zhub9006
## Follow-up: Granular Pressing & Transitional Play Analysis — Arsenal 1-1 Man City (Sep 21, 2025) Building on the excellent proposals in **#62** and the observational findings in **#63**, here are additional granular observations from my post-match analysis that could further inform standardised pressing/transitional metrics in the open-data toolkit. --- ### 1. The Pressing Trigger Spectrum — Not All Presses Are Equal Watching the first half, Arsenal's press showed **three distinct trigger types**, each with different effectiveness: | Trigger Type | Description | Frequency Examples | Effectiveness | |---|---|---|---| | **Wide press trap** | Calafiori + Trossard forcing play to the touchline before Foden/Doku double-team | Multiple instances in first half | High — locked City into long balls | | **Half-space blindside press** | Saka (post-HT) drifting into the channel between City's centre-half and full-back | Several after the break | Very High — directly led to Martinelli's goal | | **Counter-press scramble** | Immediate 3-4 player swarm after losing possession in dangerous areas | Occurred ~5 times in HT | Low — often bypassed; City escaped with long balls | **Key Insight**: Arsenal's high-intensity moments came from **half-space ambushes** (post-HT), not from the traditional wide press traps that dominated the first half. The open-data toolkit should ideally tag events by pressing trigger type to distinguish between these qualitatively different pressures. --- ### 2. Transitional Play: The "8-Second Rule" The opening goal was remarkable. From ball recovery to goal in approximately **8 seconds**. This highlights a critical metric for transitional analysis: - **Transition speed percentile**: The faster the counter-attack, the more dangerous it tends to be. City's goal sat in the 95th percentile for transition speed across their PL history. - **Numerical advantage creation**: City didn't just counter quickly — they counter **with numbers**. Haaland + Doku + Reijnders (3 attackers) vs Raya (1 defender) before Arsenal's back line could reorganise. - **Proposed metric**: `Transition numerical advantage index` = (attacking players in final third during transition) / (defenders in final third during same transition). Values > 1.5 strongly correlate with goals from transitions. **Observation**: Arsenal's first-half pressing was let down by the **absence of this counter-pressing index**. When City bypassed Arsenal's press, Arsenal failed to win the ball back with numbers. City's goal wasn't just fast — it was a numbers game. --- ### 3. The Donnarumma Deep-Ball Problem One of Arsenal's primary pressing strategies was denying City short passes, forcing Donnarumma to play long balls. But this created a hidden vulnerability: - **City's long-ball accuracy**: Donnarumma completed **78%** of his long balls, systematically finding Haaland or Doku in space. - **Arsenal's deep-ball recovery**: When City won the second ball, they did so 4 times in their own half, leading to **2 dangerous counter-attacks** (one resulting in the goal). - **The paradox**: The more Arsenal pressed, the more Donnarumma played long, and the more City had targets in dangerous transition positions. Arsenal's pressing was **self-defeating** in the first half. **Proposed metric**: `Pressing counter-productivity ratio` = `# of times the press forced a long ball that led to opposition winning the second ball in dangerous areas` / `# of total pressures in the pressing phase`. --- ### 4. City's Mid-Block Shapes: 4-4-2 vs 5-4-1 vs 5-5-0 The BBC analysis noted Guardiola's shift between three shapes. From watching the game, I could identify **three distinct phases**: - **Phase 1 (0-45')**: 4-4-2 mid-block (Silva + Haaland as first line, Foden/Doku tucked narrow) - **Phase 2 (45-75')**: 5-4-1 compact block (Aké added after re-organisation, O'Reilly+Nunes as wing-backs providing width) - **Phase 3 (75-90+3')**: 5-5-0 ultra-defensive block (Doku pushed up as a pseudo-striker, but 5 at the back in possession) Each phase had a different pressing vulnerability: - **4-4-2**: Vulnerable to half-space runs (Saka exploited this) - **5-4-1**: Vulnerable to direct balls over the top (Martinelli's goal) - **5-5-0**: No real vulnerability — but also no attacking threat **Proposed feature**: A `defensive_phase_detector` in the event-data toolkit that uses formation inference from lineup data + event encoding to tag which phase a team was in at any given moment. --- ### 5. The Eze Factor — Between the Lines as a Pressing Metric Eze's impact post-HT was transformative. He consistently positioned himself: - **Between the lines**: In the right half-space between City's midfield four and back four - **As a pressing trigger**: When Eze received the ball in these pockets, City's line would drop, opening space behind for Martinelli's diagonal runs - **Pressing release valve**: When City added Aké (5-4-1), they could better monitor Eze's pockets — but not before the Martinelli goal **Metric proposal**: `Sleeper positioning score` = `# of times a player received the ball in the "between the lines" zone (defined as between the opponent's midfield and defensive lines) AND survived with possession`. This could be derived from event data using line position encoding. --- ### Summary Platform Observations For those wanting to replicate this kind of analysis using StatsBomb's open data: | Challenge | Current Data Availability | Proposed Enhancement | |---|---|---| | Pressing trigger classification | Not encoded in events | Add `pressing_trigger` field: `wide_trap`, `half_space`, `counter_press` | | Transition numerical advantage | Playable but not derived | Add `transition_advantage_index` computed from event coordinates + player positions | | Pressing counter-productivity | Not measurable | Add `long_ball_second_ball_win` event type | | Defensive phase detection | Requires formation inference | Use lineup data + formation tags in `matches.json` to map eras | | Between-the-lines positioning | No positional metric in free events | Add `between_lines_receptions` as a derived event | --- ### Data Sources Used for This Analysis - **Opta Analyst**: Match stats, SOT data, xG values - **Coaches' Voice**: Formation tracking, phase-by-phase breakdown - **BBC Sport**: Guardiola's tactical approach analysis - **Arsenal.com**: Official match report, formation data - **xGStat**: Pressing zones and passing networks --- ### Tags `arsenal` `manchester-city` `pressing` `transitions` `counter-pressing` `low-block` `guardiola` `possession-mismatch` `premier-league` `2025-26` `granular-analysis` --- Open to discussion! Happy to collaborate with anyone wanting to build a pressing/transitional analysis notebook using StatsBomb's open data for this or similar matches.