpool.

• Arsenal play the third-most lateral passes in the league, though there’s daylight between us and City, who manage 30 more per 90.

• Arsenal play backwards at the third-lowest rate in the league (15.7%); City do it more than anyone.

• Arsenal are the league’s likeliest side to receive the ball facing play. Put another way, we’d rather go around the block than through it.

In the build-up, there have been nuances to this period.

👉 Build-up subtleties

We should begin with the fact that this Arsenal side can be quite fluid in build-up when the lineup and connections are right.

Meanwhile, personnel had a lot to do with the loss to City in the League Cup. The double-pivot was front-marked with a super-disciplined front-four wall, cutting off access. Kepa didn’t try to shoot it through, and Eze, Ødegaard, and Merino weren’t there to turn on it. Arsenal then struggled to flatten and stretch that front four out to create lanes, and couldn’t bait the City midfield forward, often opting to hit it long.

There were a few problems with this. For one, Arsenal essentially never successfully isolated high on Khusanov, who is their weakest player aerially. Here, Gyökeres lines up square-on, and Aké wins it.

The more notable thing was the positioning of the respective midfields.

Because Arsenal couldn’t bait the double-pivot forward, Rodri and Silva were never far from the backline. This meant that in a second ball situation, they usually had a 4v2 in the middle of the pitch. Look at it like boxing out in basketball.

This has been a persistent problem since the turn of the year. If you are building up deep, then you have a rushed or unplanned long-ball (i.e., you’re just hitting it to get rid of it), not only is Gyökeres unlikely to win on his own, but you are usually hitting it into less-than-ideal second-ball conditions. After all, our own double-pivot had just been involved in deeper play and won’t be up there to win on the bounce.

It was one of these hoofs that came back at us for the winner from Bournemouth.

These add up.

Hincapié, for his part, is good with second balls.

…and this shows what it looks like to hit it into a better situation.

Moments like that start a one-hand-washes-the-other effect of lower buildup and higher aerial play. If you’re able to generate regular aerial threat, the opponent will be forced to leave a +1 in the backline, which will then give you a +2 when building out from the back, which should generate time for your midfield to work.

If there is a +1 at the back, that usually means the opponent is going to shade you to one side of the pitch, and cut off access to the other side, which can then leave one player free (either the opposite full-back or a roamer centrally). The tricky part is that this requires a lifted pass like this to rip through, which isn’t always advisable.

The Southampton loss was a different problem: a problem of overcommitment. Kai Havertz was expected to play a striker-type role in-possession and Declan Rice’s role out-of-possession, and the other midfielders did as they pleased, looking to push endless numbers forward to overwhelm The Saints. The moves were very almost.

In red, I’ve highlighted the Arsenal midfield on a ball loss that came back the other way for the first goal.

Arsenal then turned down the risk dial considerably in the first away leg to Sporting, making sure not to get pulled out of structure, and circulating patiently.

Zubimendi struggled, and the team (and its finishers) gutted out a late 1-0 winner.

There is one lineup I loved more than any other.

👉 The lineup I loved

Against Everton, we saw the best of many worlds. A tall, linking box threat (Havertz), a roaming central-and-left Eze, a stretched left side that allowed Calafiori to run free, as well as dribblers, outlets, numbers forward, Saka centrally, and a strong rest shape.

This lineup can do transitions, win aerially, and pin teams. The only real risk is the back-post defensive coverage of Madueke and Calafiori on the defensive left.

Even better, this allowed some of those players who gobble up space, like Martinelli and Gyökeres, to impact the game late.

This may feel like a lineup to use at home against a deeper block. But in truth, it could work against anyone.

👉 Quick thoughts on City

I haven’t had time to do an exhaustive dive on City. I did have time to look through some footage and analytics, and I have watched many of their games. For one, I’m not expecting a replay of the League Cup.

I have some basic takes:

• We need to be brave in build-up, and force the ball to players like Saliba instead of allowing it to be pushed left, or hoofing blindly.

• When we go long, we need to isolate the aerial duels we think we can win (Havertz on Khusanov).

• I think we should try to pin and press them, and play a lineup capable of that.

• For another, I think they can be hit on diagonals in transition.

• I’ll touch on lineups next.

Above all, I’d like to see the team successfully pull Rodri and Silva forward, and then play behind them. One of the shortcomings of their squad is the recovery speed of the midfield, so ideally, you want to force them to catch up to the play.

I have seen teams be successful by playing out the back, drawing Rodri and Bernardo forward, and then staying in their cover shadow as a third man.

This is a break that gets their full-backs defending in transition, forces communication by their CBs, and triggers a dead sprint by some of their slower legs.

Valverde was also doing it to good effect. Here he was, following Rodri around.

As play progressed through the left, Rodri could never regain the high ground on him.

Here, Real Madrid pulled Silva forward.

By the time it got into the final third, he was still catching up to the play to regain his position.

And these are the diagonals where they look vulnerable.

…and here’s what I’d go with from a lineup perspective.

Some spare thoughts:

• Play to win. We can do it.

• I feel strongly about starting Havertz. I’d usually want him pressing on the right and Eze on the left, but the dynamics may be better the other way around this time (with Eze releasing Martinelli).

• My assumption is that Ødegaard, Calafiori, and Madueke won’t be fully fit. Different news may change my call there. It would, of course, be lovely to get Eze (LW) and Ødegaard on the pitch at the same time.

• I lean White, but understand those who are concerned about him physically and would opt for Mosquera here. Ideally, White can help pin Doku back. Timber would be ideal.

• There are a few players who I think are capable of A Moment here.

Again, not exactly novel or exhaustive analysis. Just some ideas for tomorrow.

🔥 Final thoughts

The characters in The Big Chill are weighed down by the realities of adult life. But, as children, we didn’t get interested in football as a continuation of the drudgery, though sometimes it can feel that way. We wanted to #touchgrass, to experience something lighter and better, to have something to look forward to every week. In fact, the lightest stretch in the movie (again, if I remember correctly) comes when the group sits down together to watch a game on television, and then by halftime they’re on the lawn to play themselves. For a few minutes, they’re laughing and running and tackling again. The weight they’ve been carrying seems to lighten.

While we get nostalgic about childhood and wax poetic about simpler joys in simpler times, those are simple joys. Simple joys are found at the beach, and there’s nothing wrong with them. But the deepest, most resonant joys are hard-won, and they come from a sense of responsibility, to our community, to our family, to our children, and maybe even to strangers on the internet. (Hi, I love you.)

On a grand scale, the same feeling applies, I suspect, for the players. Whatever resonant joy they find, if they find it this season, will come from the deeper place, from the responsibility they have entrusted in themselves, that the club has entrusted in them, and that we’ve all entrusted in them to carry on our behalf for another month or so.

I don’t know what will happen. So much depends on the lineups and the performance levels; from there, happenstance and variance will do their work. But I agree with Tom Hiddleston, and I agree with Amy Lawrence: what a fortunate place to be in, to feel the weight of such expectations. You can’t do something consequential without consequences.

In March of 1976, Thelonious happened to be listening to a special broadcast by Columbia University’s radio station, WKCR, dedicated to his music. A guest expert began droning on about how Monk created extraordinary music, in spite of “playing the wrong notes on the piano.” Perturbed, Monk dialed the Columbia switchboard and left a message to “tell the guy on the air, ‘The piano ain’t got no wrong notes.’”

We can’t stop watching football, and we can’t agree on what football is. It’s art, but it’s not fully art, because art doesn’t usually have a clock or a scoreboard. When I go to museums or galleries, I’ve noticed the most enduring works typically feature zero (0) video assistant referees. Michelangelo, for all his foresight, didn’t put David in a Hollywood Bets shirt.

But it’s not the tidy competition of a 100m dash either. Part of the appeal of throwing a javelin or a shot put (or whatever) must be the cold truth of the result. In a world of vagaries and human factors and embarrassing little negotiations over lane changes and client approvals and social calendars, there is, alas, something clear and authoritative and knowable in this life. I have to imagine the expected javelin throw distance (xJTD) and the actual javelin throw distance are the same fucking number.

