What Is xG in Soccer? Expected Goals Explained Simply
If you have watched soccer coverage in the last few years, you have almost certainly encountered the term xG. It shows up on broadcasts, in post-match analysis, on social media debates, and in the mouths of pundits who sometimes use it correctly and sometimes do not. Expected goals has become the most widely discussed advanced statistic in soccer, and for good reason: it fundamentally changes how we evaluate players, teams, and match outcomes.
But xG is also widely misunderstood. Some people treat it as a perfect measurement of who deserved to win. Others dismiss it entirely as meaningless academic noise. Neither view is correct. xG is a powerful analytical tool with specific strengths, clear limitations, and a proper way to interpret it.
This guide explains what xG actually measures, how it is calculated, what it tells you, what it does not tell you, and how to use it to understand the sport more deeply.
What xG Measures
At its core, expected goals (xG) measures the quality of a scoring chance. Every shot in soccer can be assigned a probability of resulting in a goal based on historical data. A penalty kick has an xG of roughly 0.76, meaning that historically about 76 percent of penalties are scored. A shot from 30 yards at a tight angle with a defender in the way might have an xG of 0.03, meaning that only about 3 percent of similar shots historically find the net.
When you add up the xG of every shot a team takes during a match, you get a total that represents how many goals the team would be expected to score given the quality of their chances. If a team's total xG for a match is 2.3, it means they created chances that, on average, would result in 2.3 goals.
xG does not predict the outcome of a specific shot. It tells you how likely a goal was based on thousands of similar shots taken in similar situations throughout the history of the sport. A shot with an xG of 0.10 will still be scored sometimes. A shot with an xG of 0.85 will still be missed sometimes. xG is about probabilities over large samples, not certainties about individual events.
How xG Is Calculated
The xG value for a given shot is generated by a statistical model trained on hundreds of thousands of historical shots. The model considers several factors that influence the probability of scoring.
Shot location. The most important factor. Shots taken from closer to the goal have dramatically higher xG values than shots taken from distance. The area directly in front of goal within the six-yard box produces the highest xG values.
Shot angle. The angle between the shooter and the goal line matters significantly. A shot from directly in front of goal has a wider target to aim at than a shot from a sharp angle near the sideline. Wider angles mean higher xG.
Body part. Shots taken with the foot have different conversion rates than headers. Headers are generally less accurate and powerful, so they tend to have lower xG values for equivalent positions on the pitch. Volleys, half-volleys, and other shot types are also factored in by more sophisticated models.
Game state. Whether a team is winning, losing, or drawing can influence shot-taking behavior and conversion rates. Some models incorporate this information, though its impact is relatively small compared to location and angle.
Assist type. How the chance was created affects the likelihood of scoring. A shot following a through ball that puts the attacker in behind the defense tends to have higher xG than a shot from a crossed ball, because through balls often create one-on-one situations with the goalkeeper. Cutbacks from the byline, set pieces, and open play all carry different baseline probabilities.
Speed of play. Some advanced models consider whether the shot was taken on a fast break, from a rebound, or after a sustained period of possession. Fast breaks often result in disorganized defenses, which increases scoring probability.
Goalkeeper position. The most advanced xG models, sometimes called post-shot xG or xGOT (expected goals on target), incorporate where the shot was placed relative to the goalkeeper. These models measure not just the quality of the chance but the quality of the shot itself.
Different data providers, including Opta, StatsBomb, Understat, and FBref, use slightly different models with different variables and training data. This is why you may see slightly different xG values for the same shot depending on the source. The differences are usually small and do not change the overall picture.
xG vs. Actual Goals: Understanding Over- and Underperformance
One of the most valuable applications of xG is comparing a team's or player's expected goals to their actual goals over a sustained period.
If a player consistently scores more goals than their xG suggests, they are outperforming their expected goals. This can mean the player is an exceptional finisher who converts chances at a rate above the historical average. It can also mean they are on a hot streak that will eventually regress.
If a player consistently scores fewer goals than their xG suggests, they are underperforming their expected goals. This might indicate poor finishing, bad luck, or a temporary slump.
The critical word here is "consistently." Over a single match, the difference between xG and actual goals is heavily influenced by randomness. A team can create 3.0 xG worth of chances and score zero goals. It happens. Over a full season, however, most teams and players regress toward their xG. The persistent outliers, players who outperform their xG season after season, are genuinely elite finishers.
Team-level xG comparison is one of the clearest indicators of underlying performance. A team that creates 1.8 xG per match but only scores 1.2 goals per match is probably due for positive regression, meaning they are likely to start scoring more even if nothing else changes. Conversely, a team scoring well above their xG may be due for a dry spell.
This is where xG becomes genuinely useful for prediction. A team sitting mid-table despite strong xG numbers is likely better than their current standing suggests. A team near the top of the table despite weak xG numbers may be riding unsustainable luck.
Using xG to Evaluate Players
xG provides a more nuanced evaluation of attackers than simply counting goals.
A striker who scores 15 goals from 18 xG worth of chances is a less impressive finisher than one who scores 15 goals from 12 xG worth of chances, even though their goal tallies are identical. The first player had more and better chances. The second player converted difficult chances at an exceptional rate.
For creative players, xA (expected assists) extends the concept by measuring the quality of chances a player creates for teammates. A midfielder who consistently creates high-xG chances for teammates is making a major contribution even if the strikers are not finishing those chances.
For goalkeepers, xG is essential. A goalkeeper who faces 40 xG worth of shots in a season but concedes only 32 goals has prevented 8 goals more than the average goalkeeper would have. This metric, often called goals prevented or PSxG (post-shot expected goals) minus goals allowed, is one of the best ways to evaluate goalkeeping performance independent of the defense in front of them.
