What Is xG in Football? Expected Goals Explained
Updated: 23 August 2026
xG (expected goals) is a metric that grades the quality of scoring chances: every shot gets a probability of becoming a goal from 0 to 1, and the sum of those probabilities over a match is a team's xG. A penalty is worth ≈ 0.76 xG, a shot from the centre of the box ≈ 0.1–0.4 xG, a long-range effort ≈ 0.02–0.05 xG.
How xG Is Calculated
The model is trained on hundreds of thousands of historical shots. For each new shot it weighs distance and angle to goal, body part (foot/header), attack type (open play, counter, set piece), and the positions of defenders and the goalkeeper. The output is the probability of a goal from that exact situation; the sum across all of a team's shots is its match xG. That is why "xG 2.3 — 0.4" reads as: one side created chances worth more than two goals, the other less than half a goal.
The xG Scale: What a Specific Number Means
| Shot xG | What kind of chance it is |
|---|---|
| 0.02–0.05 | long-range effort from outside the box |
| 0.07–0.15 | shot inside the box under pressure or from a tight angle |
| 0.15–0.35 | good position in the box, a half-chance |
| 0.35–0.60 | clear goalscoring opportunity |
| ≈0.76 | penalty kick |
| 0.90+ | point-blank finish into an open net |
At team level per match: xG under 0.7 means the attack created almost nothing; 1.0–1.7 is a normal creative output; 2.0+ is on-pitch dominance where failing to win is already a statistical anomaly — exactly the kind of contradiction our Crime Index flags.
Reading Match xG: Three Scenarios
Score matches the xG — the result is earned: the stronger side created more and scored more. Score above xG — the team converted nearly everything it created: either a day of clinical finishing or a run of luck that rarely repeats. Score below xG — the team dominated on chances but left without points: the classic "xG robbery". The third scenario is the one that most often means the league table is lying about a team's real strength.
Why the Same Match Has Different xG on Different Sites
There is no single official xG. Understat, Opta, FBref, SofaScore and others train their own models on their own shot databases, and the feature sets differ: some account for the goalkeeper’s position, some do not, some model headers from crosses separately. A gap of 0.1 to 0.4 xG per match between providers is normal and does not mean one of them is wrong. The practical rule: only compare xG within a single source. Taking one team’s xG from one site and their opponent’s from another, then subtracting, produces a meaningless number.
npxG, xGA and xA: the Neighbouring Metrics
npxG (non-penalty xG) strips penalties out. It matters because a single spot kick adds roughly 0.76 to the total, so a side with two penalties looks far more creative than it was in open play. xGA is expected goals against — the same metric viewed from the defensive end. xA (expected assists) values the pass before the shot rather than the shot itself: how likely that ball was to end in a goal. A forward with high xG and no xA shoots a lot and creates little for team-mates; the reverse profile is a provider.
How Many Matches Before xG Means Anything
Over a single match xG says very little about how good a side is — there is too much noise in one game: a deflection, a goalkeeping performance, a red card on twenty minutes. The metric starts to work across roughly ten matches, once the outliers average out. That is why “this team plays better than the table suggests” is a reasonable claim about a season and close to meaningless about a single round. Over ten games, xG minus xGA is a more honest measure of a side than the points they have collected.
What xG Cannot Do
xG knows nothing about what did not happen: a brilliant pass that the striker never shot from does not enter the metric. It is unreliable over a single match — variance is strong there, so conclusions about team strength are safer over 5–10 games. And xG ignores game state: a side leading 3-0 deliberately stops creating, and its low second-half xG is not weakness.
xG and the Crime Index on Bravsen
Bravsen automatically compares every match's xG with the actual score and computes the Crime Index — a 0–100% measure of the contradiction. Above 60% the match is flagged as a statistical anomaly and gets a full breakdown: who was really stronger, where the game was decided and what the numbers say about both teams going forward. The methodology lives on the about page, all breakdowns are in the match review archive, and the latest ones are on the homepage.
FAQ
What does xG 1.5 mean?
xG 1.5 means the chances a team created would, on average, produce one and a half goals: replay the same set of shots many times and they yield about 1.5 goals per match. The actual score can be 0 or 3 — xG measures the quality of play, not a guaranteed outcome.
Can a team win with a lower xG?
Yes, and it happens regularly: an xG advantage shows who created better chances, but finishing decides the match. These contradictions are exactly what Bravsen tracks with the Crime Index — the more the scoreline contradicts the xG, the higher the anomaly score.
What is the difference between xG and xGA?
xG is expected goals created by a team (attack). xGA (expected goals against) is expected goals conceded at their own end (defence). The xG − xGA difference over a run of games is one of the most honest measures of team strength.
How is xG calculated in football?
The model grades every shot by distance and angle to goal, body part, attack type and defender positions, assigning it a scoring probability from 0 to 1. The sum of those probabilities across a team's shots is its match xG. See the "How xG Is Calculated" section above for the full breakdown.
How much xG is a penalty worth?
Most public xG models assign a penalty a fixed value around 0.76-0.79 xG, based on the long-run conversion rate of penalties across leagues. Unlike open play, where distance, angle and defenders shift the odds shot by shot, a penalty is taken under the same conditions for every player, so models don't recalculate it individually.
What is xG in hockey?
The same idea as in football: every shot attempt gets a probability of becoming a goal — based on distance, angle, shot type and traffic in front of the goalie. The sum over a game is a team's xG. Bravsen covers football xG, but the metric works identically across both sports.
How do I find the xG for a specific match?
You need a source that computes xG game by game, not just season tables. On Bravsen the xG for both sides sits at the top of every match report next to the score, shots and possession — the easiest way in is the match archive or the page of the team you are interested in, which collects all of their games.
Why is xG better than a plain shot count?
A shot count treats a tap-in and a thirty-yard effort as the same unit. xG weights every attempt by how likely it was to be scored, so "15 shots for 0.6 xG" and "5 shots for 1.8 xG" describe completely different matches even though the first side had three times the attempts. That is why the team with fewer shots is often the one that actually threatened.
Where can I see xG for matches?
On Bravsen, both teams' xG is published in every match review together with full stats — shots, possession and the Crime Index. All breakdowns live in the match review archive; the latest ones are on the homepage.