Expected Goals (xG)
Expected Goals (xG) is a football statistic that measures the quality of a goal-scoring chance on a scale from 0 to 1, based on shot location, shot type, body part, defensive pressure, and the type of pass that created it. A shot from the centre of the six-yard box might be worth 0.45 xG; a 30-yard strike, just 0.03.
How is xG calculated?
Every shot is fed into a model trained on hundreds of thousands of historical attempts with known outcomes. The model weighs distance from goal, angle to goal, whether the shot was a header or a foot strike, the type of assist (through-ball, cross, rebound), defender pressure, and game state. The output is the probability that an average player would score from that exact situation.
Sum the xG of every shot a team takes in a match and you get total xG — a far better indicator of attacking performance than the actual scoreline, which is heavily affected by finishing variance and goalkeeper form.
Why does xG matter for predictions?
Football is a low-scoring sport where individual finishes are wildly random. A team consistently generating 2.0 xG per match is creating chances at a title-winning level — even if their actual goal tally lags behind. Over a long enough sample, goals catch up to xG.
Final3rd uses xG, expected assists (xA), and expected goals against (xGA) as core inputs to every match prediction and player projection. Pages like our value bets and player props rely heavily on xG-based calibration.
xG vs actual goals — what does the gap tell you?
When goals exceed xG over a sample, a team is over-performing — either through elite finishing or short-term variance. When goals trail xG, the team is creating chances but not converting, and is usually due for positive regression. This gap (sometimes called G − xG) is one of the most actionable predictive signals in football betting.
Frequently asked questions
What is a good xG per match?
+
Top European clubs average 1.7–2.2 xG per match in league play. Anything above 2.5 sustained over a season is elite attacking output. Below 1.0 typically indicates a struggling attack.
Is xG accurate?
+
xG is a probability, not a prediction. A 0.7 xG shot is missed 30% of the time. Over hundreds of shots, total xG closely tracks actual goals — but for a single match, variance is large.
Who invented xG?
+
Modern xG models trace back to academic work by Sam Green (2012) and developments by analysts like Michael Caley and StatsBomb. Major data providers now publish their own xG models with slight methodology differences.