When evaluating players, always look at xG over a full season or longer. Small sample sizes produce misleading results. A player who outperforms xG by a large margin over 10 matches might be talented or might be lucky. A player who outperforms xG consistently over multiple seasons is almost certainly talented.
Using xG to Evaluate Teams
At the team level, xG reveals tactical tendencies and underlying quality that the scoreboard often obscures.
xG difference (xGD), which is xG created minus xG conceded, is one of the strongest predictors of future league standings. Teams with strong xGD tend to be genuinely good teams, even if their current points tally does not reflect it. Teams with weak xGD despite a strong points tally are often overperforming and likely to decline.
The types of chances a team creates tell you about their playing style. A team that generates most of its xG from chances inside the six-yard box is likely playing a patient, possession-based style that works the ball into dangerous areas. A team that generates xG primarily from set pieces is more reliant on dead-ball situations. A team that takes many low-xG shots from distance is likely struggling to break down defenses.
xG against (xGA) measures the quality of chances a team concedes. A team with a low xGA is doing an excellent job of limiting opponents to low-quality chances, which reflects strong defensive organization regardless of whether those chances are being converted.
xG Maps and Visualizations
xG maps (also called shot maps) are visual representations of every shot in a match, plotted on the pitch with the size of the dot representing the xG value. Large dots represent high-quality chances. Small dots represent low-quality shots.
A single glance at an xG map often tells you more about a match than the scoreline does. If one team's map shows several large dots clustered inside the penalty area while the other team's map shows a scattering of tiny dots from distance, you can immediately see which team dominated in terms of chance creation, regardless of the final score.
xG maps are widely available on platforms like Understat, FBref, and various sports analytics accounts on social media. Learning to read them takes only a few minutes and adds significant depth to your understanding of match outcomes.
Non-Penalty xG (npxG)
Non-penalty expected goals (npxG) strips out penalties from the calculation. This matters because penalties are a somewhat arbitrary and inconsistent part of the game. A team that wins a lot of penalties will have inflated xG numbers that do not necessarily reflect their ability to create chances from open play.
When comparing players, npxG provides a cleaner measure of their open-play and set-piece finishing and creativity. A striker who scores 20 goals but 6 are penalties is performing differently from one who scores 20 goals entirely from open play, even if the headline number is the same.
npxG is particularly important when evaluating teams across leagues. Some leagues award penalties more frequently than others, and npxG normalizes for this variation.
Limitations and Criticisms
xG is a powerful tool, but it is not perfect. Understanding its limitations is essential for using it correctly.
xG does not capture everything about a shot. Standard xG models do not account for the specific skill of the shooter, the exact positioning of defenders (only whether or not the shot was contested), or the quality of the pass that set up the chance. Two shots from the same location can have vastly different likelihoods of scoring depending on context that the model does not capture.
xG is probabilistic, not deterministic. A match where one team has 3.0 xG and the other has 0.5 xG does not mean the first team deserved to win 3-0. It means that, over many repetitions of similar matches, the first team would win more often. But any individual match can deviate significantly from the expected outcome. This is part of what makes soccer compelling.
xG does not account for chances not created. A team that dominates possession but only creates shots from poor positions will have a low xG. A team that defends brilliantly by intercepting passes before they become shots will not be credited by xG for those interceptions. xG only measures what happens when a shot occurs, not the full picture of attacking and defensive play.
xG models differ between providers. Because different companies use different training data, different variables, and different algorithms, the same shot can have slightly different xG values depending on the source. This is usually not a major issue, but it means you should try to use consistent sources when making comparisons.
Small sample sizes are unreliable. xG for a single match, or even a handful of matches, should be interpreted cautiously. The metric becomes truly reliable over longer periods: half a season at minimum, a full season ideally. People who use a single match's xG to declare that one team definitively outplayed another are misapplying the tool.
Real Examples from Recent Seasons
xG regularly reveals stories that the scoreline hides.
In recent Premier League seasons, there have been numerous cases of teams sustaining winning runs despite xG data suggesting they were riding unsustainable luck. These teams often experience sharp declines in the second half of the season as their results regress toward what their underlying chance creation and prevention would predict.
Conversely, some teams have struggled early in the season despite creating excellent chances, only to surge up the table later as finishing variance corrected itself. Analysts who follow xG data can often identify these corrections before they appear in the standings.
At the player level, some of the game's best strikers, including prolific finishers in Europe's top leagues, consistently outperform their xG by significant margins, demonstrating that elite finishing skill is real and measurable. Meanwhile, players who dramatically outperform xG for a single season often fail to repeat the feat, confirming the role of variance in short-term results.
How to Start Using xG
If you want to incorporate xG into your own understanding of soccer, start with these steps:
- Follow FBref.com for free, comprehensive xG data on leagues, teams, and players worldwide
- Check Understat.com for match-level xG data, shot maps, and player xG statistics
- Compare xG to actual results after matches you watch. Notice when the scoreline tells a different story than the xG data
- Look at season-long xG data rather than individual matches to draw meaningful conclusions
- Use npxG when comparing players to get a cleaner picture of open-play performance
xG is not a replacement for watching the game. It is a complement that helps you see patterns that the eye misses and quantify impressions that would otherwise remain vague feelings. The most informed soccer fans combine what they see on the pitch with what the data reveals.
For a deeper dive into sports analytics, statistical modeling, and the data-driven methods transforming how we understand athletic performance, the Sports Analytics textbook covers the full spectrum of analytical approaches across major sports, available as a free, open-access resource.