xG (Expected Goals)
xG, or « expected goals », measures the quality of a chance: the probability that a given shot ends up in the net, between 0 and 1, based on distance, angle, body part and type of play. A team's total xG over a match estimates how many goals it « should » have scored.
xG goes beyond the raw, often misleading scoreline: a team can win 1-0 on its only chance while the other piled up 2.5 xG. Over time, xG predicts future performance far better than goals scored, which depend heavily on short-term finishing.
Each data provider (Opta, StatsBomb, Understat…) trains its own model on hundreds of thousands of shots, so the figures differ slightly from one source to another. xG for (the chances a team creates) is distinguished from xGA, or xG against (the chances it allows). Penalties are often set aside (npxG = non-penalty xG), because they weigh heavily and have little to do with open play.
For a prediction, a team's recent xG is a steadier basis than its goals scored: it shows how many chances the side creates, regardless of how clinical it has been lately. It has limits, though: it ignores moves that never end in a shot, it depends on game state (a team that leads sits deeper and concedes more) and it needs several matches before it becomes reliable.
Example
A shot from the penalty spot is worth about 0.76 xG; a strike from 30 metres, less than 0.03 xG.
How Elofoot uses it
xG is the starting point of Elofoot's Poisson model: each team's expected goals begin from its actual average xG over recent matches (or, failing that, reference values of 1.45 at home and 1.15 away), then are adjusted for league position, form, division strength and home advantage. The model derives the 1X2 probabilities, the most likely score, BTTS and over/under from them.
Frequently asked questions
- How is the xG of a shot calculated?
- A statistical model compares the shot with thousands of similar ones (distance, angle, body part, type of play) and returns the share that ended up as goals. An xG of 0.10 means a shot of that kind is scored roughly once in ten attempts.
- What is the difference between xG and xGA?
- xG measures the chances a team creates; xGA (expected goals against) the chances it concedes. A strong side combines high xG with low xGA.
- Does the team with more xG always win?
- No. xG describes the quality of chances, not the result: in a single match, finishing, goalkeeping or luck can change everything. It becomes a good indicator over several matches.
- Why does xG differ between websites?
- Each provider uses its own model and its own data (goalkeeper position, defensive pressure…). The orders of magnitude agree, but the decimals vary.
See also
Go further
Put these concepts into practice with our AI football match prediction.
Predict a match for free