Short answer: Expected goals (xG) measures the quality of a scoring chance as the probability that it becomes a goal. A shot worth 0.30 xG is scored about 30% of the time on average. Adding up a team's xG shows how many goals its chances should have produced – a much more stable measure of performance than goals themselves, and one of the most useful tools in football betting.
How xG is calculated
Models look at hundreds of thousands of historical shots and estimate the chance of scoring from similar situations. The main inputs:
- Distance and angle to goal
- Body part – headers convert less often than shots with the foot
- Type of assist – a cutback is more dangerous than a cross
- Situation – open play, set piece, counter-attack, penalty
- Defensive pressure and goalkeeper position (in more advanced models)
A penalty is usually worth around 0.75–0.8 xG; a header from a corner around 0.05–0.10; a tap-in from two metres can be above 0.80.
Key xG metrics
| Metric | Meaning | Why it matters |
|---|---|---|
| xG | Quality of chances created | Attacking strength |
| xGA | Quality of chances conceded | Defensive strength |
| xGD | xG minus xGA | Best single measure of team quality |
| npxG | xG without penalties | Removes a volatile, low-repeatability source |
Why xG predicts better than goals
Goals are rare and noisy. Over a handful of matches, finishing luck and goalkeeping can make a team look far better or worse than it is. Chance quality is much more stable, so xG over the last 10+ matches predicts future goals better than past goals do.
Example: regression to the mean
After 15 matches a team has scored 25 goals from 17 xG – eight goals above its chances. Unless it has elite finishers, the most likely future is closer to the xG rate. The market often still prices the team on its goals, which can create value on the opposite side. The same works in reverse for a team that creates plenty but keeps missing.
How to use xG in betting
- Rate teams by xGD, not by the league table.
- Estimate expected goals for a match by combining one team's attack (xG) with the other's defence (xGA), adjusted for home advantage.
- Turn expected goals into probabilities for over/under and both teams to score with a Poisson model.
- Compare with the odds and bet only when your probability is higher than the implied probability.
Limits of xG
- Game state. A team leading 2–0 sits deep and concedes shots; xG can flatter the trailing side.
- Finishing quality is real for a few elite players – but much less common than streaks suggest.
- Models differ. Providers use different inputs; compare like with like.
- Small samples. Fewer than 6–8 matches is mostly noise.
- Context is missing. Injuries, a new coach or fixture congestion are not in last month's xG.
FAQ
What does xG mean in football?
Expected goals – the probability that a shot becomes a goal, summed over all shots to show how many goals a team's chances were worth.
What is a good xG per match?
Around 1.3–1.5 per match is typical. Top teams in major leagues often average 2.0 or more, while weak attacks sit below 1.0.
Is xG useful for betting?
Yes. It is one of the best publicly available indicators of team strength and helps spot teams whose results are better or worse than their performances.
What is the difference between xG and xGA?
xG measures the chances a team creates; xGA measures the chances it concedes.