Expected Goals, Explained Without the Spreadsheet
The number on the broadcast graphic is simpler than it looks and more limited than its critics admit. What xG measures, what it cannot see, and how to read it over a season rather than a night.
Sunday, August 9, 2026/3 min read
Expected goals arrived on television faster than the explanation did, which is why it still produces arguments in which neither side is describing the same thing.
The concept itself takes one sentence. Every shot is given a value between 0 and 1 representing the proportion of historically similar attempts that were scored. A tap-in from two yards might be 0.9. A speculative effort from thirty yards might be 0.03. Add up a team's shot values and you have their xG for the match.
That is all it is: a way of counting chances that accounts for how good they were.
What goes into a shot's value
Models differ, but the main inputs are consistent and unglamorous.
Distance from goal and angle to it do most of the work. Then body part, since headers convert worse than feet from the same spot. Then the type of assist, because a cutback produces better chances than a cross. Whether the move was a counter-attack, a set piece or open play. And in better models, the position of the goalkeeper and defenders at the moment of contact.
Note what is absent. Almost no public model knows who is taking the shot. A chance worth 0.2 is worth 0.2 whether it falls to a prolific striker or a full-back. That is deliberate, because the point is to measure the chance rather than the finisher, but it is the source of a great deal of misreading.
What it is genuinely good at
Its real strength is prediction over time, not description of a single match.
Goals are rare, so results are noisy. A team can play well and lose to a deflection. Over a handful of matches, actual goals tell you as much about luck as about quality. xG accumulates faster because every shot contributes something, so it stabilises sooner.
Consequently, a team substantially outperforming its xG over ten matches is usually a candidate to regress, and one underperforming it is usually a candidate to improve. That is the single most useful thing the metric does, and it is why recruitment departments care about it more than pundits do.
What it cannot see
Four things, and honest analysts say so.
The shots not taken. A side that works the ball into the box repeatedly and never shoots records low xG. So does a side that cannot get near the box at all. The number does not distinguish them.
Finishing quality, mostly. Because shooter identity is usually excluded, a genuinely elite finisher will beat their xG persistently, and the model will call it luck every season. Distinguishing real finishing skill from variance requires several seasons of data, not several matches.
Game state. A team two goals up defends deep and concedes low-value shots by design. Their xG against rises and their performance was fine.
Goalkeeping. Standard xG says nothing about whether a save was extraordinary. That is what post-shot models are for, and they answer a different question.
How to use it without embarrassing yourself
Do not quote it for one match as though it were the true result. A single game contains too few shots for the number to mean much, and 1.4 against 1.1 is not a moral victory.
Do look at it across ten matches or more, and look at xG for and against separately rather than only the difference. A side creating 1.8 and conceding 1.6 is a different problem from one creating 0.9 and conceding 0.7, even though the margin matches.
Do check the shot count behind the number. Two xG from four shots is a team carving a defence open. Two xG from twenty-five is a team shooting from anywhere.
And be aware that different providers publish different values for the same match, because they use different models and different training data. Comparing one provider's figure against another's is meaningless.
Why it matters beyond the graphic
The reason clubs invest in this is money. Transfer fees are the largest discretionary spend in football, and a metric that separates a striker who had a lucky season from one who reliably gets into good positions is worth a great deal.
That logic applies wherever squads are being assembled quickly and expensively, which is precisely the situation across the Gulf, where the Saudi Pro League is now selling more players than it buys and recruitment has visibly shifted from acquiring available stars to working a positional shortlist. Understanding how those lists are built starts here, and it is a useful antidote to the reporting that surrounds every transfer window.
Published in The Outspoken Digest
Newsletter
The Digest, in your inbox
One edition, sent when it is ready. No noise, and your address is never passed on.
Read Next
More Sports →
FIFA Charges Argentina Players After the World Cup Final: What Happens Next
Jul 30, 2026/2 min read