Expected goals (xG), explained without the hype

Expected goals, or xG, is the stat that took over football and then got shouted about until nobody was sure what it meant. Strip away the noise and it is a simple idea, honestly useful and easy to over-read. Here is what it is, what it is genuinely good for, and the list of things it quietly cannot tell you.

A number on every shot

The core move is this: every shot is given a number between zero and one, the probability that it becomes a goal. A model works it out by comparing the shot with a huge database of past shots from similar situations. How far out was it? At what angle? A header or a foot? A clear one-on-one or a crowd of defenders? From a cross, a through ball, or a rebound? If shots that looked like this one went in 20 times out of 100, the shot is worth 0.20 xG.

Add up the value of a team’s shots and you get its xG for the match. A team that racks up 2.4 xG built the kind of chances that, on average, yield about two and a half goals. It says nothing about whether they actually scored them. That gap, between the chances created and the goals taken, is where all the interesting arguments live.

What it is genuinely good for

The best use of xG is telling apart two teams that look identical on the scoreboard. One side wins 2-1 having carved out five clear chances and conceded one half-chance. Another wins 2-1 off two long-range deflections while being battered. The score is the same; the xG is not, and over the next ten games the second team’s luck tends to run out.

xG separates a team finishing hot from a team creating good chances. A striker outscoring his xG for a month is usually finishing above his true rate, not reinventing himself, and the numbers will drift back. Over a season, a team’s xG describes its real level better than its actual goals do. It is a fairer summary than the scoreline, precisely because it strips out the coin-flips.

What xG misses, honestly

This is the part the hype skips, and the part worth reading twice. xG treats a shot as the average of every shot like it, which means it is deliberately blind to some real things.

The finish. xG does not know this shot was curled into the top corner and that one was scuffed at the keeper. Placement and technique are exactly what it averages away, so a genuinely elite finisher will beat their xG for reasons that are not luck.

The goalkeeper. A 0.30 chance against a world-class keeper is not really 0.30. Standard xG scores the shot, not the save behind it.

Pressure and game state. It struggles to price how tight the defending was, and it does not understand context: a team two goals up shoots less and takes worse chances on purpose, so their xG can undersell a game they were controlling.

Small samples.One match of xG is noisy. A single big chance missed or a deflection in can swing it hard. It earns its keep over dozens of games, not in a single Saturday’s headline.