In football, the nerds assure us the scoreboard does, often, lie; at very least, it leaves itself open to interpretation.

Sounds like art to me. Maybe that’s why so many people insist Arsenal are playing the wrong notes.

Occasionally, I put on old games, partially because I don’t sleep much, and partially because I like anchoring myself and getting a sense of what has actually changed (versus what just feels like it has). The other night, I put on an 89/90 Arsenal match. The first two goals were scored from inswinging corners.

Time is a flat circle. In The Simple Art of Murder (1944), the novelist and screenwriter Raymond Chandler wrote a critical essay examining the evolution of the detective story.

The classic detective story has learned nothing and forgotten nothing. It is the story you will find almost any week in the big shiny magazines, handsomely illustrated, and paying due deference to virginal love and the right kind of luxury goods. Perhaps the tempo has become a trifle faster and the dialogue a little more glib. There are more frozen daiquiris and stingers and fewer glasses of crusty old port, more clothes by Vogue and décors by House Beautiful, more chic, but not more truth.

Chandler wrote about the explosion of popular detective fiction, from the sheer volume of bad novels published, to how closely their contours can resemble those of the good ones. The stinkers fail for a million ordinary reasons, but largely because it’s hard to find a writer who can hold the cards of art and science in the same hand.

I suppose the principal dilemma … is that for any approach to perfection it demands a combination of qualities not found in the same mind. The coolheaded constructionist does not also come across with lively characters, sharp dialogue, a sense of pace, and an acute use of observed detail. The grim logician has as much atmosphere as a drawing board. The scientific sleuth has a nice new shiny laboratory, but I’m sorry I can’t remember the face.

The topics, the themes, the narrative thrust can help ‘platform’ a book. But really, it comes down to the skill of the author.

[S]ome very dull books have been written about God, and some very fine ones about how to make a living and stay fairly honest. It is always a matter of who writes the stuff, and what he has in him to write it with.

Like a detective in a Chandler story, Arteta and Arsenal have spent the season building a difficult case in a league that doesn’t give anything away. It has been a decidedly grind-it-out period, punctuated only by an obligatory thrashing of Spurs. With the discourse clanging, every week produces new post-mortems that proclaim the notes are all wrong, all while the team keeps stubbornly persevering. This late in the season, the particulars of the plot matter less than the character study underneath, that lingering question of who Arsenal will be at the end.

👉 I. What good looks like

“I always say that my ideal is to get pole with the minimum effort, and to win the race at the slowest speed possible.”
— Alain Prost, four-time Formula 1 Champion

In the 70th minute at the Amex, that sweeping modern egg laid into the South Downs, the Seagulls gained possession and cleared it over the top. The ever-present Piero Hincapié was playing RCB for Arsenal, because of course he was, and headed it back to Raya so the Spaniard could fetch it with his mitts.

Raya, then, didn’t feel the need to play it back to where they got it. Instead, he threw it over to a left triangle made up of Gabriel, Calafiori, and Rice.

Coaches will tell you to look for overloads, but they don’t always tell you what actually constitutes an overload. Truth is, many situations are interstitial like this. It’s kind of a 3v3, kind of a 4v4, mostly a 3v2, depending on where you draw your box.

Rice then turns and does something that isn’t always advisable in build-up, because the range of outcomes widens. Instead of keeping it grounded, he lifts the ball over the press to switch it to Hincapié.

Hincapié again recycles it back to Raya. But there’s a difference now: Havertz drops from the #9 slot, and he isn’t followed. This means that Brighton are looking to keep numbers on the backline for now. As a byproduct, it will be easier to find a numerical advantage in the build-up phase.

If you count around, this is essentially an 8v5 now. Brighton’s two wide players are hybrid as they reset; nobody has fully followed Havertz, and Kai is feverishly pointing to the soft spot. Once Raya gains a commitment, it’s off to the races.

Because Welbeck has committed to Raya, Zubimendi can fall in behind him and demand that Pascal Groß jump up to meet him. This triggers Havertz to run away, giving Groß the impossible job of marking them both.

Arsenal broke the press, Brighton fell back and regrouped, and possession was sustained from there.

Among other things, this illustrates how backpassing and regrouping are more often signs of composure than of conservativeness. Hoofing, at least hoofing without the requisite talent or numerical advantage, is the easier, more ruffled alternative.

Such sequences reverberate. To prove my point, I will deface the Sistine Chapel (i.e., a Scott Willis graphic) to illustrate where that run of play took place.

Arsenal largely controlled things from that point forward, something they had trouble doing for most of the match. Per James Benge:

Brighton’s final shot of the game, per Opta, came with 68:28 on the clock. After that they had four touches in the box. Arsenal had nine in theirs.

Given our previous traumas, I didn’t give myself full permission to exhale.

But I was close.

So, yes, we know what those sequences should look like. And that situation gave a much-needed reminder of how capable this team can be at building up possessions like that. But we know too well that it hasn’t been the rule of late.

This has been more likely.

Oy.

Wolves, the bottom dwellers of the league, were down. They committed their press forward, a man for every man. By our powers of deduction, we know that means the advantage is more likely to be long, so Raya does what he’s trained to do, and happily sends it over the top to their backline. Especially in today’s game, this is widely viewed as the correct decision.

While the numbers are fine, the delivery isn’t ideal, the jockeying is nonexistent, other players look 10% more tired than they did at the beginning of the season, the ball is lost, possession returns to the opponent, and we’re left wondering why we can’t (fucking) sustain a (fucking) attack (fucking).

This has been a familiar chorus since the beginning of the calendar year.

“It was a moment after another moment after another moment. Even though we scored the second goal, we never got the grip and dominance of the game. That’s the reality,” said Arteta after Wolves. “That’s credit, obviously, to Wolves. I don’t want to underestimate that, but I think we played a big part in that. As I said, it’s very basic things, very simple things, that today we did really wrong and that’s why we had the feeling, without really conceding much, but the moment is always the game open to anything that can happen.”

There was the Spurs drubbing, then, promptly, some familiar bobbles late against Chelsea. In another lovely Benge Factoid, he shared that after Pedro Neto’s red card, Arsenal completed 55 passes … but Chelsea completed 114. It never really felt like we had the man advantage.

After Brighton, Arteta again delivered a summary that felt ctrl+v’d.

“I think that was down to especially what happened certainly in the game as well and the lack of composure that we had in certain moments - probably our heart rate is a little bit too high in certain moments because we are trying to recover as well.”

Mansfield was a thing to itself, then Leverkusen had some throughlines that we’ve seen before: conservatism in possession, some gaunt-looking play, and an opportunistic opponent goal.

Still, the amount of concern over all this can feel disproportionate, given the struggles of our rivals and where Arsenal continue to sit in the league and elsewhere. Across competitions, this same squad, the one causing all the handwringing, is unbeaten in twelve straight.

But it sure feels more complicated and anxious than that, and these issues of play-style and “dominance” keep bubbling up.

So maybe we’ve identified the wound. Stitching it is another matter.

👉 II. What to do when everybody knows everything

“In football, everything is complicated by the presence of the opposite team.”
―Jean-Paul Sartre (via Paul Grech)

Let’s start with the fact that there is an unprecedented amount of money in the high-level game. The market value of the 20 Premier League squads is up 285% since 2015.

With this influx, there is a torrent of downstream effects from the volume of data, analysts, and opposition prep that follows.

This desire (and ability) to research and limit opponent strengths may have even contributed to Thomas Frank’s demise.

Sources at Tottenham’s training ground told ESPN that players became surprised at the volume of work done on nullifying opponents in training, rather than focusing on developing their own style of play.

It’s a fitting portrait of a league that is drowning in talent and physicality and money, yet starved of spontaneity.

This level of opposition prep is nothing new. Sky Sports reporter Tim Thornton recalled a wild press conference in which Marcelo Bielsa went peak Marcelo Bielsa.

