Expected goals, shortened to xG, is a number between 0 and 1 that estimates how likely a single shot was to result in a goal, based on thousands of similar shots taken in the past. A tap-in from two yards might carry an xG of 0.95, while a speculative strike from the halfway line might sit closer to 0.01.
Add up the xG of every shot in a match and you get a team’s total xG for that game, a figure that has quietly become one of the most-quoted stats in modern football coverage.
Key Takeaways
- xG is a number between 0 and 1 that estimates how likely a shot was to result in a goal, based on thousands of similar past shots.
- Key factors behind an xG value include shot distance, angle, body part, assist type, and defensive pressure.
- A team can dominate on xG and still lose, which separates chance quality from actual finishing.
- Different providers such as Opta and StatsBomb run their own models, so totals can differ between sites.
- xGOT (expected goals on target) sharpens the picture further by only scoring shots that were actually on target.
How Is xG Actually Calculated?
xG is built from a model trained on a very large set of historical shots, each one tagged with the factors that affected whether it went in. Stats Perform says Opta’s model is trained on nearly one million shots. The model then scores every new shot against that history rather than against any single rule.
The main factors that push an xG value up or down include the following.
| Factor | Effect on xG |
|---|---|
| Distance from goal | Closer shots score a much higher value |
| Angle to goal | A central position scores higher than a tight angle |
| Body part | Shots with the foot generally outscore headers |
| Type of assist | A cutback or through ball scores higher than a long ball |
| Defensive pressure | A crowded box or a tight marker pulls the value down |
| Goalkeeper position | An exposed keeper raises the value; a set keeper lowers it |
None of these factors work alone. A close-range header under pressure from two defenders can still carry a modest xG, while an unmarked shot from the edge of the box after a cutback can climb surprisingly high.
What Does an xG Score Actually Tell You?
An xG score tells you the quality of the chances a team created or conceded, separate from whether the finishing was clinical or wasteful on the day. A team can lose 2-1 while producing 2.4 xG to their opponent’s 0.6, which is the analytics way of saying they were the better side and just finished poorly.
That gap between actual goals and xG is often described as overperformance or underperformance. A striker who consistently outscores their xG total is finishing better than the average player would from the same chances, which is either a sign of real quality in front of goal or a hot streak that is unlikely to last. The same logic works defensively: Spain held opponents to tiny xG totals throughout the 2026 World Cup, down to just 0.13 for Argentina in the final.
Where You Will See xG Used
Broadcasters now show a live xG bar alongside the scoreline on most major matches, and it shapes analysis well beyond the highlights package.
- Judging performance beyond the scoreline. A dominant xG total with no goals to show for it is treated very differently to a scrappy 1-0 built on a single shot.
- Recruitment and scouting. Clubs use a player’s shot quality and xG output over a full season to judge finishing ability more fairly than raw goal counts allow.
- Setting expectations. Pundits use season-long xG trends to flag a team riding a run of results that its underlying numbers do not support.
What Are the Limits of xG?
xG measures shot quality, not the value of everything that happens before the shot. A team that strings together twenty passes to manufacture one clean chance does not get extra credit in the model for the buildup, only for the shot itself.
It also cannot account for context that a human eye picks up instantly, like a forward through on goal in the eighty-ninth minute with legs gone, or a wet pitch turning a routine finish into a scramble. Treat xG as a strong guide to chance quality, not a verdict on a match.
Why Different Sites Show Different xG Numbers
Opta and StatsBomb each run their own model, trained on their own shot data, so the same chance can carry a different xG value depending on which provider you are looking at. Most sites that publish xG license it from one of these providers, which is why two broadcasters can show two different totals for the same match without either one being wrong.
A related stat worth knowing is xGOT, or expected goals on target, which only scores shots that were actually on target and adjusts for exactly where in the goal the ball was heading. It strips out the shots that never troubled the goalkeeper and gives a sharper read on finishing quality in the moments that mattered.
xG, npxG, xA and xGOT: The Related Stats
xG is the starting point for a family of chance-quality metrics. The table shows how the main ones differ.
| Stat | What it measures | Good to know |
|---|---|---|
| xG | Chance quality of a shot | A value between 0 and 1 |
| npxG | xG with penalties removed | Lets you compare strikers like for like |
| xA | How likely a completed pass is to become an assist | Credits the passer, not the shooter |
| xAG | The xG of the shots that follow a pass | Isolates passes that lead to a shot |
| xGOT | Quality of an on-target shot, including where it was heading | Separates good finishing from good chances |
| xGA | The xG a team concedes | The defensive mirror of xG |
How Much Is a Penalty Worth in xG?
Every penalty is taken from the same spot, so providers give it a flat value. FBref’s explainer, written when the site carried Opta data, put a penalty at 0.79 xG, while a Stats Perform article quotes 0.76. The exact figure depends on the model and its version, which is why analysts often quote npxG when comparing strikers. FBref stopped publishing Opta’s advanced stats, including xG, in January 2026.
Shot Quality Beats Shot Volume
Stats Perform gives a clean example of why xG exists. In one Opta example, Tottenham produced 2.14 xG from 22 shots, while Nottingham Forest scored twice from just four shots worth 0.48 xG. Many shots do not mean many good chances, and a few clear chances can be enough to win.
xG Versus xGOT: A Free Kick Example
Opta’s xGOT explainer uses a James Maddison free kick against Aston Villa. The free kick position carried an xG of just 0.12, but the way he bent it around the wall and inside the near post earned an xGOT of 0.73. In other words, xG says the chance was difficult, and xGOT says the execution was excellent.
Where to Find xG
Opta’s xG model powers many broadcaster graphics, and Opta’s own analysis is published on The Analyst. Because each provider trains its own model, only compare numbers that come from the same source.
Quick Questions About xG
Can a team win with a lower xG than its opponent?
Yes. xG measures chance quality, not results. A side can create fewer or poorer chances and still win through finishing, a goalkeeper error or a set piece.
Is xG the same as shots on target?
No. xG scores every shot, on target or not, by how likely it was to score. Shots on target only count the ones that tested the goalkeeper, which is the idea behind xGOT.
How does xG relate to goal difference?
Goal difference records what happened, while xG difference estimates what should have happened. A team with a much better goal difference than xG difference may be running hot, and the clean sheet record of its defence can be read the same way.
Making Sense of xG Next Time You Watch
Once you know what the number is built from, an xG figure on screen stops being a mystery stat and starts being a shortcut for a question you were already asking: were those chances actually good ones? A team piling up shots from distance can look busy while their xG stays low, and a side with three shots all season can post a healthy total if every one of them was a clear-cut chance.
That is the whole point of the metric. It puts a number on chance quality so that a dominant, wasteful performance and a lucky, backs-to-the-wall win no longer look identical on the scoresheet.