Learn what expected goals (xG) measures, how to add shot values and why expected goals are not a promised final score.
What is expected goals (xG)?
Expected goals (xG) represents the estimated probability that a given shot becomes a goal, calculated from comparable past shots. A chance scored at 0.20 xG is one that a particular model estimates would convert around 20 times across 100 similar shots. Shot distance, angle and shot type are common inputs, while more advanced systems include defender and goalkeeper positioning. See Hudl / StatsBomb — Expected Goals definition.
Worked example: adding shot quality
| Fictional shot | xG value | Interpretation |
|---|---|---|
| Close-range chance | 0.40 | 40% modelled conversion likelihood |
| Mid-range chance | 0.20 | 20% |
| Header | 0.12 | 12% |
| Long-range shot | 0.05 | 5% |
| Difficult angle | 0.03 | 3% |
| Total for five shots | 0.80 xG | Sum of chance probabilities |
Illustration only: 0.40 + 0.20 + 0.12 + 0.05 + 0.03 = 0.80 xG. It does not mean the team will score exactly 0.8 goals or any specific number of goals in a match.
Compare goals, xG and xG difference
| Statistic | Example (fictional) | Meaning |
|---|---|---|
| Actual goals | 2 | Observed goals scored |
| xG for (xGF) | 0.80 | Model-estimated chance quality created |
| xG against (xGA) | 1.10 | Model-estimated quality conceded |
| xG difference | −0.30 | 0.80 minus 1.10 |
| Result | 2–1 | Actual score can exceed xG |
MLS describes xGF, xGA and xG difference as standard football analytics concepts: Major League Soccer — Soccer Analytics 101.
Why two xG providers might disagree
There is no single universal xG formula. Models may use different historical samples and variables. A provider including goalkeeper position and defender pressure may give a different probability than a simpler distance-and-angle model for the same shot. Compare model definitions before treating two providers’ xG totals as identical.
xG is not a pre-match prediction
After a match, recorded xG describes the quality of chances actually created during play. Pre-match expected-goals forecasts are separate predictions of what chance creation might occur. A team can outperform xG in a single match, and individual results vary. See the statistics guide for observed-versus-derived measures.
Evidence, scope and further reading
Sources below support the concepts explained here; teaching tables and calculations are fictional and do not represent current team data or official Forebet forecasts. Recheck external source methods before making real-world decisions.