“He told us that he and his team watched all of Derby County's 51 games last season, so he broke it down into Derby in this instance, and was calling out into the room ‘pick a game, pick a game’ and they were all numbered on the wall chart that he had … He talked about 360 hours of research, four hours of analysis on each game from last season. This is an incredible level of preparation for each opponent and he made the point that watching training nobody could tell him anything he didn’t already know.”

Similar stories of Mourinho’s work in preparation are easy to remember. While it’s always existed, that level of detail was perhaps more concentrated in a few sickos then, and is more efficient and attainable now. Everybody is in on the act.

The Champions League knockouts still offer a partial antidote, as the game is more likely to tilt back in the direction of raw player quality and brilliance. But the level of maniacal nerdery is everywhere. Leverkusen’s coach Kasper Hjulmand proudly talked about the homework he did on Arsenal, particularly on kickoffs and set pieces, starting the second half in a way that led to the seal-breaker.

Not to be outdone, Arteta hammered home that his team was prepared for the kickoff routine.

“And the other one is us because we knew, we showed them three clips from last weekend in three different ways and we weren’t ready for it and we got caught.”

Balancing that amount of detail with a sense of freedom and proactiveness is a significant problem for the modern manager, especially when there’s little time to train. Eze singled it out when asked for the biggest adjustment when joining Arsenal.

“Probably the level of information around games. There’s a lot of detail that goes into how we approach a game and prepare for a game and it’s been a lot more than what I’m used to.”

(As an aside, the base level of tactical preparedness in the league is what makes some of the Cup setups so compelling to watch. It was unmooring and fun to watch Arteta put out an underdetailed, vibes-based back-three with six attackers against Mansfield.)

Outside of the occasional Mansfield fever dream, things are still methodical. And my feeling is that this analytical bent is exacerbating some of the differences between game-states. In other words, teams prepare more precisely for specific situations, so the game changes more dramatically when those situations shift.

We still see this at the highest levels elsewhere. The well-prepared Leverkusen played great, and once they went ahead, they closed the shutters, logging 0 shots on target and 40% possession.

Hamzah Khalique-Loonat, who is doing some of the best football journalism and analysis anywhere, had this to share:

I asked an intl manager how they scout opponents/what they look for:

“First thing: write down their formations,” they said.

“Ok, in possession and out?” I replied.

“No — how they are set up when level, ahead or behind, and how they change.”

Teams quickly learn the limits of rolling up with the same gameplan regardless of the situation or opponent. Just ask Ange last year. Or Slot this year:

“I’m the same manager as last season. It’s so hard to create tempo in a game if the other team doesn’t want to have tempo,” Slot says. “It’s so hard to create so many chances in the current Premier League where everybody is so physical and has a certain playing style. Everybody is struggling with that and we are struggling with that as well. I would not deny that at all.”

This world makes pragmatists of us all.

While the feel and demands of a 0-0 scoreline in the first half were always different from a late-game at, say, 2-1, the difference seems to be getting more extreme. It’s a different sport based on the clock, so properly attuning your talents and approaches to different phases of the game feels increasingly important.

This, along with other factors (namely, fatigue, intensity, and more refined recruitment by the bottom-half), helps explain the caginess early in games and the explosion of goals late. Looking at the highest-level Champions League games, we still see the influence of the changing Premier League. Any realist must acknowledge that there’s a difference between getting kicked by Caicedo and Joelinton types once or twice when you’re rested and ready, versus facing their kind every week, from all angles, up and down the table. Arteta shared his perspective today.

“For me, Barcelona are the most exciting team in Europe in many moments the way they play, and they face one of the Premier League teams, Newcastle, who are exceptional in their intensity and the high pressure, everything man-to-man, a huge amount of tools, a really good team in transition and we saw a completely different game I’ve never seen Barcelona play and that’s a huge credit to Newcastle, but this is the league that we are playing in.”

But we must also balance that with the triumphs of PSG, Real Madrid, Bayern, and even Galatasaray this week. One can’t watch those performances, ones marked by bravado and worldies and front-footedness, without considering whether the Premier League is lost in its own sauce. At a certain level, everybody has analysts, but everybody does not have Kvara and Barcola and Doué and Dembelé.

After regaling journalists like Thornton with his sprawling tactical masterclass, Bielsa put a bow on his presentation:

“He closed it by saying he was stupid because he didn’t need all this information and then he got up and walked off.”

👉 III. What the numbers say

For our sake, the first thing I wanted to do was try to isolate intent.

When we zoom out, we get some clues.

When Arsenal are ahead, a few things catch my eye:

• Possession drops ten points from ~60% to ~50%. Remember, Arsenal aren’t the only team playing. The opponent has a say in this and will want to get on the ball more.

• Arsenal cross less. This is one of those game-state changes that I fully endorse. If you set up for big, looping crosses from wide areas, there’s a good chance you’re hitting it out of your defensive cocoon. Those balls have a tendency to fall to the top of the box, where they can be countered quickly. Still, it does give you one less route to goal.

• Progressive passes are slightly less accurate, and match tempo is very marginally slower, but nothing too significant. I wouldn’t read too much into that.

• What is significant is the spike in long-balls. In an even game-state, Arsenal hit long-balls 8.15% of the time. While winning, that number jumps to 14.27%. That seems clearly intentional.

• It’s not in that table, but your eyes don’t deceive you lately. In the second halves against Brighton, Chelsea, and Wolves, the average long-ball percentage was over 14%. In the second halves against Chelsea and Brighton, Arsenal could only muster 2.78 and 2.79 passes per possession, respectively.

There is also some other worthwhile context.

• Arsenal are fairly easily the best second-half team in the league, believe it or not. That means the most goals (37), fewest goals against (14), and best goal differential (+23). The second-half goal differential is more than twice as good as the second-best team (Chelsea, at +11). Man City have generally struggled with late goals and goals from subs.

• Arsenal are still the best team in the Premier League after taking a lead. That record so far is 20-3-1, averaging 2.63 points per match.

• I’ve thought some of the away approaches were overly conservative. But on a points-per-match basis, Arsenal are top of the home table, and top of the away table.

• Here’s an interesting one I stumbled across: Arsenal have been behind for 266 total minutes this year in the league. In those minutes, they’ve scored 7 out of 8 times; we saw that capacity again at Leverkusen. Meanwhile, David Raya has only faced two total shots on target this season while trailing in the league, both against Man City. That means that, apart from Man City, no team has logged a single shot on target against Arsenal while leading. It goes to show a) how little Arsenal have trailed this year, and b) how teams close up shop once they score.

And that brings us to our most important point, which is what I always fuckin’ say: yesterday’s price is not today’s price, and the opponent is never static. PPDA (Passes Allowed Per Defensive Action) measures how many passes an opponent can make before a team makes a defensive action in the advanced 60% of the pitch (so, the lower the number, the more aggressively a team presses, at least in theory). Relatedly, Arsenal face the lowest defensive lines in the league, and it’s especially true when the score is level.

When the score is even, Arsenal face a 20.85 PPDA, the most relaxed press in the league by a distance. When Arsenal are winning, that jumps up to 14.66. This, again, shows how differently teams play based on the scoreline. While I was writing this, The Athletic did a great job of covering how that alters the picture.

👉 IV. Raya bypassing the midfield

When I’m looking to understand Arteta’s intention on the pitch, my eyes usually drift to two players: Martin Ødegaard up front (pressing, attacking patterns, etc), and David Raya in back. Because Raya was a long-standing target who can put the ball anywhere, and usually has breaks in play to communicate and confirm before doing so, I have little reason to believe he often defies his manager’s wishes.

(Edit: It should be noted that those are seemingly goal-kicks directly attributed to Raya. I’m still trying to find the ones that get tapped over to him by the CB).

The intention behind this is understandable and not without value. The opponent is more likely to man-mark and come out of their trenches when down, committing players forward in the press, as we saw in the PPDA numbers. This will leave 1v1s along the backline, and the opportunity to win bouncing balls more easily. It also helps avoid costly mistakes in dangerous areas. At worst, it allows you to press on the other side.

