What xG actually is (and what it isn’t)
xG is the most used and most misunderstood number in football. Pundits wave it away, broadcasters flash it up without explanation, and half of the arguments about it online are really arguments about what people think it means. Here's what it actually is, in plain English, and what it can and can't tell you.
The idea in one paragraph
Expected goals (xG) measures chance quality. Every shot is compared against hundreds of thousands of historical shots taken from similar situations, and given a value between 0 and 1 representing the probability that an average player scores it. A tap-in from two yards might be 0.9 xG: scored nine times out of ten. A hopeful hit from 30 yards might be 0.02: scored once in fifty attempts. Add up the values of every shot a player or team takes and you get their xG total, which is the number of goals an average finisher would have scored from those chances.
What goes into the number
The value of a chance is driven by the things that genuinely change scoring probability: distance from goal, the angle to it, the body part used (feet score more than heads), the type of assist (a cutback beats a hopeful cross), and the situation (open play, counter-attack, set piece). What matters is that xG is measured before the shot is struck. It judges the chance, not the execution; a brilliant save and a wild miss leave xG unchanged.
What xG is genuinely for
- Judging process over results. Goals are so rare that a match or even a month of them is mostly noise. A team creating 2.0 xG a game while conceding 0.8 is doing the right things, whatever the recent scorelines; a striker on a goalless run who keeps racking up xG is getting chances, and chances keep arriving before goals do.
- Measuring finishing. Compare goals scored to xG over a real sample and you get the cleanest available read on conversion skill. That's the entire basis of our live best-finishers ranking.
- Separating the player from the team. A striker's raw goal tally is half service; his xG shows the service directly, so you can tell a starved finisher from a wasteful one.
What xG is not
- It's not a prediction about one shot. A 0.4 xG chance missing isn't a failure of the model; it's the more likely outcome. xG only means anything summed over samples.
- It's not the same everywhere. Different providers use different models, so one player's xG varies slightly by source. Trends and gaps are comparable; decimals between websites aren't.
- It's not all goals. Penalties are worth roughly 0.79 xG every time, regardless of the taker, which is why serious analysis uses npxG (non-penalty xG). A player's totals can be quietly inflated by spot-kick duty, and stripping penalties out is the first thing FDscout does when measuring finishing.
- It's not proof by itself. "He scored above xG this season" can be skill or luck. Repetition and volume turn it into evidence; a single hot month doesn't. (How many minutes before you can trust a stat is a rabbit hole of its own.)
How FDscout uses it
Every player profile carries xG, npxG, xG per shot (are his chances good, or just many?) and goals versus expectation, each ranked as a percentile against players in the same position across 30 leagues, so you always know whether a number is genuinely high or just looks it. Sustained overperformance at volume earns the xG out performer trait badge; consistently good shot locations earn Shoots from good positions. The follow-up question, whether a player strikes the ball better than his chances deserve, has its own metric and its own explainer: xG vs xGoT.
Look any player up on FDscout and the xG picture is one glance. And when a pundit next says a player "should have scored" a 0.3 chance, you'll know exactly how much that claim is worth.