Expected goals assigns every shot a probability of scoring based on historical outcomes of similar chances — distance, angle, body part, assist type. Summing them gives a team's xG for a match or season: a measure of chance quality that strips out finishing luck. A team winning 1–0 with 0.4 xG against 2.3 rode its goalkeeper, whatever the table says.
For bettors, xG is a form signal that stabilises faster than results: teams massively over-performing their xG tend to regress, and the market sometimes lags that regression. Its limits are equally important — it ignores game state and shot context beyond its model, differs between providers, and one match of xG is nearly as noisy as one match of goals. Use it as an input, never a verdict.
Over the last six matches a side scored 11 from 6.1 xG. Betting markets pricing them off those 11 goals may be overrating a finishing hot streak that xG says is running ~80% above sustainable rate.