The problem is also easy to understand. With Merino and Havertz out, and Gyökeres not sprinting away from his mark in channel runs, Arsenal have no qualitative advantage in these situations. You can generate a 1v1, but your player still has to win the 1v1, and so far, we’re not really doing it. As I wrote last time:

A target man is a warm blanket. You can build up with confidence knowing that if things get hairy, you’ve got a great option ahead. On the flipside, if you know a long-ball is likely to become a turnover, it can cause you to overthink and take unnecessary risks when deep.

Gyökeres’ pre-duel muscling has improved since we dissected it in full, and he had moments early against Leverkusen. But he’s still winning just over a quarter of his aerial duels, and often struggles to make it stick on the ground.

He’s looked best on first-touch, chancey flicks, which aren’t always easy to do from goal-kicks. This often leads to Raya popping it long to other options, who also don’t carry enough of an advantage in their duels. Do we really want our strategy for progression to be high balls to Saka, who doesn’t look at his physical best?

Martinelli and Madueke have more potential to unlock there, but it’s still a work in progress.

And Jesus, previously so scrappy and honey-badger-effective as a target up front, isn’t winning at the moment, either.

We can pull some advanced Gradient stats on this for our two most-played options at striker:

• For Jesus, 68.3% of aerial duels are below expectation.

• For Gyökeres, 61.9% of aerial duels are below expectation. Only 18.6% are above expectation, which is the third-lowest for players with 1000+ minutes.

In sum, Raya is going long when Arsenal are winning, and Arsenal’s current crop of forwards is giving the ball right back.

On that point, it’s hard to imagine a clearer impact than the subs that Arteta threw on against Brighton. Raya hung onto the ball until gaining a commitment forward from Rutter, then ripped it over to Havertz. Havertz didn’t even win the ball, but credibly contested it. Rice brought it down, then the two other subs (Calafiori and Trossard) controlled it, establishing possession for some of the first spells of the game.

Otherwise, it’s difficult if your striker isn’t winning by going up aerially or dropping low into build-up. One or the other has to happen.

If Havertz isn’t ready to start, it sure feels like Arsenal should continue to try and play out from the back with a little more volume. Hoofing it without an advantage aggravates the problem.

👉 V. I’m tired, boss

There is tactical strategy, or at least tactical comfort, in dropping back to a more constrained mid-block after achieving a lead. I am generally not opposed to the idea of scoring, defending a bit like Aston Villa, then subbing runners on (players like Martinelli and Gyökeres), winning the ball, and hitting them on the counter.

But some of the “finishers” have served as “starters” a bit too much this season, which can mess with the late-game dynamics. And there’s a real cost baked into the approach of giving up possession too much.

A few recent moments stick in my mind. They’re almost unexplainable outside of fatigue. The first was this weird, unpressured flip by Timber that landed out of play in the 86th minute to Wolves.

Another was this simple read, with Zubimendi drawing out the press to play wide to Gabriel, who will have a 2v1. For reasons unknown, Zoob hit it right back to Raya.

And the final one was this Rice recovery run, in which he won the spot before yeeting it over the byline for a corner. He was hobbling a bit after.

Players can overcome cardiovascular fatigue through force of will, but hitting a ball with some touch and technique is more rooted in biomechanical fact. In my view, the most heavily impacted player has been Timber, who can still work himself into good positions. His touches often look frayed from there.

It’s not abnormal to feel fatigue at this stage in the season. There is a likely culprit.

👉 VI. Why defending costs more

When you have the ball, the ball moves. When you don’t have the ball, you move.

This shows up in the numbers. A 2023 study in the European Journal of Sport Science found that…

Ball possession markedly affected physical intensity, with the rates of distance covered running (+31%), at high-speed (+30%) and in total (+7%) greater out than in possession.

FIFA’s own World Cup physical data backs it up, with some other little nuggets, like this one: centre-backs and defensive midfielders do 70-85% of their sprinting without the ball.

The CIES Football Observatory found that while out of possession, high-intensity effort is spread almost evenly across the pitch, whereas in possession, it’s heavily concentrated on wingers and forwards.

And in those tight scorelines (the 1-0s), your best players are on the pitch longer. So your starters play more minutes, they perform more high-intensity runs, and they accumulate more fatigue across a season. The Athletic flagged that no Premier League player has run more this season than Zubimendi.

So the tradeoff between defensive solidity and cumulative physical load is real. The question is whether a) we can keep the ball a bit more to lower the toll, b) bigger leads can be forged in the first half, allowing players to rest, c) there can be a bit more rotation in midfield, and d) players can fight through the asks of them.

👉 VII. What Zubimendi is actually doing

There’s another reason for fatigue. Here, I sped up a 10-second sequence of Arsenal build-up play. Watch Zubimendi throughout.

He wasn’t finished; I reached the file-size limit, which probably tells you something on its own. Over the next ten seconds, he kept rotating.

In that short time of looking to pull Chelsea out of shape, he fills zones typically associated with a #6, an LCM, a RCM, a LW, and more. Rodri, he is not.

Two summers ago, I looked into Zubimendi’s numbers and play-style at Real Sociedad. Some of it may feel familiar.

When researching Merino I looked into why Zubimendi and his teammates have lower progression than it feels like they "should." Some context:

• Highest line in LaLiga, very disruptive to opponent build-up, lot of mid-to-high wins and loose balls = no progression needed

• 58.3% of goal-kicks launched (hence Merino aerials) = progression goes wheeee

• Zubi is content in the 'cover shadow' dragging players around, creating space for others. Isn't a #6 who demands every single ball. It's not a knock

• As a result, deep circulation is split between four different defenders, three midfielders, and a GK. The numbers get spread out; not a lot of stat-padding to be done

Then, I investigated further in the big scouting report before he signed.

My only question with this signing is that I worry that a Rice/Zubimendi/Ødegaard midfield is a little light on indefensible killer balls, and am always hugely intrigued by Wharton types for that; Bruno Guimarães always struck me as the right balance, but alas, fuck Bruno Guimarães. Zubimendi doesn’t necessarily offer that on his own — he’s more about enabling others, drawing pressure, timing things right. I think an acquisition of his profile would raise the bar in another signing (say LW) to be able to provide some nasty crosses, or else we’re a little too reliant on Saka again.

Here’s an example of how his focus is to “manipulate the press as much as possible,” and not “receive the ball from deep as much as possible.” After he’s joined in the pivot, he loops around, sees the jump by Barella, and then escapes behind him for a free carry.

Given how often he’s recovering possession or involved in duels, he tends to play off his first touch to keep things flowing or force the issue. That probably explains some of the pass accuracy dings you’ll see in the data, which is a common theme at Real Sociedad, but there is a risk of some bounciness settling in.

Something Arsenal supporters will like, and which became more clear as I watched the “deeper cuts” of his tape, is that he feels more aggressive (and less concerned with pure security) against overmatched sides, taking risks while trying to push forward for goals.

I like the idea of them switching and changing, as it’d offer Rice the opportunity to work a little more on the right: he can run and combine with Saka, freeing up his angles for three things: whipped, inswinging crosses; overlaps and cutbacks; and big cross-body shots like the one he scored against Newcastle. Zubimendi, then, has a lot of untapped potential to push forward on carries, setting up plays.

His passing has been clean, pragmatic, and somewhat measured from a volume standpoint.

Zubimendi is yet to have an 80-pass game for Arsenal, something he’s known to do for Spain.

He has exceeded any expectations defensively. His reading of the game has perfectly scaled to the Premier League, and it’s worth reinforcing how rare it is for somebody to settle in as a possession player so quickly, while winning so much off the ball.

In terms of individual goal-threat, he’s also exceeded expectations.

But he has very much passed in a duo with Rice. As we saw, he’s usually trying to artfully move the block around for others, and we’ve rarely seen him switch into Jorginho Mode to take the bite out of a game. This puts his advanced passing metrics at the baseline.

While he’s unlocked his potential as a goal-scorer, he hasn’t regularly delivered those final balls that turn calm moments into easy goals for the forwards. We’ve only seen flashes.

