Football has always had a scoreboard problem: the final result rarely tells the whole story. A team can dominate for 90 minutes, hit the crossbar twice, force a string of saves, and still lose 1-0 to a deflected shot from nowhere. Expected goals, almost always shortened to xG, is the metric built to close that gap between what happened and what probably should have happened. Over the past decade it has moved from spreadsheets on niche analytics blogs to the graphics on primetime television, and it has quietly changed the way millions of fans watch the game.
What Exactly Is Expected Goals?
Expected goals is a measure of chance quality. Every time a player takes a shot, an xG model estimates the probability that the shot will become a goal, expressed as a number between 0 and 1. A value of 0.10 means a chance like that is scored roughly once in every ten attempts. A value of 0.90 means it is a near-certain goal, the kind of tap-in almost anyone would finish.
Crucially, the number comes from history, not opinion. Analysts feed a model hundreds of thousands of past shots, each labelled with whether it ended up in the net. The model learns how often shots with similar characteristics are scored. When a new shot is taken, it is compared against all those historical attempts to produce its xG value. Add up every shot a team takes in a match and you get their total xG for the game, which is a rough estimate of how many goals an average team would have scored from those same chances.
What Goes Into an xG Value
Not all shots are equal, and a good model knows it. A tap-in from two yards and a hopeful strike from 30 yards both count as one "shot" in the old-fashioned stats, but their xG values are worlds apart. Modern models weigh a range of factors, including:
- Distance from goal — the single biggest driver; closer shots are worth far more.
- Angle to goal — a shot from directly in front is easier than one from a tight angle near the byline.
- Body part used — headers are converted less often than shots taken with the feet.
- Type of pass — a cutback or through ball that leaves a clear sight of goal raises the value; a whipped cross usually lowers it.
- Type of play — open play, fast break, corner, direct free kick, and penalty all behave differently.
- Defensive pressure — how many defenders, and the goalkeeper's position, sit between the ball and the net.
Penalties make a handy anchor. Because they are taken from the same spot under near-identical conditions every time, they carry a fixed value of roughly 0.76 to 0.79, reflecting the simple historical fact that about three in four penalties are scored.
Typical xG Values at a Glance
To build intuition, it helps to see how different chances are priced. The figures below are approximate and vary between models, but they show the enormous range hidden inside the word "shot."
| Type of chance | Approximate xG |
|---|---|
| Tap-in into an empty or open net | 0.90+ |
| Penalty kick | ~0.76 |
| Close-range shot inside the six-yard box | 0.35 – 0.50 |
| Header from a cross around the penalty spot | 0.10 – 0.15 |
| Shot from the edge of the box, central | 0.05 – 0.08 |
| Speculative effort from 25+ yards | 0.02 – 0.03 |
How to Read xG Like an Analyst
At team level, the trick is to compare xG with actual goals over many matches, not one. If a side consistently scores more than their xG suggests, they either possess exceptional finishers or they are riding luck that tends to fade. Comparing a team's xG created against the xG they concede often delivers a fairer verdict on a match than the scoreline itself.
At player level, comparing a striker's goals to their xG is essentially a finishing report card. Consider two forwards:
- A striker with 12 goals from 8.0 xG is either genuinely clinical or overdue a cold streak.
- A striker with 5 goals from 9.0 xG has been getting into excellent positions and will, more often than not, start converting.
There are sibling metrics too. Expected assists (xA) applies the same logic to the pass that creates a shot, rewarding the players who manufacture high-quality chances even when a teammate fails to finish them. The single most important habit in all of this is to think in samples rather than single games. Over one match, xG is noisy and easily skewed by one big chance. Over a full season it becomes one of the most reliable guides to how good a team really is.
The Limits of xG
xG is powerful, but it is a model, not a verdict, and it pays to know where it stops being useful:
- It does not fully capture finishing skill. Most models treat every shooter as an average finisher, so a genuinely elite striker may beat their xG season after season.
- Providers disagree. Companies such as Opta, StatsBomb and sites like Understat build their models differently, so the same shot can carry different values depending on where you look.
- It ignores what happens after the shot. A rocket into the top corner and a scuffed effort straight at the keeper from the same spot receive the same xG.
- Small samples mislead. A handful of shots, or a single dramatic game, can produce numbers that look meaningful but are mostly noise.
Used sensibly, xG is a lens rather than a law. It tells you whether a result was earned or fortunate, whether a striker's drought is bad luck or bad play, and whether a hot start to the season is built to last. That is why it now sits on the touchline graphics of matches watched by tens of millions of people. It has not replaced watching football with your own eyes, but it has given fans a sharper, more honest vocabulary for talking about what they see. The next time your team loses a game they controlled, glance at the xG. The scoreline may say one thing, but the underlying numbers might tell you the performance was better than you feared.