How does Mikel Arteta’s 4-3-3 deal with Diego Simeone’s transition-baiting 5-3-2 when the away goal is gone and the tie is level on the road back from Madrid? On Tuesday at the Emirates, the answer has to be different from the one Arteta gave at the Metropolitano. The first leg finished 1-1T1 - UEFA.com, Viktor Gyökeres converting from the spot before the interval and Julián Álvarez levelling from the same spot after itT1 - UEFA.com. Arsenal stay unbeaten in UEFA club competition this seasonT1 - UEFA.com; Atlético have lost one of their last twelve away in EuropeT1 - UEFA.com. The geometric problem the second leg poses is narrower than the scoreline suggests, and the channel where it will be solved is the half-space.
Begin with what Madrid actually showed. Simeone set up in a 5-3-2 out of possession that became a 3-5-2 when Atlético had the ball and a 5-3-2 again the instant they lost it. Reinildo Mandava and Nahuel Molina, the wing-backs, played at the touchline when Atlético were settled, and tucked into a back five at the first sign of a turnover. The two strikers, Álvarez and Antoine Griezmann, did not press Arsenal’s centre-backs. They held the half-way line and waited. The midfield three of Koke, Pablo Barrios and Rodrigo De Paul stepped onto Arsenal’s pivot only when the ball entered a specific zone: ten yards either side of the centre-circle, on the floor, with the receiver facing his own goal. The trigger was geometric, not territorial.
That is the bait. Atlético want Arsenal to pass into midfield in a posture that forces a turn. The turn is the cue. The press from De Paul and Barrios, two-on-one onto whichever of Declan Rice or Mikel Merino has received, is timed to arrive as the receiver swivels. The ball comes loose. Álvarez and Griezmann, who have not pressed and are therefore not winded, run the channel between Arsenal’s centre-backs and the recovering full-back. This is the pattern that produced Atlético’s best two open-play sequences at the Metropolitano, in the 28th and the 51st minutes. The second of those won the foul that Álvarez converted from twelve yards.
Arteta’s first-leg answer was to take the bait fewer times. Arsenal’s progression through midfield in the second half, after the equaliser, dropped to short combinations in wide areas: Bukayo Saka and Jurriën Timber on the right, Gabriel Martinelli and Riccardo Calafiori on the left. The full-backs held the touchline; the wingers came narrow into the half-space; Martin Ødegaard, the ten, dropped between the lines whenever Atlético’s midfield three stepped. Arsenal’s xG in the second half was 0.71T1 - StatsBomb model, via UEFA broadcast graphic, against Atlético’s 0.42T1 - StatsBomb model, via UEFA broadcast graphic. The geometry was Arsenal’s. They did not convert.
The Emirates problem is different in two respects. First, Arsenal need to score, not draw. A goalless ninety sends the tie to extra time and then penalties, where Jan Oblak’s record makes the maths uncomfortable. Second, Atlético no longer need to invite the press to set the trap; they can sit deeper, defend the box, and counter from a more compressed shape. Simeone’s 5-3-2 at home, with possession, was a 3-5-2. Simeone’s 5-3-2 away, without possession, is likely to be a 5-4-1, with one of Álvarez or Griezmann dropping to make a midfield four and the other holding alone on the half-way line as the counter outlet.
The half-space is where the geometry matters. Atlético’s 5-4-1 leaves two channels exposed by construction: the corridors between the wide centre-back and the wing-back on each side, ten to fifteen metres wide, sitting on the edge of the box. Saka, when he is at his most dangerous, plays exactly there. The first leg saw him take an average position 9 metres infield from the right touchline (SofascoreT2 - Sofascore positional average); he will need to play narrower at the Emirates, perhaps 6 metres infield, which puts him between the right wing-back and the right-side of the back-three. The same logic on the other flank gives Martinelli the channel between Reinildo and the left-sided centre-back.
The diagram-paragraph is the 38th minute of the first leg. Saka received from Timber on the right, eight metres infield. Reinildo, the left wing-back for Atlético when defending, stepped to him. Behind Reinildo, the left-sided centre-back held his line. Saka cut inside Reinildo onto his left foot. The pass he wanted was a low cross to Gyökeres, who had peeled off the right-sided centre-back into the channel just vacated. The pass did not arrive; the centre-back recovered; the chance died. That sequence, repeated with the cross arriving, is Arsenal’s clearest path on Tuesday. The geometry exists. The execution did not.
There is the question of the press. Arteta cannot press Atlético’s back three high without giving Álvarez and Griezmann the running space they want; he cannot drop the line to fifteen yards inside his own half without surrendering the territory needed to threaten Oblak. The middle solution, the one Arsenal trialled in the second half at the Metropolitano, is a 4-4-2 mid-block triggered by the back-pass to the Atlético goalkeeper. The line-height in that phase was 38 metres from Arsenal’s own goal (approx., from the broadcast tactical camera; not a tracked figure). It is high enough to compress Atlético’s build-up and low enough that a clipped ball over the top has 30 metres to travel before reaching Álvarez. The bet is on the percentage of those balls Gabriel and William Saliba defend. Arsenal’s centre-back pair won 79% of aerial duels in the Premier League this seasonT2 - Opta, via Premier League official site. Atlético’s two strikers won 47% of theirs in La LigaT3 - Opta seasonal average, via Marca summary. The maths favour Arteta. The variance is the problem.
The Ødegaard role is the third axis. Simeone’s mid-block has a soft point at the seam between Koke and the right-sided centre-back, the place a ten can find on a half-turn. At the Metropolitano, Ødegaard touched the ball there fourteen times and turned with it fiveT2 - Sofascore touch map. Three of those turns ended in shots; none were on target. The Emirates instruction, inferred from the pattern rather than a presser quote, will be to play the ten higher and let one of Rice or Merino take the dropping role. That gives Ødegaard a starting position closer to the Atlético line and a shorter distance to travel for the ball that arrives in the half-space.
Arteta has not, in the post-first-leg presser, named his eleven for TuesdayT1 - UEFA.com matchday wire. The selection question that matters geometrically is whether Calafiori starts at left-back. Calafiori’s progressive carry numbers (4.7 per ninety in the Premier League this seasonT2 - FBref, season aggregate) are higher than Myles Lewis-Skelly’s, and the half-space where Martinelli will operate needs an underlapping full-back to occupy Reinildo. Calafiori does that. Lewis-Skelly stays wider.
The forward-looking constraint is this. If Arteta plays the same shape as the first leg, with Saka and Martinelli at the touchline and Ødegaard between the lines, Atlético will defend the same 5-4-1 they prepared for and the tie will go to extra time. If Arteta narrows the wingers, drops Calafiori in to underlap, and lets Rice take the dropping pivot role to free Ødegaard higher, the half-space chance Saka missed in the 38th minute at the Metropolitano will arrive again, in a different shirt, against a back-line that will be twenty minutes more tired than the one in Madrid. The geometry is solvable. Whether Arsenal solve it for the first time in twenty years, after last reaching the final in 2006T1 - UEFA.com, is the only thing that will be remembered on Wednesday morning. Budapest, on 30 May, waitsT1 - UEFA.com.