Arsenal’s likely starting trio (Rice, Zubimendi, Ødegaard) can project like a more physical version of PSG’s flowing setup last year. But it should be said: whereas Zubimendi is yet to top 80 passes for Arsenal, his counterpart at PSG (Vitinha) averages 105.8 passes per 90, and it’s not surprising to see him with 120, 130, 140-pass games. There are league effects there, but also a difference in strategy and, likely, temperament and skillset.

Crucially, his role is another thing that is affected by the timing and score of the game. In second halves, with the ball going long so often, and with Zubimendi not bringing a domineering, give-me-the-fucking-ball streak, he hasn’t been able to pull the possession strings to the same level.

He has shown an urgent vibe, slapping it on his first touch, hoping to keep the ball moving forward, or if nothing else, away from our goal.

It’s all understandable, but you’d like to see him toggle between modes as well. Some of this may come down to an adjustment period. A lot of this comes down to the team’s philosophy of attack.

👉 VIII. The pivot

I’m not sure it was the most strategic grouping, but here are 10 good midfielders in the league.

There are a few takeaways:

• A reminder not to look at the volume of defensive actions as a meaningful metric when thinking about individual quality.

• Rodri is still quite good. This shows the degree to which he is taking responsibility and single-pivoting in an era when most are sharing responsibilities.

• Rice and Zubimendi fare well against the league’s best. What you’ll notice is that other players have some more highs and lows, but Arsenal’s duo fare pretty well on everything.

• Once a weakness, Rice has really cleaned up his aerial work: his 70% win rate is the highest in the sample. It’s another data point for how extreme he can be about removing weaknesses from his game. I used to watch some of his long-throws with a bemused look, and he’s recently been heaving them into the goalmouth.

• I wondered if Zubimendi would be lower for “received passes,” but he comes out fine.

• Zubimendi is second-highest for goals, and is low for shot assists.

At the same time, I spent a couple of hours poring through stats on the Gradient platform and taking notes, hoping to get a better sense of the nuances of Arsenal’s build-up play. It should be said, I wasn’t looking at this dispassionately, given recent attacking form. I wanted to be critical to see what can be improved.

Here are a few data points I found:

• Arsenal currently have the sixth-most passes completed under pressure. The gulf is pretty large, though: Man City average 252.37/90 to Arsenal’s 186.31. My inkling is that Arsenal can usually handle passing through the press, but due to personnel limitations and fatigue, don’t proactively welcome it to the extent that a Rodri or Kroos-style team might.

• Man City are 85.9% when passing under pressure, best in the league. Arsenal are fifth in the league, at 80.7% (after City, Chelsea, Villa, and Liverpool).

• Arsenal are 10th in total line-breaking passes attempted (36.85/90). Arsenal are first in balls breaking the final line of pressure (3.23/90); they are 17th in success rate at breaking lines in the midfield (67.8%). Arsenal, to my eye, have kept the ball on the edge of the final third a bit more this year, and are good at those little slip passes that get players like Saka and Timber past the backline, but are not trying to push it through the opponent’s pivot.

• A look at how Arsenal possessions break down:

  • 30.4% through the left (19th-most). This matches the eye test. More on it later.

  • 35.3% through the middle (5th-most). Probably better than you may think.

  • 34.3% through the right (3rd-most). To be expected.

• Arsenal have the highest percentage of touches facing the goal (49.6%), and the lowest percentage of touches received back to goal (18.8%).

• Man City have 50+ more touches per 90 facing laterally, and carry the highest percentage of backward passes in the league (17.9%). Arsenal have the 17th-most backward passes (15.7%).

• As far as risk goes, Arsenal are fairly low on miscontrols or failed passes. They are fairly high in through-balls and over-the-top balls, and low in switches (1.63/90).

• Most notably, Arsenal have the sixth-fewest “passes below expectation,” but are down in 18th in “passes above expectation,” at 3%. Again, that’s something that may confirm our priors: many good and acceptable passes, not many mistakes, but not many KDB-style world-beaters, either.

👉 IX. On Rice

Arteta views his duo as a duo.

“Yes [I think of them as a pair]. I mean, we always talk about these relationships and chemistry. There is one, and it’s very strong off the field as well. That makes what happens on the field very, very special.”

Rice devised a brilliant career plan that I can’t believe more players haven’t thought of. Why not just be good at everything?

There isn’t much that is unexpected in his passing numbers. Everything has been comfortably above-average as he scales up his volume.

There are a few interesting nuggets as you go through Rice’s year-over-year passing data:

• His passing volume is up 29% year-over-year (41.95 to 54.03). Basically, his deeper role shows up in the numbers, and he’s inherited a good chunk of the touch count of Ødegaard and the previous pivot players. He’s receiving nine more passes per 90.

• His passes to the final third are up 38% (from 5.55 to 7.68), but his accuracy is down a touch (from 82.7% to 80.5%). Little blips in accuracy don’t really matter that much on their own, but it’s worth observing.

• I think he’s proven he can be a great, detailed ‘double-6’ while deep, and is a safe, retentive, and occasionally creative passer in open play. The growth areas are about what happens from there. Most interestingly:

  • His long passes are up 58% (from 2.73 to 4.3), but his success rate on those pings has really fallen off (from 63.8% to 51.5%). This aligns with what we’ve seen elsewhere. The team isn’t particularly successful as a long-ball threat in the league.

  • His through passes have also doubled (0.29 to 0.59), but the accuracy has fallen off (50% to 38.9%).

  • Likewise, his forward passes are up 34% (10.89 to 14.55), but the accuracy on those is also down significantly (83.5% to 76.7%).

• He’s still a near-perfect back-passer and has made very few mistakes.

His work off the ball is looking better than ever.

Here’s Rice and Hincapié competing for the hustle award.

Try as Rice might, Arsenal are still missing something.

👉 X. Eze and the Leverkusen setup

My last big piece, Triangles, was basically a big critique of Arsenal lineups. It covered how any starting eleven is a compilation of small relationships through the pitch, and how many Arsenal selections have been lacking in terms of their complementary characteristics. The TL;DR:

I’ve never wavered from my optimism about our chances for the big stuff this year. Arteta has generally rotated well, and I have few critiques of how the squad has been deployed in the cups or in the Champions League.

But I’ve also been frustrated with some setups and dynamics in the league, even when results eventually go our way. With returns from injury comes a bigger squad. With more options, we’re less resource-constrained, less obligated to do things just because they’re the only things available. It should make it easier to get it right, but it also makes it easier to get it wrong.

There have been times when I felt lost when the Premier League line-ups drop. Too often, they’re comprised of small clusters of players who don’t feel fully complementary, and who narrow the number of coherent arguments we have towards a goal. Gone are the days of simply hoping to see some players and hoping not to see some others. Now, we’re scanning for compatibility.

I spent particular time on lineups like the one against Bournemouth, in which “finishers” were deployed from the get-go, and Arsenal had trouble generating momentum earlier.

Here’s what I’ve been hoping to see.

It’s oversimplified, but in any outside trio, I’d like to see a few adjacent profiles:

• (Below) An orchestrator: somebody who can dictate play and pick apart the opponent shape with passes.

• (Beside) A dribbler: somebody who can drive at opponents and collapse the shape, starting a cascading effect.

• (Beyond) A runner: somebody who can interpret space and pull it apart: “decompacting” the shape horizontally, diagonally, or vertically.

Ideally, a player has the flexibility to play multiple roles based on rotations.

Changes have been slow in coming. The next game saw a lineup against Wolves that had a lot of the characteristics I was worried about. The running trio of Martinelli/Rice/Hincapié on the left had individual moments but not regular dominance; the Martinelli/Gyökeres pairing in the frontline isn’t the right choice against a low-block. Those selection issues compound because you’re unable to push the advantage in the first half, and then you have the wrong profiles available in the second half. Arsenal ultimately gave up a two-goal lead.

