Where, exactly, can Arsenal hurt Paris Saint-Germain in Budapest? The question is worth asking now, while the semi-final footage is still warm, because what Luis Enrique’s side showed against Bayern Munich over 180 minutes was not the performance of a team without vulnerabilities. PSG are in the final on merit, but they arrive with cracks. Mikel Arteta’s staff will have watched the Allianz Arena second leg and seen the architecture of those cracks in detail.
Start with what PSG do well, because it sets the geometry of the problem Arsenal must solve. Enrique’s base shape, in possession, is a 4-3-3 that becomes a 3-2-5 when the left-back, Nuno Mendes, pushes high and Achraf Hakimi tucks inside to form a back three alongside Marquinhos and the left centre-back, Lucas Beraldo or Presnel Kimpembe. The single pivot, Vitinha, sits behind the two number eights, Warren Zaïre-Emery and Fabián Ruiz, who occupy the half-spaces. Ousmane Dembélé plays the false nine, drifting right. Bradley Barcola holds the left wing. The sixth attacker is whoever Enrique picks on the right, usually Kang-in Lee or Gonçalo Ramos.
In the first leg at the Parc des Princes, this structure overwhelmed Bayern for 70 minutes. PSG’s xG in that opening period, per the StatsBomb model, was 1.87; Bayern’s was 0.22. The overloads came through the left side. Barcola, Nuno Mendes, and the left-eight, usually Ruiz, formed triangles that isolated Joshua Kimmich, Bayern’s right-back, and forced the centre-backs to choose between stepping to the ball and covering the central channel.
But the second leg, at the Allianz Arena on Wednesday, told a different story. Bayern manager Vincent Kompany made one adjustment that changed the entire defensive picture: he dropped his defensive line by eight metres and asked his two eights, Aleksandar Pavlović and Leon Goretzka, to stay narrow, within nine metres of each other, rather than pressing the half-spaces. The effect was to compress the zone where Vitinha likes to receive and turn. In the first leg, Vitinha completed 11 progressive passes. In the second leg, the figure was four. The supply line into Dembélé’s dropping runs was choked.
Here is where the diagram-paragraph technique matters, because one sequence in the 58th minute showed the vulnerability Arsenal should study. Bayern cleared a PSG corner. The ball fell to Zaïre-Emery on the right side of the centre circle, facing his own goal. He turned and played a short pass to Vitinha, who was already under pressure from Pavlović. Vitinha laid it back to Marquinhos, who was now the deepest outfield player. Marquinhos looked up, saw Barcola’s run on the left, and tried a 40-yard diagonal. The pass was underhit. Jamal Musiala intercepted, drove forward, and played in Harry Kane. Kane’s shot was saved by Gianluigi Donnarumma, but the sequence was instructive: PSG’s build-up, when the pivot is compressed, routes through Marquinhos, and Marquinhos does not have the passing range to beat a structured press from deep.
Arteta’s Arsenal press differently from Bayern. Where Kompany asked his eights to stay narrow and let the centre-backs have the ball, Arteta’s trigger is typically the goalkeeper’s first pass into the back line. In the 2024-25 Champions League group stage, Arsenal’s PPDA (passes per defensive action) in the pressing phase was 8.3, the second-lowest in the competition behind Manchester City. The front three press in a stagger: the centre-forward, likely Kai Havertz in Budapest, takes the ball-side centre-back; the ball-side winger presses the full-back; the far-side winger tucks in to cover the pivot. The ten, Martin Ødegaard, covers the space between the pivot and the ball-side eight.
The geometry of this press suits the problem PSG showed in Munich. If Havertz forces Donnarumma to play short to Marquinhos, and Saka or Leandro Trossard (depending on who starts on the right) presses Nuno Mendes or Hakimi on the first touch, the passing lane into Vitinha is covered by Ødegaard’s position. The question then becomes: can Marquinhos play through the press, or does he go long?
Against Bayern, the answer was long. And long balls from Marquinhos, under pressure, had a completion rate of 38 per cent in the second leg, according to StatsBomb’s event data. PSG’s aerial duel win rate in the defensive third was 41 per cent, seventh-worst among Champions League semi-finalists over the last five seasons. Havertz, at 6‘4” and Gabriel Magalhães at set pieces, give Arsenal a target the Bayern front line lacked.
The second vulnerability is more structural, and it lives on PSG’s right side. Enrique’s system asks Hakimi to tuck inside when the team has the ball, forming that back three. But Hakimi’s defensive recovery, when possession is lost, is slower than Nuno Mendes’s on the opposite flank. In the 34th minute of the second leg, Bayern won the ball in midfield. Serge Gnabry, who had held width on Bayern’s left all evening, received a switch pass with Hakimi 14 metres infield. Gnabry drove to the byline, cut back, and found Kane. The shot was blocked, but the pattern repeated twice more before half-time. Hakimi’s average recovery time on defensive transitions in the second leg was 4.2 seconds; the Champions League median for full-backs this season is 3.1.
Arsenal’s left side is where this becomes actionable. Gabriel Martinelli, if fit, or Leandro Trossard starting on the left, can exploit Hakimi’s positional indiscipline in the same way Gnabry did. The key is the trigger. Arteta’s team does not switch the ball as a default; they probe the right side through Saka and Ødegaard, draw the press, and then release the left winger on the counter-transition. The pass that unlocks it is Ødegaard’s diagonal from the right half-space to the left channel, a ball he has completed at a rate of 72 per cent this season in the Premier League, per StatsBomb.
The third and final crack is Donnarumma’s distribution. PSG’s goalkeeper attempted 14 long passes in the second leg. Six found a teammate. The other eight went to Bayern players, three of them in PSG’s defensive third. In a match where Arsenal’s press is functioning, those turnovers are the difference between a half-chance and a clear opportunity. Arteta will know this. Arsenal’s pressing structure is designed to force the goalkeeper long and then win the second ball in the opponent’s half. Their second-ball recovery rate in the Champions League this season is 54 per cent, the highest in the competition.
Enrique, for his part, will have solutions. His half-time adjustments in Munich, when he moved Dembélé wider and asked Zaïre-Emery to drop deeper to create a double pivot, were effective in possession if not in outcome. The question for Budapest is whether Enrique adjusts to Arsenal’s press before Arteta’s team finds the first goal. PSG’s record when conceding first in the Champions League this season is one win in five.
The constraint Arteta must solve before the final is different: what happens when PSG beat the press? If Vitinha receives in space, his progressive carry distance is 6.3 metres per possession, the highest among midfielders in the knockout rounds. If he turns, Dembélé’s movement between the lines creates the kind of chaos that Arsenal’s centre-back pairing, Gabriel and William Saliba, have rarely faced in the Premier League. Saliba’s one-on-one duel win rate is 74 per cent, but Dembélé does not attack in straight lines. He drifts, he feints, he changes the angle of attack three times before committing. The geometry of the final will be decided by whether Arsenal’s press arrives before Vitinha’s first touch, or after it.
