Quick answer: Expected goals (xG) is a number from 0 to 1 that rates how likely a shot was to score, based on thousands of similar chances. Add up a team’s xG and you get how many goals their chances were worth — stripping out luck and finishing. It predicts future results better than the scoreline, which is why every AI soccer model is built on it.
Expected goals (xG) is the single most useful stat in modern soccer and the foundation of almost every AI football model. This guide explains xG in plain English, why it beats the final score, and how to use it when you bet.
What is expected goals (xG)?
Expected goals (xG) measures the quality of a chance: a tap-in from two yards might be worth 0.9 xG, a speculative 30-yard shot just 0.03. Sum a team’s xG across a match and you get how many goals they “should” have scored from the chances they created, removing luck and finishing variance.
| Type of chance | Approximate xG |
|---|---|
| Penalty | ~0.79 |
| Close-range tap-in | 0.70+ |
| Header from six yards | ~0.30 |
| Edge-of-box shot | ~0.05 |
| Speculative 30-yard strike | ~0.03 |
Why xG beats the final score
The scoreline lies more often than you think — a team can win 1-0 while being outplayed 0.4 xG to 2.1 xG, and that team is likely to lose the rematch. xG exposes these mismatches, which is why models and sharp bettors trust it. A side overperforming its xG usually regresses; one underperforming is often undervalued by the market.
| Team | Goals | xG | Read |
|---|---|---|---|
| Team A | 1 | 0.4 | Lucky — likely to regress |
| Team B | 0 | 2.1 | Unlucky — undervalued |
How AI models use xG
AI soccer models use xG and xG against (xGA) as the core measure of team strength, then adjust for form, lineups and venue to simulate a match thousands of times. The output is a probability for each outcome and a fair price for every market — exactly how the tools in our best AI for soccer guide work, and the process in our how AI predicts soccer deep dive.
Using xG when you bet
Use xG to find teams the market is mispricing: back sides that consistently create more than they concede but have poor recent results, and fade lucky teams riding unsustainable finishing. Always combine it with confirmed lineups, and only bet when the implied edge is real — our expected value guide shows how.
Related reading: how AI predicts soccer · best AI for soccer · expected value
Frequently Asked Questions
What is expected goals (xG) in soccer?
Expected goals (xG) rates how likely each shot was to score, from 0 to 1, based on similar historical chances. A team’s total xG shows how many goals their chances were worth, removing luck and finishing variance.
Is xG better than the final score?
For predicting future results, usually yes. The score reflects one match’s luck, while xG reveals who actually created and conceded the better chances, which is more repeatable.
How do AI models use xG?
AI soccer models use xG and xG against as the core measure of team strength, then simulate matches to produce outcome probabilities and fair prices for betting markets.
Can I bet using xG?
Yes. Back teams creating more than they concede but underperforming results, and fade lucky teams, but only when the price offers real value. Combine xG with lineups and context.
What is a good xG in a match?
There is no fixed number, but roughly 1.5+ xG suggests strong chances created, while under 0.7 xG is poor. Compare both teams’ xG rather than judging one in isolation.
What is the difference between xG and xGA?
xG measures the quality of chances a team creates; xGA (expected goals against) measures the quality of chances it concedes. Together they describe attacking and defensive strength.
