Expected Goals (xG)

Expected Goals — commonly abbreviated as xG — is a statistical model used in football analytics to quantify the quality of scoring opportunities. Rather than simply counting shots, xG assigns each shot a probability value between 0 and 1, representing the likelihood that the shot results in a goal under typical circumstances.

A value of 0.1 xG means there is a 10% chance the shot becomes a goal. A penalty kick, taken from 12 yards directly in front of goal, typically carries an xG of 0.76 — meaning it results in a goal roughly 76% of the time.

The xG model was developed by football statisticians to evaluate performance beyond the scoreline. A team can lose 1–0 while generating 2.4 xG to their opponent's 0.3 — suggesting the outcome was driven by luck rather than quality.

xG models are trained on vast datasets of historical shots. Each shot is evaluated on multiple variables:

Distance from goal — Shots from inside the six-yard box have much higher xG than long-range efforts.

Shot angle — Central positions provide better xG than wide angles.

Assist type — A headed shot from a cross has different baseline probability than a shot after a through-ball.

Game state — Some models factor in whether the shot came from open play, a set piece, or a counter-attack.

Body part — Header vs. foot significantly affects goal probability.

Modern xG models, built by companies like Opta, StatsBomb, and FBref, incorporate dozens of variables per shot to produce highly accurate probability estimates.

OddsWiki uses xGoals — our AI-enhanced interpretation of expected goals — as a core input to our match prediction engine. Unlike raw xG, which focuses purely on shot quality, OddsWiki's xGoals metric integrates:

• Rolling 10-match form data
• Home vs. away performance splits
• Opponent defensive strength (measured by our Defense Power metric)
• Predicted lineup and tactical shape

This means our xGoals figure for an upcoming match is forward-looking rather than purely historical — it predicts the number of quality chances each team is likely to create, not just what they've produced in the past.

Experienced bettors use xG in several ways:

1. Identify value in Over/Under markets. If a match has low expected goals from both teams' recent form, the "Under 2.5 Goals" market may offer better value than the bookmaker's odds imply.

2. Evaluate team performance beyond results. A team consistently generating 2.0+ xG per match but losing due to poor finishing is likely to improve — making their future matches higher value.

3. Spot overpriced favourites. A team on a winning streak but with an xG deficit (they're being outchanced) may be overvalued in the market.

4. Combine with OddsWiki predictions. Our AI football forecast shows xGoals for each upcoming match, allowing you to cross-reference our model output with bookmaker odds to identify potential value.

No model is perfect. Key limitations of xG include:

• Does not capture goalkeeper quality — Shot-stopping ability varies enormously between keepers.
• Small sample sizes — xG becomes reliable over 15–20 matches, not single games.
• Does not model chance creation probability — xG only measures shots that already happened; it can't tell you a team "should have" created more.
• Model variance — Different xG providers (Opta vs. StatsBomb) often produce different values for the same shot.

OddsWiki addresses these limitations by layering our Attack Power and Defense Power metrics on top of xG to create a more complete picture.