Inside the match.
Compare chance creation and revisit the detail behind each result.
INTELLIGENCE
What decided this match?
Checking the result, chance creation and response to goals…
Evidence-ranked explanation. The review questions test possible tactical causes; they are not claims established by the event data.
Leyton Orient 2 – 3 Wycombe Wanderers
Full-time score
| Leyton Orient | Metric | Wycombe Wanderers |
|---|---|---|
| 0.88 | Expected goals | 1.16 |
| 0.88 | Non-penalty xG | 1.16 |
| 1.41 | Expected assists | 0.90 |
| 13.00 | Shots | 16.00 |
| 4.00 | Shots on target | 7.00 |
| 27.00 | Touches in opposition box | 23.00 |
Connections, context and recurring routes
Start with the relationships. Then inspect one recurring sequence at a time.
Building contextual passing evidence…
Who connects the team?
Routes worth examining
At most 12 directed links are drawn. Select a connection to see its individual passes.
One pattern. Multiple occasions.
Three consecutive linked passes, grouped by phase and route through a 3 × 3 grid. A repeat must occur in at least two distinct control spells within the selected team, period and time band.
How to read this analysis
Feed continuity: for a completed pass, the immediately following item in the original liveData.event list identifies the receiver when it is an on-ball action by a different named teammate in the same period. We do not skip items or impose distance or time thresholds. Opponent actions, administrative items, stoppages, restarts and period boundaries do not create teammate links. The original cached feed supplies the order; no network is inferred from database row order when that feed is unavailable.
Coordinates: pass origins use the supplied x/y; destinations use qualifiers 140/141. Receiver identity does not depend on proximity to the next action. Missing coordinates affect plotting only, not the connection. Network nodes average the supplied passing and receiving locations. This is an event-continuity rule, not a separate receiver field.
Context: transitions cover the first 10 seconds of an inferred open-play regain. Restart phases cover the first 20 seconds, separating goal kicks from other restarts. Other actions after at least 10 seconds of observed control are circulation. Early uncertain origins stay separate. These temporal labels do not claim to detect defensive pressure or tactical intent.
Repeated routes: non-overlapping three-pass chains stay within one inferred control spell and one phase. Grouping uses the first origin and three destinations, with defensive/middle/attacking thirds labelled D/M/A and lateral coordinate bands 1/2/3. Different players can execute the same route. Repeated zone routes are descriptive matches, not learned tactical classifications. Coordinate interpretation assumes x increases towards the opponent goal; verify against video.
One team and period are shown at a time. Narrow the time band to inspect personnel changes; this view does not assume every player was simultaneously on the pitch. Node positions use all eligible connections in the selected context before the minimum-link and player-focus filters. Clip links use exact stored anchor events and may show only part of the sequence. Network filters are independent of the transition and event labs below.
The rhythm of the contest
How quickly does possession change hands—and what do teams create when they win it?
Reconstructing changes of control…
Where the game changed hands
Confirmed open-play switches per five-minute clock band. Select a band to inspect its regains.
How long did control last?
Completed regain-to-loss intervals only. Stoppages and incomplete spells are excluded.
Follow the transition
| Regain | Team / first controller | Time held / ending | First shot | Early xG | Review |
|---|
Definitions, confidence and interpretation
Orient inferred control v1. A pass attempt, take-on, shot, recovery, keeper claim or pickup establishes control. An opponent's next controlled action confirms a turnover. Tackles, interceptions, saves, clearances and duels alone do not establish control. Event clocks have one-second resolution; timestamps order ties.
Outs, fouls (conservatively including advantage), corners, offsides, goals and explicit restart qualifiers break sequences. Period boundaries and long gaps censor observations. Rapid-shot rates exclude short incomplete spells unless a shot is already observed. The median and middle 50% describe completed regain-to-loss intervals only; zero-second intervals are excluded. Quick regains credit the team winning the ball back within five seconds after its loss.
Rapid attacks are a counterattacking proxy. Opponent organisation requires video or tracking. Switch frequency describes possession volatility; it does not measure physical workload. Five-minute bands use the match clock, including stoppages, and separate each period; final bands may be partial. No league ranking is implied.
Coordinate gain uses successive controlled-action origins on the supplied 0–100 grid, assuming x increases toward the opponent goal. Validate direction with video. It is not ball velocity or player running speed. Missing shot xG stays unavailable. Clip links match existing clip anchor events; they may cover only part of a sequence.
Definition reference: event definitions. This is an in-house reconstruction, not an official possession feed.
Explore the event data
Loading event data…
Sample and time filters update the metrics below. Pitch layer and outcome controls filter the map, event table and CSV only.
Where the actions happen
For passing structure and repeated phases, use the pass network above ↑. Individual passes here support event auditing.
How the chances developed
Cumulative recorded shot xG · selected events
What this selection tells us
Selected events
| Fixture | Player | Period / time | Action | Outcome | xG | Location |
|---|