The left trio, plus Gyökeres, was trotted out again at Brighton. It was the seventh game that Martinelli and Gyökeres had started together in the Premier League. In those games, they’ve combined for 0 goals and 0 assists, and Gyökeres is yet to have a single shot on target. Through 23 league starts, Gyökeres is still awaiting his first open-play goal in the first half.

These issues have led to some conservative first-half away performances in the league.

• Brighton: 1 shot (1 on target)

• Spurs: 12 shots (3 on target)

• Wolves: 6 shots (3 on target)

• Brentford: 1 shot (0 on target)

• Leeds: 2 shots (2 on target)

• Forest: 7 shots (0 on target)

• Bournemouth: 5 shots (1 on target)

• Everton: 6 shots (1 on target)

To help expand the advantages, here were some of the lineups I was pitching back then.

The familiar groupings were chosen against Leverkusen: Hincapié/Martinelli/Rice on the left, Gyökeres up front. In truth, I would have done something a bit different, but I wasn’t grumpy this time. For one, Martinelli and Gyökeres are better-suited to away days in the Champions League, especially against a side like Leverkusen, who are bound to have some of the ball. For another, that XI does run a lot, and it does have a good performance floor for a first leg. And finally, there were availability issues elsewhere, and the deployment of Eze has improved.

Eze has had a few too many games locked into the right pocket, playing like the lowest-touch version of Ødegaard imaginable. After his Spurs brace, Arteta said, “I start to understand how we’re going to get the best out of him now.” That showed up in his heatmap.

Those closely-knit central pairings can befuddle back-threes. It was so nice to get a couple of quintessential Gyökeres goals. They do hit different.

(Same goes for his penalties.)

Arteta expanded on platforming Eze before Brighton.

“He’s a player that needs a lot of movement around him to activate those spaces, to create a space for himself, and for him in order to be efficient in those spaces. But it’s certainly an option that we have in the team.”

Against Leverkusen, it wasn’t all perfect. These kinds of rotations don’t look fully complementary. All dribblers on the right, all runners on the left.

But what was seen much more often, which I like plenty, is this: Rice driving from deep, Hincapié running up, Martinelli going in and out, and most importantly, Eze available in the pocket on the left, and Zubimendi available in the pocket on the right.

The shape often wound up looking like this, which I have very few complaints about.

These are promising outlines. There’s enough safety in the back for the occasion, plenty of runners, and complementary profiles on paper. As I wrote in Triangles:

In the meantime: a Rice / Zubimendi / Eze midfield is also quite balanced. The issue is how we’ve been using them. We should seriously consider Rice and/or Zubimendi in the right midfield role. Zubimendi can replicate much of Ødegaard’s orchestration. As needed, Rice could play a more aggressive, Henderson-style advanced role, especially on underlaps.

Eze’s action map looked good, aside from receiving in Zone-14 a bit more.

I don’t think this was one of those games with dynamics that were fundamentally off. I didn’t even mind the patience: you don’t want to rush up to a back-five and beat your head against the wall. And it’s an away first leg, of course.

There were good little moments like this.

It’s great to have Eze as a resource on transitions, as he can usually figure something out. Timber has been great at pounding in those diagonals, as he did against Spurs, and it’s great to get Martinelli running like that in the Champions League.

This was another transition that nearly turned into a goal. Timber again hit that diagonal ball through, which Eze dummied, Gyökeres laid off, and Martinelli hit off the crossbar.

Through the middle, there were some snappy moments.

Why didn’t it turn into more?

To my eyes, there were some speed issues and occasionally funky groupings of players. But mostly, it was about individual performances, timing, and rhythm. It was just a little tired and sloppy.

Martinelli stopped his run here.

The timing for moves like this was just a bit off.

And Saka was simply losing his battles.

About a month ago, I did a deep-dive on Saka’s play and found that concerns over his performances were massively overblown. He was playing great, unlucky with assist variance, but sublime at everything else, except for missing shots. I do think he’s #actually struggled more lately. It’s OK. He’s looked good in the #10, and it’s why we have Madueke.

For Timber, aside from those great diagonals, we saw some tired whacks.

…and a missed header here.

People are understandably wary about our best chances being shipped to our full-backs, but look above. It’s not luck or tactical happenstance on Timber’s part: he gets into those spots, and that’s a huge part of the battle. It’s a bit like imagining Fabio Vieira being able to whip balls in from the spots that Ødegaard gets free in. That’s well and good, but he wasn’t getting into those spots.

So it was all that … and Leverkusen played quite well. I’m not sure I have more analysis.

👉 XI. Calling all passers

I really liked Tim Stillman’s description of the differences between Ødegaard and Eze:

If Ødegaard works an opponent by weaving, ducking and diving and landing lots of jabs to the abdomen, then Eze is your uppercut guy.

Jabs and uppercuts; that seems apt. And maybe the problem is that you need a few jabs to open up the opportunity for an uppercut. Those jabs, particularly from deep, have been few and far between.

The simple reality is that Arsenal are a deeper squad than last year, have more legs to throw on late, more ability to rest players and compete across fronts, and better defenders overall. I also think that much of the tactical intention has improved, especially given the realities of the day. But Arsenal have taken a step back as a passing team.

It’s not terribly difficult to deduce why:

• First, the full-backs. For fitness or preference, Hincapié has gotten the call over Calafiori and Lewis-Skelly. Timber, previously a massive-volume buildup CB at Ajax, was a bit behind White as a combining attacking force, and has gotten more harried as the season has gone on. He’s still getting in good positions and deleting wingers from existence.

• Per Gradient, Ødegaard graded out as the best passer in the league among attacking midfielders and wingers last season. This was widely seen as a down year for him.

• This year, the top line-breakers have been Ødegaard (4.6/90), Zubimendi (4.32/90), Saliba (4.06/90), Rice (3.57/90), and Mosquera (2.95/90). There aren’t a lot of line-breakers coming from full-backs or players who have filled in for Ødegaard. Saliba, we should remember, has also missed some annoyingly big games.

• Ødegaard is also statistically (and actually) our best player at breaking multiple lines, and breaking the midfield lines. Our second-best player at breaking the midfield line is Merino, also out.

• A few positions have taken a step back. Gradient’s league-wide rankings are rough viewing, take them as you will, and tell an important story. Gabriel Magalhães, Martinelli, Madueke, and Gyökeres are all near the very bottom of their respective positions for passing and passing under pressure right now.

• There’s also a receiver issue. Arsenal have the second-lowest cutback success rate in the league (26.3%), only above Burnley. I think many of those passes look plenty “on.”

I don’t see a massive quality difference in the passing analytics of Zubimendi and last year’s sixes. They got on the ball more, but weren’t at their best, especially when under pressure. I do miss those 80-touch Jorginho games, though, and think Zubimendi is capable of tilting more in that direction.

My synopsis here is simple: more technicians need to hit the pitch.

👉 XII. Hincapié and the left

Hincapié has been brute-forcing his way into the left-back spot.

Here was the one-pager from before his league debut.

It can be exhausting just to watch him. Here are twenty sped-up seconds in the life of the Ecuadorian.

I’ve noticed him doing a lot of good things with his body orientation.

Here, he wins the spot, lets it run across, forces out the wide CB, and meets the ball on time. Against a back-three, this has the added bonus of emptying out the middle. A single control touch would kill this move.

Here, he doesn’t swivel his head or give away that he’s looking to release Rice. He fully sells the inside carry to gain the commitment.

And here are the good hips. When you get fired a ball like this, it’s hard to swivel enough and get touch on the release. But he’d already scanned Trossard’s position and just opened up.

He’s not a dribbler, but he’s also more interesting than a Hustle Chad. He’s got a good relationship with the ball, especially with first touches.

Still, a lot of that impact is still in progress, and I’m not sure his best use is sticking him to the inside of Martinelli. If you’re wondering why he’s been chosen over the likes of Lewis-Skelly, there were clues in this Arteta conference.

“Before, when you used to do a game plan and you just invert a full-back and bring an extra player in midfield, or a false-nine, opponents are finished, big overload, 4v3 inside, 2v1 inside, time on the ball, so dominant, 70-80% of possession, the other opponent, two counter-attacks, set-pieces, the game is done. Now, teams are adapting.”

