What does xG mean in football—and can you trust it?

What does xG mean in football?

xG stands for expected goals. It measures the likelihood of an individual shot becoming a goal. Every attempt is given a value between zero and one by comparing it with previous shots taken in similar circumstances.

A shot valued at 0.20 xG is the sort of chance that would be expected to produce approximately two goals across ten similar attempts. That does not mean the player has scored one-fifth of a goal, and it does not mean the chance definitely should have been scored. It is a probability—not a verdict on the player.

How is xG calculated?

Different data companies operate their own expected-goals models. The model looks at what happened to a large number of comparable shots. If 20 out of every 100 similar attempts became goals, the chance may receive a value around 0.20 xG.

More advanced models can include detailed tracking information. Research using Bundesliga data has shown how player positioning—particularly the goalkeeper’s position—can materially change the quality of two shots taken from roughly the same place.

• The distance from goal and angle of the shot.
• Whether the ball was struck with a foot or headed.
• The type of pass or action that created the chance.
• Whether it came from open play or a set piece.
• The position of defenders and the goalkeeper, when that data is available.

A simple xG example

Imagine a team has the following three attempts. These figures are purely illustrative.

• The combined match total is 0.80 xG.
• The side might still score twice, once or not at all.
• The total describes the overall quality of the chances; it does not dictate the result.

How can a team win on xG but lose the match?

Because chances are not goals. A goalkeeper can make an excellent save. A striker can miss from six yards. A speculative effort can fly into the top corner.

Suppose one team creates six reasonable chances worth a combined 1.8 xG but fails to score. Its opponent has one difficult attempt worth 0.08 xG and scores. The final result is still 1–0. There are no imaginary league points for winning the spreadsheet.

What xG tells us is that the losing team generally created the better shooting opportunities. If the same pattern continued across numerous matches, it would be worth investigating rather than dismissing it as irrelevant.

Does higher xG mean a team deserved to win?

Not necessarily. ‘Deserved’ is a judgement. xG is a measurement.

Expected goals normally assess shots—not every promising attack. A brilliant pass across an empty goal can produce no xG if nobody reaches the ball and takes a shot. Equally, a team protecting a lead may deliberately concede possession while restricting its opponent to poor attempts.

Tactics, game state, red cards and individual decisions all affect how a match develops. It is more accurate to say a team created the better shooting opportunities than to claim xG proves it deserved to win.

Why do websites show different xG figures?

There is no single universal xG model. One provider might concentrate on shot location, angle, body part and the preceding action. Another might also have access to goalkeeper position, defensive pressure and detailed player-tracking information.

Providers can also define and record match events differently. The same chance might therefore receive slightly different values on two websites. It does not automatically mean either figure is wrong; the models are making estimates using different information and methods.

For sensible comparisons, use figures from the same provider wherever possible. A difference between 0.18 and 0.22 is rarely worth starting a family argument over.

Is xG accurate?

xG is useful, but it is not perfect. A properly constructed model should be tested against large numbers of real shots. It can then show whether chances given a particular probability are converted at roughly that rate over time.

Individual outcomes remain unpredictable. A 0.70 chance can be missed and a 0.03 chance can be scored without the model being broken. Unlikely things happen in football every week.

The value becomes clearer over larger samples. One match can be chaotic. A pattern across ten, twenty or thirty matches may reveal much more about how regularly a team creates and concedes good chances.

A peer-reviewed expected-goals study based on more than 100,000 Bundesliga shots found that positional and event data could provide a useful process-based measure of team and player performance. That does not make every commercial xG model identical or infallible.

What is xG genuinely useful for?

Ten hopeful efforts from outside the box are not necessarily better than three excellent chances close to goal. A basic shot count treats them equally; xG attempts to separate them.

A team repeatedly creating strong chances but scoring fewer goals than expected may be experiencing poor finishing, good opposition goalkeeping or ordinary short-term variation. A side continually scoring from difficult chances may be enjoying outstanding finishing, but that level could be difficult to maintain.

Average xG per shot can indicate whether a team regularly works the ball into dangerous areas or settles for lower-quality attempts. The statistic also adds context: a 2–0 victory can come from a dominant display or two isolated moments in an otherwise even match.

xG does not provide the entire explanation. It identifies the question worth asking.

What does xG miss?

Most expected-goals models begin when a shot is taken. That creates important limitations. Models are improving as better event and tracking data become available, but no single number can contain an entire football match.

• Dangerous attacks that never produce a shot.
• A final pass narrowly missed by every attacker.
• A player choosing to pass instead of shoot.
• Tactical control without frequent attempts.
• The influence of the score on each team’s approach.
• Every aspect of pressure, positioning and shot technique.

Does xG include penalties?

Usually, yes. Penalties are normally assigned a fixed xG value based on historical conversion rates. The precise value can vary by provider.

Statistics labelled non-penalty xG, often shortened to npxG, remove penalties. This can be useful when comparing the open-play and set-piece output of teams or players who receive very different numbers of penalties. Always check which version is being displayed before comparing figures.

What do xGA, xGD and post-shot xG mean?

xGA means expected goals against: the combined quality of the chances a team allowed its opponents to take.

xGD means expected-goal difference: a team’s xG minus its xGA. A positive figure means it created more expected goals than it conceded.

npxG means non-penalty expected goals: expected goals with penalty attempts removed.

Post-shot xG is calculated after the shot is struck. It can include where an on-target effort was placed and is commonly used when analysing goalkeeping. Naming and methodology vary between providers.

How to use xG without overcomplicating football

Expected goals should make football easier to understand—not turn every conversation into a maths lesson.

• Remember that the score is the result. xG explains chance quality; it does not rewrite the table.
• Look for patterns rather than obsessing over one match. A longer run is usually more informative than one chaotic afternoon.
• Use it alongside what you watched. Statistics provide evidence and context, not a replacement for understanding tactics, decisions and game state.

The result and the performance can differ

The next time your team loses despite producing the higher xG, you do not have to pretend the result never happened. You can simply recognise that the performance and the outcome may be telling different stories.

The score tells you what happened. xG helps explain the route taken to get there.