xG vs. xA: Unpacking Football's Key Performance Metrics
Expected Goals (xG) measures the quality of a shot, while Expected Assists (xA) measures the quality of a pass that leads to a shot. Together, they help evaluate chance creation and finishing skill.
Expected Goals (xG)
AnalystShort definition
A model that assigns a probability (from 0 to 1) to any given shot resulting in a goal, based on historical data of similar shots.
Primary use
Coaches use season-long xG totals to determine if their team's attacking structure is creating high-quality chances, or if a striker's goal drought is due to poor luck or poor positioning.
Expected Assists (xA)
AnalystShort definition
A model that assigns the xG value of a given shot to the player who made the final pass before the shot was taken.
Primary use
Scouts use xA to identify creative midfielders who may be underperforming on traditional assists due to their teammates' poor finishing, suggesting they could be more productive in a better team.
Side-by-side comparison
| Criterion | Expected Goals (xG) | Expected Assists (xA) |
|---|---|---|
| Definition | A model that assigns a probability (from 0 to 1) to any given shot resulting in a goal, based on historical data of similar shots. | A model that assigns the xG value of a given shot to the player who made the final pass before the shot was taken. |
| What it measures | The quality of a scoring chance. It evaluates the likelihood of a shot being scored, independent of the shooter's ability. | The quality of chance creation. It evaluates the likelihood that a pass will become an assist, independent of the receiver's finishing. |
| Data inputs | Shot distance/angle to goal, body part used (head/foot), type of attack (e.g., open play, set piece), pattern of play preceding the shot. | The data inputs for the resulting shot's xG (shot location, angle, etc.), thereby crediting the quality of the chance created by the pass. |
| Typical use | To evaluate a team's or player's long-term performance in creating and converting high-quality chances. Used to compare goals to chance quality. | To identify creative players who consistently generate high-quality shooting opportunities for teammates, even if official assist totals are low. |
| Strengths | Provides a more stable and objective measure of attacking performance over time than raw goal counts, which can be subject to luck. | Isolates the contribution of the passer from the skill of the finisher, rewarding creative players for making dangerous passes. |
| Weaknesses | Most public models do not account for the position of defenders or the goalkeeper, a key piece of contextual information. | Is entirely dependent on the resulting shot. A brilliant pass that a teammate controls poorly before shooting receives a low xA value. |
| Common misuses | Using a single match's xG total as definitive proof of which team 'deserved' to win, ignoring finishing skill and randomness. | Assuming a high xA player is always making brilliant passes; it could reflect a simple pass to a player who creates a great shot for themself. |
| Where it excels | Evaluating a striker's movement and shot selection over a season by seeing if they consistently get into high-probability scoring locations. | Spotting undervalued playmakers whose teammates are poor finishers, thereby hiding their true creative impact from traditional assist stats. |
Advantages · Expected Goals (xG)
- Quantifies the quality of a shot, not just its outcome.
- More stable indicator of attacking process than goals.
- Helps evaluate long-term finishing skill (goals vs. xG).
- Useful for assessing a player's positioning and shot selection.
Advantages · Expected Assists (xA)
- Isolates and credits the passer's creative contribution.
- Rewards dangerous passes even if the shot is missed.
- Identifies creative players whose assist totals may be misleading.
- Fairer measure of a playmaker's impact than raw assist counts.
Limitations · Expected Goals (xG)
- Basic models ignore defender and goalkeeper positioning.
- Does not capture the quality of the shot execution (e.g., power).
- Can be skewed by a single high-value chance like a penalty.
- Less reliable for single-game analysis than over a long period.
Limitations · Expected Assists (xA)
- Directly linked to the xG of the resulting shot, not the pass itself.
- Doesn't credit passes that break defensive lines but don't lead to a shot.
- Can be unfairly low if the shot-taker does something to worsen the chance.
- Can be inflated by a simple pass if the receiver creates a great shot.
When each is useful
Reach for Expected Goals (xG) when
Coaches use season-long xG totals to determine if their team's attacking structure is creating high-quality chances, or if a striker's goal drought is due to poor luck or poor positioning.
Reach for Expected Assists (xA) when
Scouts use xA to identify creative midfielders who may be underperforming on traditional assists due to their teammates' poor finishing, suggesting they could be more productive in a better team.
Real football examples
Expected Goals (xG) and Expected Assists (xA) help us understand performance beyond the scoreline. They measure the process, not just the result.
Elite Finishing vs. xG
Historically, players like Son Heung-min and Lionel Messi have consistently scored more goals than their xG models predict. This isn't just luck; it's a quantifiable sign of elite finishing talent. While an average player might score from a position with a 0.1 xG (a 10% chance) one out of ten times, these players might convert two or three times. Comparing a player's Goals to their xG over multiple seasons can help identify those with exceptional shooting ability.
The Creative Playmaker and xA
Kevin De Bruyne is a modern example of a player whose value is clearly shown by xA. He consistently ranks among the leaders in Expected Assists because he doesn't just pass the ball; he delivers it into areas where teammates have a very high probability of scoring. Even if his striker misses the shot, De Bruyne's xA figure still gets credited with the high quality of the chance he created, reflecting his creative contribution more fairly than the traditional assist metric would.
Analysing Team Dynamics
In some seasons, teams might demonstrate a disconnect between creativity and finishing. For instance, a team could have midfielders generating a high cumulative xA, proving they are creating good chances. However, if the team's total goals are much lower than their total xG, it points to a problem with the forwards' finishing. This tells analysts the creative foundation is strong, but the final piece of the attack is failing. It distinguishes between a team that can't create and a team that can't finish.
Frequently asked questions
?Can a player have a higher xG total than goals scored?
Yes, absolutely. This is common and indicates that a player has missed chances with a high probability of being scored, suggesting a run of poor finishing or bad luck.
?What is considered a good xG for a single shot?
A penalty kick is typically ~0.76 xG. A shot from inside the six-yard box could be >0.40 xG, while a shot from outside the penalty area is often <0.05 xG.
?Are xG and xA directly related?
Yes. The Expected Assists (xA) value of a pass is simply the Expected Goals (xG) value of the shot that is taken as a result of that pass.
?Why not just use goals and assists?
Goals and assists are outcomes that can be influenced by randomness. xG and xA measure the underlying process of creating chances, offering a more stable and often more predictive view of performance.
?Does a player receive an xA value if the shot is missed?
Yes. The xA is credited to the passer based on the quality of the chance they created, regardless of whether the shooter scores, misses, or has their shot saved.
?Can a shot have an xG of 1.0?
No. Since xG is a probability based on thousands of historical shots, no chance is considered 100% certain. Even an open-goal tap-in has a tiny, non-zero probability of being missed.
Continue Learning
Hand-picked next reads from Academy, Compare and Research Insights.