The score says how a match ended. xG says how it should have ended. Here is what every metric on this site means, on today's numbers.
xG (expected goals) is the sum of the chances behind every shot a team took. Each shot gets a probability of becoming a goal from where and how it was struck: a shot from the six-yard box is worth about 0.4 goals, one from outside the box about 0.03. Add them up and you get xG. It is arithmetic over shots, not an opinion.
The same thing at your own end: the chances a team allowed its opponents to take. A low xGA means opponents only get bad positions — not that the keeper is saving everything.
Expected points: how many points a performance like that is worth on average. From the xG of both sides you get the probability of a win, a draw and a defeat, and from those the average points per match. Summed over a season, xPTS shows how many points the play deserved.
Δ = points − xPTS. A plus means results run ahead of performance: something went in more often than such chances usually allow. A minus is the opposite. Over a season the gap tends to close, which makes a big plus a regression warning and a big minus an upside signal.
Because football is a low-scoring game: two or three goals decide it and randomness matters. xG does not cancel the result, it tells you how much to trust it. One match proves nothing, ten start to.
From about ten matches xG predicts the future better than results do. Below that the numbers are noisy: we flag those clubs and say plainly that it is an early signal, not a verdict.
Right now in the RPL, Krasnodar is +7.4 points above expectation: results are ahead of the play. Fakel Voronezh has the opposite, −5.6. One is due a correction, the other is owed points if the play holds.