Match analysis
Burnley 2–0 Leeds: why xG and win probability tell different stories
Burnley won despite recording less xG. Follow the model from a 25.3% pre-match chance to 58.3% after the opener, and learn what each metric measures.
By xG Lab ·
The result and the chances answer different questions
Burnley beat Leeds United 2–0 at Turf Moor on 18 October 2025. In xG Lab’s dataset, Burnley recorded 0.43 xG and Leeds 2.66. Leeds generated the larger estimated scoring opportunity total; Burnley scored the goals. Both statements can be true without making expected goals useless.
Expected goals assigns a scoring probability to each recorded shot and adds those probabilities together. It describes the chances taken. A win-probability model instead asks how likely each final result is, given the score, time remaining and the team-strength estimates it uses. Once a goal is scored, it is part of the current state of the match.
Watch Burnley’s model probability change after each goal
Our replay starts with Burnley at 25.3% to win, a draw at 29.7% and Leeds at 45.0%. After Burnley’s opener in the 18th minute, the home-win probability rises to 58.3%. The second goal in the 68th minute takes it to 96.6%.
| Minute | Match state | Burnley win | Draw | Leeds win |
|---|---|---|---|---|
| 0′ | Kick-off | 25.3% | 29.7% | 45.0% |
| 17′ | Before the opener | 24.4% | 33.7% | 41.9% |
| 18′ | Burnley lead 1–0 | 58.3% | 24.8% | 16.9% |
| 67′ | Before the second goal | 77.0% | 19.4% | 3.6% |
| 68′ | Burnley lead 2–0 | 96.6% | 3.0% | 0.3% |
| 90′ | Full-time endpoint | 100.0% | 0.0% | 0.0% |
These are xG Lab model estimates, not observed frequencies. The replay uses a simplified 90-minute clock; its full-time endpoint is settled at 100%, rather than a prediction made before the referee’s whistle. Rounded rows may not sum to exactly 100%.
Why Leeds’ higher final xG does not undo Burnley’s lead
A missed chance can raise a team’s accumulated xG, but it does not put a goal on the scoreboard. Once Burnley led, Leeds needed actual goals to draw or win. With less time left to score them, their win probability fell even though their final xG total was higher.
Our in-match model estimates future goals from pre-match team strengths, remaining time, score and recorded red cards. It does not update team strength from the live shot-by-shot xG total. That modelling choice explains why the xG timeline and probability replay can move differently. It also limits what the replay can say about changes in attacking pressure.
The useful reading is therefore two-part: the xG view helps you review chance creation; the probability view shows how the score and clock changed the path to the final result. Neither one is a complete account of the match on its own.
Sources and next steps
The xG values and probability checkpoints above are a fixed snapshot of xG Lab’s match data, captured on 12 October 2026. xG is supplied through our licensed TheStatsAPI feed; win probabilities are calculated by xG Lab. xG figures vary between providers. The Premier League’s match report confirms the result and the two goal minutes used here.
Open the full match replay, xG timeline and shot map, or read how the probability model works. To compare players and teams yourself, start with our guide to xG Lab’s football data tools.