It sounds like he believes a fluid midfield double-pivot is the best response to the rise of man-marking, which “un-inverts” the full-back and keeps them wider, which also gives them longer transitional runs. That tilts his equation toward Hincapié.

I’m not 100% sold on that in all situations, and yes, I think Lewis-Skelly’s skillset is well-suited for the floating “double-6” thing that Rice and Zubimendi are doing. It may just take some time to get there.

(Health permitting, my choice is still generally Calafiori, especially in small-space games).

🔥 Final thoughts

For all the consternation, not least of which by me, Arsenal have so far been the best in the world at navigating this moment. It is the highest-floor squad in an age where the strength of that floor is constantly battered and tested. An eighth-place Everton knocks next. It’s hard to point to any Arsenal player right now and say they’re at their best, but the results are still flowing, mostly.

There has been stubborn perseverance, but also, a compounding loop. When certain first-half groupings start, the lead isn’t as wide as it could otherwise be, which leads to starters falling back and defending more often, which leads to fatigue, which leads to worse passing, which leads to long-balls, and repeat. Ødegaard and Havertz will help if able, but the solution is always the Injured Guy, and things can be improved elsewhere. Zubimendi can rack up more touches, and more technicians can hit the pitch, more often, and earlier, and in more complementary triangles. Arsenal need to reclaim our passing bona fides.

I don’t know about you, but this was my most stressful moment in a minute.

We’re lucky to live with this tension. We’re lucky to be sprawled behind the couch in positions our bodies have never quite experienced before, too anxious to watch. We’re lucky to be fighting on this many fronts at once. And oh, I almost forgot: we’re lucky to have Max Dowman.

It’s art, it’s science, it’s pseudoscience, it’s competition, it’s disorder, and it’s luck. Arteta was cuttingly straightforward when asked about play style.

“It’s playing the best football that you possibly can and the game demands to give you the best possibility to win the game and that's it.”

In his final analysis of those detective tales, Raymond Chandler said their proper backdrop is “not a fragrant world, but the one we live in.”

Then, he offered his conclusion on what makes such a story compelling.

All this still is not quite enough. In everything that can be called art there is a quality of redemption. It may be pure tragedy, if it is high tragedy, and it may be pity and irony, and it may be the raucous laughter of the strong man. But down these mean streets a man must go who is not himself mean, who is neither tarnished nor afraid.

Since we last spoke, only a month ago now, we’ve looked toothless and goalless in a draw against Nottingham Forest, played Inter off the pitch at the San Siro, lost in angering fashion to a clinical United, indulged in beautiful 3-2 slopart against Kairat Almaty, destroyed Leeds 4–0, vanquished Chelsea from the cup in cagey-then-riotous fashion, popped on our handmasks for a second-half Gyökeres brace, felt deflated again at Brentford, and regained some joy in the FA Cup. There have been own goals turned olímpicos, injuries and returns from them, lineup consternations, familiar concerns in attack, explosive celebrations, yellow cards for sarcastic laughter, annoying City comebacks, and meandering rants from humanity’s most smooth-brained pundits. There have been understandable recriminations and their worthwhile rejoinder: Arsenal still sit atop all the tables.

In a previous, pre-cloud age, this file would have been titled newsletter-final-v2-final-FINAL-v4-final.doc. I’ve done a lot of writing and rewriting and deleting and rewriting during all this. In keeping with the times, we are presented with a new swipeable story or stimulus every minute. Oh, Kai is back? And gone again? Why isn’t Merino on the bench? Oh, we signed an 18-year-old Scot? And a Stoke defender? And Dowman? Why is all the data on fbref suddenly gone? Why is Liam Delap playing right wing against us? Is that Saka in the #10? Did Arteta just do a second-half goalkeeper change?

“There it is again,” Bo Burnham sings, “that funny feeling.”

A couple of weeks back, Jon Mackenzie linked to a study titled “Quantifying individual and teammate effects on small-sided football performance through repeated-game observations.” In it, researchers had 31 players compete in 3v3 games, then repeatedly reconfigured the teams to estimate the effects of different groupings. The findings were interesting.

[T]eammate combinations explained more variance in team success (20–23 %) than individual players (11–12 %), though substantial residual variance (64–69 %) indicates performance depends on multiple factors beyond these measured effects.

Continuing:

We found that while individual players contributed meaningfully to a team’s goal differential (11–12 %), their impact was consistently outweighed by the effect of teammate composition (20–23 %).

In short, teams won more because of how players fit together than due to individual quality.

The other intriguing thing was how much having a mixture of different player profiles may matter. The researchers created four archetypes: goal scorer, team catalyst, defensive specialist, and role player. For men, at least, teams composed of three distinct player archetypes outperformed less diverse teams.

The scope of this project was fairly limited and should be kept in proportion.

Still, it makes sense, no?

This effect is most pronounced in international play, when coaches have less time to create marginal gains through endless drilling and tactical detail. There are shotgun marriages throughout the pitch, and as a result, underlying player characteristics matter more and are harder to override. If your eleven doesn’t have a fundamental complementarity, it will be sussed out.

(I fully expected “complementarity” to get a red squiggly line under it. But apparently it’s an actual word.)

In the last Euros, we can remember the ramrod efficiency of the Spanish eleven.

This was just a group of players standing where they should be, and doing things they were good at. End-of.

Contrast that to the England side, who would often slap redundant profiles together. I remember Kane, Bellingham, Rice, Mainoo, Foden, Saka, and a wrong-footed Trippier all requesting the ball to their feet, vying for the same 20 square yards of grass.

Many stars can still generally succeed in such an environment, but few can reach their full potential without the surrounding conditions helping make it so. As in life, we are heavily impacted by what is around us. And lineup composition is a tricky, mysterious trade.

I’ve written for Tactx in the past, and the bright minds over there often talk about the “Three Connections,” which they observed at La Masia. Once you start seeing the game this way, it’s very difficult to unsee it.

For England, Thomas Tuchel’s tactics so far have been fairly easy to digest, which, for me, means they are easy to admire. I prefer international tactics that are devastatingly simple. The quicker you establish a firm tactical identity, the sooner you can hammer home the details that make it hum.

In the above study, we saw how conjoining profiles reign supreme in 3v3 games; when you look at a wide-angle view of the full 11v11 pitch, you start to really think about how those 3v3 player groupings enhance or detract from one another. In these early days, you can see those roaming groupings, with Kane empowered to join any at will.

The players form the rest shape, and also ensure that there is a “below,” “beside,” and “beyond” option at every turn.

When we look at passmaps, it’s hard not to think of something like a network or a power grid. In fact, they’re often called passing networks. Those subjects include many relevant sub-topics that I very much understand, so long as you don’t ask me too many follow-up questions.

There’s the idea of load balancing: spreading demand across multiple routes so no single corridor gets overloaded.

Load balancing can optimize response time and avoid unevenly overloading some compute nodes while other compute nodes are left idle.

There’s the idea of rerouting, of parallel paths, of contingencies. When one path fails, it puts extra stress on the others. It can look like this:

I’ve also written in the past about “triple modular redundancy,” a concept in which “three systems perform a process, and that result is processed by a majority-voting system to produce a single output. If any one of the three systems fails, the other two systems can correct and mask the fault.”

For football, this can be extended to depth: you want three possible options at every position. But it can also be extended to balance: you want three routes to every goal: left, middle, and right. And it can also be seen in these recurring triangles throughout the pitch.

There are players who make these kinds of connections easier. Through his return and all-too-soon absence, we were reminded clearly that Kai Havertz is one.

“His main quality is that he makes you (as a team-mate) better somehow,” said Arteta. “The way he moves, communicates with you, gives you information and moves around. He’s a very difficult player to mark, and especially he opens up a lot of spaces for the rest of the players. So I think they’re going to be very happy to have him around.”

