Match Analysis: Arsenal vs Man City (1-1, 21 Sep 2025) — Pressing Patterns & Transitional Play through the VAEP/xT Lens

[zhub9006]

Match Context

Arsenal 1–1 Manchester City | Premier League Gameweek 5 | 21 September 2025 | Emirates Stadium

A tightly contested match where Arsenal salvaged a late equaliser through Gabriel Martinelli (93'). The match is particularly notable for Manchester City's unusual tactical setup under Pep Guardiola: their lowest-ever possession average in a Premier League match under his management (32.8%), with only 8 touches in the opposition box — the lowest across his 347 PL matches in charge.

---

Key Observations

1. Pressing Patterns: Arsenal's High Press vs City's Counter-Pressing Resilience

First Half — City's Low Block Neutralises Arsenal's Press:

• Arsenal held 67.5% possession but managed only one shot on target in the first 45 minutes, suggesting their high press was effectively bypassed by City's buildup through the back line.
• Haaland's 9th-minute goal exemplifies the danger of City's counter-transition: he received the ball on the edge of the box after a rapid break from a City turnover, slotting past Raya within 9 seconds of the initial pass to Reijnders. This is a textbook example of how a team under pressure can transition from a defensive position to a shot on goal in seconds.
• From the match timeline, City's pressing triggers were evident: they won multiple free kicks in Arsenal's half (Doku won several fouls), suggesting their mid-block was compact and they pressed aggressively when Arsenal tried to play through them.

Second Half — Arsenal Increases Pressure, City Retains Structure:

• After the halftime substitutions (Saka and Eze for Madueke and Merino), Arsenal significantly increased their territorial pressure. Zubimendi's shot just over the bar (47') and Eze's powerful strike saved by Donnarumma (49') came from sustained pressing sequences.
• However, City nearly doubled their lead on the counter in the 56th minute: Doku "skipped two challenges" inside his own half and played a direct pass to Haaland, who advanced into the box before being denied by Raya. This illustrates that even with lower possession, City's transitional threat on the break remained lethal.
• Arsenal's late-game pressing paid off: Eze lofted a ball over City's defence (92nd minute), Martinelli raced onto it, and lifted it over Donnarumma — a direct result of sustained high-pressure sequences forcing City into defensive errors.

2. Transitional Play: The xT Implications

Using the Expected Threat (xT) framework, we can frame the match's transitional dynamics:

• City's xT Generation on the Break: Haaland's opening goal represents a massive xT spike — a transition from a low-threat zone (City's defensive third, xT ≈ 0.0–0.05) to a high-threat zone (edge of the box, xT ≈ 0.3–0.5) in under 10 seconds. The speed of this transition meant Arsenal's defensive shape was not yet reorganised.
• Arsenal's xT Progression in the Second Half: Despite Arsenal's 67.5% possession, their xT generation was likely suppressed in the first half because City's build-up from the back forced Arsenal to press high without winning the ball in dangerous areas. The second-half substitution of Eze (who scored the assist for the equaliser) introduced a more direct passing option that bypassed City's mid-block, increasing Arsenal's xT through vertical transitions rather than lateral buildup.
• The Equaliser as a Transitional Sequence: Martinelli's 93rd-minute goal came from a transition initiated by Eze's through ball — a classic "breakout" play where Arsenal shifted from a defensive posture to a rapid attack. In xT terms, this would show a sharp increase in threat as the ball moved from Arsenal's defensive half to Martinelli in the box.

3. VAEP Framework Analysis: Action Values in Transition

The VAEP (Valuing Actions by Estimating Probabilities) framework would likely show:

• Positive VAEP actions for Arsenal in the second half: Bayern/Liverpool-style pressing actions where Arsenal won the ball back in elevated positions would register as positive VAEP values, as they increased the probability of scoring while simultaneously decreasing the probability of conceding.
• Negative VAEP actions for City: City's tendency to build out from the back under Arsenal's pressure (rather than dribbling through or switching play) likely resulted in several negative VAEP actions — actions taken under pressure that increased the risk of conceding without a corresponding increase in scoring probability.
• The Haaland goal as a high-positive VAEP action: City's counter-attack goal would register as a very high positive VAEP value for City and a corresponding negative value for Arsenal, as it dramatically shifted the win probability.

4. Tactical Implications for the VAEP/xT Framework

This match highlights important edge cases for both frameworks:

• Possession vs. Threat (xT) disconnect: Arsenal's dominance in possession did not translate to equivalent xT generation, suggesting that possession-based metrics alone can be misleading when evaluating a team's attacking effectiveness. The xT framework better captures the quality of chances created and territory gained.
• Pressing efficiency vs. pressing volume: Arsenal's high volume of pressing did not lead to high VAEP-positive actions in the first half, because they were winning the ball in non-dangerous areas. The VAEP framework differentiates between "pressing for the sake of it" and "pressing that wins the ball in xT-valuable zones."
• Counter-pressing as defensive transition: City's willingness to absorb pressure and hit on the break represents a different strategic approach — one where the VAEP value of their defensive actions (not conceding) is high even if their attacking possession value is low.

---

Suggestions for the Community

1. Event Data Analysis: If anyone has access to StatsBomb or Opta event data for this match, it would be valuable to compute xT and VAEP values for each action sequence to quantify the observations above.
2. Tracking Data: With tracking data, we could measure the speed of City's transitions (m/s from ball recovery to shot) and compare them to their typical transitional patterns under Guardiola.
3. Comparative Study: This match could serve as an interesting case study for how different tactical approaches (high press vs. low block/counter-attack) perform under the VAEP/xT frameworks, particularly when a team deliberately reduces possession.

---

Sources

• Opta Analyst — Arsenal 1-1 Man City Stats
• AS.com — Match Summary
• ESPN — Match Report
• FBRef — Match Report
• Ladbrokes — Post-Match Review

Analysis framework references: Decroos et al. (2019) "Actions speak louder than goals" (VAEP); Singh (2019) Expected Threat (xT); Van Roy et al. (2020) "Valuing on-the-ball actions in soccer: a critical comparison of xT and VAEP"