What Is xG (Expected Goals)? A Simple Explanation
If you've watched any football coverage in the last few years, you've almost certainly heard the term "xG" thrown around. Pundits mention it, post-match graphics display it, and Twitter arguments rage over it. But what does it actually mean?
Let's break it down properly — no maths degree required.
The Basic Idea
Expected Goals, or xG, is a way of measuring the quality of a chance. Every shot in football gets assigned a number between 0 and 1, representing the probability that the average player would score from that exact situation.
An xG of 0 means there's basically no chance of scoring. An xG of 1 means it's a guaranteed goal. Everything else falls somewhere in between.
Here are some real benchmarks to give you a feel for the scale:
- A penalty has an xG of about 0.76 — roughly three in four are scored
- A one-on-one with the keeper from 8 yards might be around 0.40
- A shot from the edge of the box with no defenders blocking could be about 0.08
- A header from 15 yards out has an xG of roughly 0.04 — one in twenty-five
- A long-range effort from 30 yards might be as low as 0.02
So when you see a match where Team A had 2.3 xG and Team B had 0.8 xG, it means that based on the chances created, you'd expect Team A to score about twice and Team B to score once. If the final score was 0-1 to Team B, something unusual happened — Team B were clinical with limited opportunities, or Team A were wasteful, or the goalkeeper had a blinder, or all three.
How Is xG Calculated?
xG models are built using enormous datasets of historical shots — hundreds of thousands of them. Each shot is logged with a load of contextual information, and then machine learning works out which factors most influence whether a shot becomes a goal.
The key factors that go into most xG models include:
Distance from goal
This is the big one. The closer you are to goal, the higher the xG. A shot from six yards is worth far more than one from twenty-five. This should be obvious to anyone who's ever played football, but it's surprising how much this single variable explains.
Angle to goal
A shot from directly in front of goal has a much higher xG than one from a tight angle near the byline. The narrower the angle, the smaller the target the shooter can see, and the easier it is for the goalkeeper to cover.
Body part used
Shots with the foot convert at a higher rate than headers, which convert at a higher rate than shots with other body parts. This makes intuitive sense — you have far more control over a football with your foot than with your head.
Type of assist
Was it a through ball? A cross? A cutback? A set piece? The type of pass that preceded the shot has a significant impact. Cutbacks, for instance, tend to produce high-xG chances because the ball is being played across the face of goal to a player who's already in a good position.
Game state
Some models factor in whether the shooting team is winning, losing, or drawing. A team chasing a goal may take more speculative shots, which would affect the context.
Shot type
Is it a volley? A placed shot? A header from a corner? Different techniques have different conversion rates historically.
Defensive pressure
More advanced models also account for how many defenders are between the shooter and the goal, and how close those defenders are. An open shot from 12 yards is worth far more than a shot from the same distance with three players in the way.
What xG Is NOT
This is where people get confused, and where a lot of the arguments start. Let's clear up the biggest misconceptions.
It's not just shot count
This is the most common misunderstanding. Having 20 shots and 1.2 xG is not the same as having 5 shots and 1.2 xG. The first team took a load of low-quality potshots from distance. The second team created fewer but much better chances. xG tells you about shot quality, not shot volume.
It doesn't account for the individual shooter
Standard xG models use the average player as their baseline. They don't know that Mo Salah is taking the shot from the right side of the box or that a centre-back is having a speculative dig from 30 yards. That's by design — it's measuring the quality of the chance, not the quality of the player.
There are variations like xGOT (expected goals on target) that factor in shot placement, and some models try to incorporate shooter quality, but classic xG deliberately strips this out.
It's not meant to "predict" individual matches
xG is a probabilistic tool. A team with 3.0 xG won't always score three goals. Sometimes they'll score none. Sometimes they'll bag five. Over a single match, anything can happen. But over a season of 38 games, xG is remarkably good at showing which teams are genuinely good at creating chances and which have been riding their luck.
It doesn't measure everything
xG can't capture a brilliant dribble that doesn't end in a shot. It can't measure the through ball that was slightly overhit and never became a chance. It focuses specifically on shots and their likelihood of becoming goals. Other metrics like xT (expected threat) and progressive passing stats try to capture those earlier phases of play.
Why Should You Care About xG?
Here's the really interesting bit. xG is one of the best tools we have for separating genuine quality from luck.
Spotting teams that are better (or worse) than their results suggest
In the 2022/23 Premier League season, Brighton under Roberto De Zerbi were a perfect example. They created chances worth about 60 goals across the season based on xG but actually scored considerably fewer. Their underlying numbers suggested they were performing at a level far above where they finished. The following season, with similar underlying metrics, they continued to be one of the most watchable sides in the league.
Conversely, a team that's winning matches despite low xG numbers is often living on borrowed time. They might be relying on a goalkeeper in the form of his life, or a striker who's converting every half-chance. That kind of performance is very hard to sustain over a full season.
Understanding tactical quality
xG helps you see past the scoreline to understand what actually happened in a match. A 1-0 win doesn't tell you much. But if the winning team had 0.4 xG and the losing team had 2.1 xG, you know the result was a smash-and-grab. If the winning team had 3.2 xG and the losing team had 0.3 xG, it was a dominant performance where the margin of victory probably should have been larger.
Making better predictions
This is where it gets really practical. Teams that consistently create high-xG chances are more likely to score goals in future. Teams that consistently allow low-xG chances against them are more likely to keep clean sheets. If you're trying to predict what will happen in upcoming matches, xG data from recent games gives you a far better foundation than just looking at the league table or recent results.
A team on a four-match winning streak where all four wins came from low-xG performances is probably not as strong as their record suggests. A team on a four-match losing streak where they created loads of chances but couldn't convert is probably better than their record suggests.
A Quick Worked Example
Let's say Manchester City play Wolves. In the match, these shots are taken:
City's shots:
- Header from 6 yards off a corner: 0.15 xG
- Shot from 20 yards, deflected: 0.03 xG
- One-on-one after a through ball: 0.42 xG
- Penalty: 0.76 xG
- Shot from edge of box, open play: 0.07 xG
- Total: 1.43 xG
Wolves' shots:
- Counter-attack, shot from 14 yards: 0.12 xG
- Long-range free kick: 0.05 xG
- Header from a corner: 0.06 xG
- Total: 0.23 xG
If the final score is City 2-0 Wolves, that's broadly in line with what the chances suggested. City created significantly better opportunities and converted a reasonable proportion of them.
If the final score is Wolves 1-0 City, that's a major overperformance by Wolves and underperformance by City. It happens — that's football. But if City keep creating 1.4 xG per game and keep failing to win, you'd expect their results to improve. And if Wolves keep creating only 0.23 xG per game, their results will almost certainly get worse.
The Bottom Line
xG isn't perfect. No single metric can capture the full complexity of football. But it's the best widely-available tool we have for understanding the quality of chances a team creates and concedes, and for cutting through the noise of individual match results to see what's really happening underneath.
Once you start thinking in xG terms, you'll never watch football the same way again. You'll spot the team that's riding their luck before the inevitable slide. You'll see the team creating brilliant chances before their results catch up. And you'll win a lot more arguments on the internet.
Welcome to the world of expected goals.