The idea that some players quietly lift the level of everyone around them is easy to agree with in theory, but can sometimes be hard to fully appreciate in the moment. Football happens too fast, and we’re too busy looking at the ball, arguing about shots and subs and referees to notice the small mercies that thoughtful players provide us with.

In the sixty-fifth minute against Sunderland, Havertz rolled over from his spot as an “RCM” to join this wide left triangle.

He combined and delivered the ball out to Trossard; in this case, he was the “below” man in the triangle. But this situation also met the conditions of an overlap: because Trossard was single-marked by a stationary man, and the other defender would have to arrive at pace, and force a timely handoff, Havertz could provide the overlap and test their connection.

Sunderland responded well, as they do, so Havertz again looped around and provided an option to Rice, who had joined as the “below” player. Havertz was “beyond” as the ball was recycled.

The ball made it out to the other side, which I won’t screenshot for now, but Havertz essentially played the role of a box #9, looking for crosses.

As the ball made its way back to Trossard, Havertz again showed an option: floating back to provide an underlap. Trossard, as he was wont to do, tried another so-called risky pass, but Sunderland were able to jump it.

Sunderland won possession and tried to play out, but Rice and Havertz smothered them. The ball again went to Trossard. Winning on the counterpress is nice for a lot of reasons, but one undercelebrated facet is that the offside line is garbled, and there are more opportunities to sneak around. Havertz ghosted a run in-behind, Trossard hit him this time, and the rest was history.

Here it is, in full flow, with Gyökeres hitting the exact kind of messy striker’s goal one would hope for from the long-awaited striker signing.

It shows just how much relentlessness is required for a goal against a block like this.

Last time, I listed a litany of small grievances about our Arsenal, from the lack of an aerial presence, to pressing follies, to some issues connecting the middle of the pitch. I laid out some of my preferred tactical solutions, but ended it with this.

If you want my simpler opinion on it, it’s this: Save us, Kai Havertz.

Havertz made a brief, teasing cameo against Portsmouth before being put back on the shelf in the name of caution. After that, he made four appearances, playing all of 191 minutes (the equivalent of 2.15 90s), and he tabbed 4 G+A. Looking back, that may be underselling it. It’s almost bewildering how many little plays he made.

He switched his bodyshape against Kairat, ghosting up into the right wing to dissolve into the blindspot, before White hit him over the top.

It led to a footrace with Gyökeres, with the Swede, ahem, screening defenders, before Havertz cut back onto his left and hit it home.

Against Sunderland, Raya freed him up for a fifty-yard carry that almost resulted in a curling Goal of the Month contender.

The Kairat destruction started early, with Eze showing deep (he was defending in the left-8), line-breaking up to Havertz, who delivered it up to Gyökeres.

The pass was similar to the one he later delivered against Sunderland for the narrowly-offside Jesus penalty shout.

What’s the point, besides Kai propaganda? (And isn’t that reason enough?)

Arsenal have a few “below” players, a few “beside” players, and a few “beyond” players. Some, like Calafiori, Saka, and White, can capably float between roles. But none can combine the ability to interplay with the ability to truly gallop in-behind to the level of Havertz.

This “beyond” ability led him to the badge-kissing finale of the Carabao semi-final.

With Merino likely out for the bulk of the year, it’s clear how big a physical vacancy this leaves.

We can hope for his speedy return, while acknowledging the ability to improve elsewhere.

After the United loss, Arteta hoped to see the players perform with more lightness.

“We took a moment to bring the temperature down, to pause, to reflect and ask two questions. One is how do we feel, and how do I feel myself, and then how we want to live the next four months. It was so encouraging and beautiful, because what came out of there is very simple. We have earned the right to be in a great position in four competitions, and in the next four months we’re going to live and play with enjoyment, with a lot of courage and with the conviction that we’re going to win it.”

The return of Kai briefly helped. But looking more broadly, many of these connections we speak of, those little nodes that transmit attacking energy through a passmap, have sputtered this season. Things simply don’t feel figured out in attack, and we were cagey again at Brentford. This subject is where much of the variability in the remaining games will lie.

I’ve never wavered from my optimism about our chances for the big stuff this year. Arteta has generally rotated well, and I have few critiques of how the squad has been deployed in the cups or in the Champions League.

But I’ve also been frustrated with some setups and dynamics in the league, even when results eventually go our way. With returns from injury comes a bigger squad. With more options, we’re less resource-constrained, less obligated to do things just because they’re the only things available. It should make it easier to get it right, but it also makes it easier to get it wrong.

There have been times when I felt lost when the Premier League line-ups drop. Too often, they’re comprised of small clusters of players who don’t feel fully complementary, and who narrow the number of coherent arguments we have towards a goal. Gone are the days of simply hoping to see some players and hoping not to see some others. Now, we’re scanning for compatibility.

Nail it, and we can win anything. Are we on our way to doing that?

👉 Faulty wiring

My criticisms were sharpest against Forest.

Arteta was not short on resources for that one. Aside from the XI, Arteta had Saka, Trossard, Merino, Jesus, Eze, and Havertz available at the very least. Still, we could see a series of issues from the lineup drop.

• The right triangle had White, Ødegaard and Madueke. Pretty solid, though Ødegaard would be without his partner-in-crime, while facing the exact kind of matchup he struggles with the most: being a pocket player behind a physical DM like Elliot Anderson.

• Without Merino or Havertz, there was no aerial outlet.

• The middle had no coherent method for generating threat on its own.

• The left didn’t either, and notably so. Facing Dycheball, that grouping consisted of Rice as a pinning player in the left half-space, Martinelli running at a fairly settled block, and Timber inverting into midfield, onto his wrong foot as a left-back. Who is dribbling to unsettle this block? Who is providing wide help? Who is providing the killer ball?

• With Timber staying low as a pivot, and Rice high on the left, this also constrains Zubimendi’s movement. He’s not going to join the high attack as much because he has less rotational coverage.

• The front-three had two off-ball runners and Madueke.

The intent to “overload to isolate” is clear. But what if that doesn’t work out? Where are the reroutes, the parallel paths, the redundancies? What would an attack on the left look like?

In practice, it often looked like this.

It was a little too similar to Bournemouth, when the “finishers” started the game, but couldn’t get going on the left.

Against Wolves, who sit deepest in the league and have won one (1) time, Arteta’s frontline included two in-behind threats (Gyökeres and Martinelli), as well as Rice in the pinning LCM role.

All of these may be explainable in isolation, but collectively, they paint a pattern.

The Forest game ended 0-0, the second scoreless draw in a row. A contrarian may point to some missed chances by Martinelli and Gyökeres, but one was on a second-phase after a set piece, and another was after a Forest error. Neither, in other words, was due to structural superiorities in open play, and the first half finished with 0.35 open play xG and 0 shots on target. More importantly: it just felt like players were fighting to overcome the setup.

The floor has remained higher than that of other teams. But there is no doubt that the attacking ceiling can be raised.

👉 What would “better” look like?

It’s oversimplified, but in any outside trio, I’d like to see a few adjacent profiles:

• (Below) An orchestrator: somebody who can dictate play and pick apart the opponent shape with passes.

• (Beside) A dribbler: somebody who can drive at opponents and collapse the shape, starting a cascading effect.

• (Beyond) A runner: somebody who can interpret space and pull it apart: “decompacting” the shape horizontally, diagonally, or vertically.

Ideally, a player has the flexibility to play multiple roles based on rotations.

The first example that comes to mind of such a trio is Trent Alexander-Arnold, Mo Salah, and Jordan Henderson. Henderson would always try to peel a yard or two of space for his colleagues, and then the group would pounce on the advantage with their unique talents. If they were all orchestrators, or all dribblers, or all runners, it wouldn’t have worked nearly as well.

Another shining example? For a year or two, one could argue that the most productive attacking trio in the world was Saka, White, and Ødegaard. At their physical best, Ødegaard could orchestrate from deeper positions (and is actually a great decoy runner on the inside), White could orchestrate or run, and Saka could generate paranoia in all three points of the triangle, but especially on the dribble. It left defenders with nothing but bad choices. So much was happening in blindspots.