Expected goals, or xG, has become one of the most widely used statistics in football. You see it on TV graphics, in match reports and in club analysis departments. But what does it actually measure, and how can a coach use it without getting lost in numbers? This guide explains xG in simple terms.
What is expected goals (xG)?
Expected goals measures the quality of a chance. Every shot gets a value between 0 and 1 that represents the probability that a shot like this one ends in a goal. The value is based on thousands of historical shots with similar characteristics.
For example, a shot with an xG of 0.10 is the type of chance that is scored about once in ten attempts. A penalty usually has an xG of around 0.75–0.8 in most public models, while a header from a tight angle outside the six-yard box might be worth less than 0.05.
What goes into an xG model?
Each data provider builds its own model, but most use similar factors:
- Distance from the goal
- Angle to the goal
- Body part (foot or header)
- Type of assist – through ball, cross, cut-back, rebound
- Game situation – open play, counter-attack, set piece, penalty
- In advanced models: defender and goalkeeper positions from tracking data
Machine learning is used to learn how each of these factors changes the probability of scoring.
How to read xG in a match
Add up the xG of all shots and you get a team’s expected goals for the match. If your team wins 1–0 but the xG is 0.6 vs 2.1, you probably rode your luck. If you lose 0–1 with an xG of 2.4 vs 0.5, your team created the better chances and the performance may have been better than the result.
Over a season, xG is a strong indicator of how well a team creates and prevents chances – often a better predictor of future results than the current table.
Why xG matters for coaches
- Evaluate performance, not just results: separate good processes from lucky outcomes.
- Improve chance quality: find out which attacking patterns create high-xG chances (for example cut-backs) and train them.
- Defend smarter: track the xG you concede and where those chances come from.
- Shot selection: show players why a pass into the box can be worth more than a long-range shot.
- Recruitment and player feedback: compare a striker’s goals with their xG over time.
Limits of xG
- It describes chances, not the whole game – build-up play and pressing are not directly included.
- Small samples are noisy: one match says little, ten matches say much more.
- Different providers use different models, so values are not always comparable.
- Finishing quality and goalkeeper performance are partly outside the basic model.
Related metrics
- xA (expected assists): the xG value of the shots a player creates.
- xGOT (expected goals on target): includes where the shot is placed in the goal.
- xT (expected threat): values passes and carries by how much they increase the chance of scoring.
A simple xG routine for your team
- Collect the location and type of every shot in your matches.
- Use a public xG table or simple model to estimate the value of each chance.
- After each match, compare goals vs xG – for and against.
- Review the highest-xG chances on video with your players.
- Track the trend over 5–10 matches before drawing conclusions.
Want to go further? Read how AI is changing tactical analysis and how computer vision tracks players to feed advanced models.
Go deeper: Modern Football Intelligence
From xG to computer vision and AI decision support – the book by Anas Skiti explains the methods behind modern football analytics in a practical way.