What Is Expected Goals Against (xGA)?

Expected goals against, or xGA, is the total expected goals value of the shots a team concedes. Each opponent shot gets a scoring probability from an xG model, and summing them shows how many goals those chances would typically produce. It measures a team's chance prevention, not its goalkeeper, and is usually compared per 90 minutes.

How xGA is computed

Expected goals against is not a separate model. It is the same expected goals model applied to the shots a team faces rather than the shots it takes.

Every shot gets a probability of becoming a goal, between zero and one, estimated from large numbers of past shots with similar features: distance, angle, body part, type of assist, whether it came from open play or a set piece, and in richer models the positions of defenders and the goalkeeper. Sum those probabilities across every shot the opponent takes and you have the team's xGA for the match.

A worked example: an opponent takes three shots valued at 0.05, 0.12 and 0.40. The team's xGA for the match is 0.57. If one of those shots goes in, goals against is 1 and xGA is 0.57. If none does, goals against is 0 and xGA is still 0.57. xGA describes the chances conceded, not what happened to them.

Penalties and own goals

Penalties are usually assigned a fixed value from their historical conversion rate; Opta, for example, gives every penalty 0.79 xG. Because a single award adds a large chunk, many analysts quote non-penalty xGA when judging open-play defending. Own goals are not shots by the opponent, so they count in goals against but not in xGA.

Team defence, not goalkeeping

xGA is a measure of chance prevention. A low xGA means the team limits the number and quality of shots its opponent gets: fewer shots, from further out, under more pressure. That is the work of the whole side, including pressing, shape and the midfield screen, not only the defenders.

The goalkeeper barely affects pre-shot xGA, because most models value a shot at the moment it is struck, before the keeper does anything. That is why comparing a keeper's goals conceded with xGA is a weak test of shot-stopping. A shot struck straight at the keeper and one placed into the top corner carry the same pre-shot value.

For goalkeepers, analysts use post-shot expected goals, also called xGOT or PSxG, which values only shots on target and accounts for where the shot ended up. The usual form is PSxG minus goals allowed: positive values mean the keeper conceded fewer goals than the shots they faced would typically produce. FBref carried this figure until January 2026, when it removed its Opta-powered advanced stats. The comparison of xG and xGOT explains the difference in detail.

xGA per 90 and other rates

Season totals reward teams that have played more matches, so xGA is usually expressed as a rate.

Per match or per 90 for teams. Total xGA divided by matches, or by minutes divided by 90 when extra time is involved. A team allowing 1.0 xGA per 90 is conceding chances worth about one goal a game.

On-pitch xGA per 90 for players. The team's xGA while a given player was on the field, divided by that player's minutes over 90. It describes the defensive environment around a player rather than the player's individual contribution, and it is heavily shaped by teammates and opponents, so treat it as context.

xGA per shot. Total xGA divided by shots conceded. It separates two defensive styles: teams that allow many low-quality shots from distance, and teams that allow few shots but some very good ones.

Expected goal difference. xG minus xGA, per 90. Over a season it is one of the more reliable public indicators of underlying team quality.

Comparing xGA with goals against

The gap between goals against and xGA is the most quoted use of the metric and the most over-read.

When a team concedes far fewer goals than its xGA, the usual explanations are strong goalkeeping, opponents finishing below their usual standard, or plain variance. When it concedes far more, the reverse. Over a handful of matches the gap is mostly noise, because goals are rare events and a single deflection moves the total by a whole goal. Over a full season the gap shrinks relative to the totals, and what remains is more likely to reflect goalkeeping or a systematic feature of the defence that the model misses.

A practical reading rule: treat xGA as the better guide to how a defence will perform next, and goals against as the record of what happened. Check post-shot figures before crediting or blaming the goalkeeper, and remember that different providers' models produce different xGA for the same match, so compare like with like. The page on goals versus expected goals covers the general case.

xGA in hockey

Hockey uses the same idea with different inputs. Public models from MoneyPuck and Evolving-Hockey value unblocked shot attempts using distance, angle, shot type, the preceding event and the strength state. Team xGA is commonly quoted per 60 minutes at five-on-five, where most play happens and special teams do not distort the comparison.

One difference matters. In hockey, goaltender evaluation is commonly built directly on pre-shot xGA: goals saved above expected, or GSAx, is expected goals against minus goals actually allowed, as Evolving-Hockey's glossary defines it, and MoneyPuck describes a positive number as a goalie stopping more than an average goalie would. Soccer analysis leans on post-shot models for the same job. The guide to expected goals in hockey covers the model inputs.

Using xGA with results and odds

In modelling, xGA is most useful as a rolling input: a team's recent non-penalty xGA per 90, weighted toward recent matches and adjusted for opponent strength, predicts goals conceded better over short horizons than goals conceded does. Testing that claim on your own data needs match results and historical prices on the same keys. Parlay Ledger provides game results alongside real-time and historical odds in a unified format, so an xGA-based totals model can be checked against closing lines rather than against intuition. None of this is a betting recommendation; it is a way to measure whether a metric carries information the market has not already priced.

Frequently asked questions

What does xGA mean in football?
xGA stands for expected goals against. It adds up the expected goals value of every shot a team concedes, where each shot's value is the probability that a typical shot of that kind is scored. It measures the quantity and quality of chances a defence allows, regardless of how many actually went in.
Is a low xGA good?
Yes. A low xGA means a team concedes few chances or only poor ones, which indicates strong chance prevention. Compare it per 90 minutes rather than as a season total, check the non-penalty figure, and look at xGA per shot to see whether the team limits shot volume, shot quality or both.
Why is a team's goals against lower than its xGA?
Usually a mix of good goalkeeping, opponents finishing below their usual level, and variance. Over a few matches it is mostly variance, since one goal shifts the total sharply. Over a season, a persistent gap is more likely to reflect the goalkeeper, which post-shot expected goals can confirm or rule out.
Does xGA measure the goalkeeper?
Mostly not. Standard xGA values shots at the moment they are struck, before the goalkeeper acts, so it reflects the team's chance prevention. In soccer, goalkeeper evaluation uses post-shot expected goals on target. In hockey, goals saved above expected compares pre-shot xGA with goals allowed, and is widely used for goalies.
How do you calculate xGA per 90?
Divide total expected goals against by minutes played, then multiply by 90. For a team that played only full matches, that equals xGA divided by matches. For a player, use the team's xGA while that player was on the pitch, divided by the player's minutes and multiplied by 90.