xA, or expected assists, measures the likelihood that a pass becomes a goal assist, based on the quality of the chance it creates for the player who receives it. It is essentially the xG value of the shot that follows, credited to the passer, used to judge chance creation regardless of whether the finish goes in.
xA grew out of the same data-tracking boom that produced xG in the 2010s. Companies such as Opta were already logging where every pass and shot happened on the pitch, so building a model for chance creation, not just chance taking, was a natural next step.
Kevin De Bruyne is the player analysts reach for most often. In the 2019-20 season he recorded 20 Premier League assists, level with Thierry Henry's record from 2002-03, and his underlying chance-creation numbers across his best Manchester City seasons ranked among the highest in the division year after year.
Like xG, xA is a guide rather than a verdict. It says a pass created a good chance, not that the pass itself was inventive or difficult to play, so the two numbers are best read together rather than as a final judgement on a player's creativity.