What is xA (Expected Assists) in football

xA (expected assists) rates how good a goalscoring chance a pass created. The model takes the xG of the shot that followed right after that pass and credits it to whoever played the ball. Set up a teammate for a shot worth 0.35 xG, and you’re credited with 0.35 xA — whether or not it goes in. Add up a team’s xA over a match and you get how many assists its passing quality deserved, not how many times a teammate finished well.

How it’s calculated: the value travels from the shot back to the pass

The starting point is always the xG of the specific shot. The model rates how dangerous the situation was: distance, angle, which part of the body, how many defenders stood in the way. That turns into a probability of scoring. The same number, already credited to the shooter, also travels to whoever played the ball right before the shot.

The first consequence follows from that: xA doesn’t exist without xG. It’s the same model, looked at from the other side of the move. Not “how dangerous was the shot” but “how good was the chance you handed someone”.

The second consequence is practical. A pass only counts if a shot follows it right away. A perfect ball that a teammate controls badly and never shoots from earns zero xA. That holds even if the pass was the best moment of the match.

How xA differs from a normal assist

A classic assist is binary. Either the shot goes in and the passer gets a point, or they get nothing. xA is continuous: it rates the quality of every pass before a shot, regardless of the outcome.

That difference cuts both ways. A midfielder can keep finding teammates in the box with clean passes. They can still go several matches without a single assist, because the forwards waste clear chances. The match report won’t show that. xA will: a high value next to zero assists in the table signals that the creation is there and the finishing is letting it down.

It works the other way too. A player can pick up an assist off a pass that took a lucky deflection and landed at a teammate’s feet. A classic assist logs that the same way it logs a defence-splitting through ball. xA tells the two apart, because it rates the quality of the chance created, not what came of it.

xA against xG: two halves of the same move

Worth reading side by side, because together they tell the full story of a goal.

xG answers one question: how good was the shot? xA answers another: how good was the pass that made it possible? A goal from a one-touch move scores high on both: good pass, good shot. A goal from a scrappy, individual move is different. A forward conjures a chance out of nothing, so xG can be high while xA stays low or at zero. The decisive moment was the shooter’s own doing, not the passer’s.

Comparing a team’s xG and xA in the same match shows where its goals come from: individual finishing quality, or collective work building the chance. Those are two different sources of an edge. Mixing them up leads to the wrong conclusion about who really decides matches for that side.

What xA doesn’t show

The move before the pass. A player sprints half the pitch to drag a defender out of position, then plays a simple ball into open space. That run gets nothing: only the final pass counts, not the work that set it up.

The second-to-last pass. Both the classic assist and xA count only the final touch before a shot. A player who opened up the move with a key pass three touches earlier is left out of the number, even though they often decided the numerical advantage.

Small samples. A handful of matches isn’t enough to tell stable creativity apart from a run of lucky passes that landed on a well-placed teammate. xA starts to mean something over a season, not over a single round.

How it enters a match analysis

Example of an analysis card. Figures are illustrative.

xA is useful exactly where the assist table stays silent: it shows who systematically creates good chances, independent of how sharp their attacking teammates are that month. In a Sharkline read, team and player statistics — including the quality of chances created — come from a sports data API alongside xG and form, the same sources covered in how to read football match stats. What comes out isn’t a table of raw numbers but a read: the probability of each outcome, a confidence percentage, and a rating of the data behind it, the way it’s laid out in how AI predicts a match result.

The full chain from raw data to a finished read is on how Sharkline works.

Frequently asked questions

What is xA in simple terms? It’s a rating of how good a goalscoring chance a pass created. The number comes from the xG of the shot that followed right after — the more dangerous the teammate’s shot, the higher the xA for whoever played the ball.

How does xA differ from a normal assist? An assist is a binary outcome: the shot goes in or it doesn’t. xA rates the quality of the pass regardless of whether the teammate finished well. You can post high xA and zero assists in the same match.

What’s the difference between xA and xG? xG rates the shot. xA rates the pass that made it possible. They’re two halves of the same move, calculated by the same model from two different angles.

Does high xA mean a player is good? It means they’re a good chance creator — someone who regularly finds teammates in good positions. It says nothing about their defending, their work off the ball, or moves that never ended in a shot.

Over how many matches is xA worth judging? Over a single match the number carries a wide margin of error. It starts to mean something over ten to twenty matches, once randomness stops deciding the outcome.