What is xG (Expected Goals)?

The simple definition

Expected goals (xG) assigns a probability to every shot in a football match. A shot from six yards out, straight at an open goal, might have an xG of 0.85, meaning a player in that position would score roughly 85% of the time. A long-range effort from 35 yards might carry an xG of 0.02.

At the end of a match, each team has a total xG figure: the sum of all their shot probabilities. A team with 2.3 xG created far more dangerous chances than a team with 0.7 xG, even if the scoreline does not reflect that.

How xG is calculated

Data companies like StatsBomb, Opta, and Understat build xG models using thousands of historical shots. Each shot is characterised by dozens of factors, including:

• Distance from goal: closer shots have higher xG
• Angle: central positions produce better chances than tight angles
• Body part: headers have lower conversion rates than shots with the dominant foot
• Assist type: goals following crosses, through balls, or set pieces each have different base rates
• Defensive pressure: shooting under pressure reduces xG
• Goalkeeper position: advanced models account for keeper positioning

The result is a consistent, objective score for every shot that removes the subjectivity of traditional analysis.

xG vs actual goals: what does the gap tell you?

When a team consistently scores more goals than their xG suggests, one of two things is happening: they have an exceptional finisher, or they are getting lucky. Over time, most teams regress toward their xG baseline.

This regression is exploitable. If a team is on a winning run but their underlying xG figures are mediocre, the bookmaker odds may still be underestimating the probability of a result reversal.

Conversely, a team losing games but generating strong xG each week is often a value bet at inflated odds. They are unlucky in the short term, not bad.

xGA: expected goals against

xGA is the defensive mirror of xG: the total expected goals of chances conceded. A team with a low xGA is genuinely restricting the quality of chances their opponents create, rather than just benefiting from poor finishing.

The most complete picture of a team comes from looking at both:

• High xG + low xGA: genuinely strong team
• High xG + high xGA: open, attacking team (good for over/under markets)
• Low xG + low xGA: defensive side, backs up under bets and draw markets
• Low xG + high xGA: struggling team likely to concede

Using xG for football betting

Professional quantitative bettors use xG as a core input in their probability models. The basic process:

1. Collect each team's rolling xG for and against (typically last 6–10 games)
2. Build an expected goals model for the upcoming match
3. Convert the model output to match probabilities using a Poisson distribution
4. Compare those probabilities to the bookmaker's implied probability
5. Bet when the model finds a meaningful discrepancy

KiqIQ's Poisson calculator uses this exact framework. Enter each team's average goals scored and conceded per game and it outputs match probabilities and fair odds you can compare directly to bookmaker prices.

Limitations of xG

xG is a powerful tool but it has real limitations you should understand before relying on it:

• It ignores player quality. Elite finishers like Harry Kane consistently outscore their xG because they are genuinely better than average. Pure xG models undervalue great strikers.
• Context matters. A team protecting a 2-0 lead and deliberately inviting pressure will rack up conceded xG without genuine defensive weakness.
• Small samples. xG needs 10+ matches to stabilise. Do not over-index on two or three games of data.
• Post-shot xG. Basic models use pre-shot data only. More advanced post-shot models account for where the ball is placed within the goal frame.