What is xG? Expected Goals Explained for Football Fans

Expected Goals — abbreviated as xG — is one of the most widely discussed statistics in modern football. First developed in academic sports analytics research and brought to mainstream attention through platforms like Opta, StatsBomb and Understat, xG has moved from niche data tool to broadcast staple. Sky Sports, the BBC and most major football media outlets now display xG figures as a routine part of match coverage. But what does the number actually mean?

What xG Measures

xG measures the quality of a goal-scoring chance, expressed as a probability between 0 and 1. A shot with an xG of 0.10 is estimated to result in a goal 10% of the time, based on historical data from comparable shots. A shot with an xG of 0.50 would be expected to result in a goal roughly half the time. An xG of 1.0 is theoretically certain to result in a goal — a near-impossible standard rarely reached in practice.

The model assigns an xG value to each shot based on factors including:

• Shot location: Distance from goal, and angle relative to the centre of the goal — shots from close range and central positions have higher xG than those from distance or tight angles
• Assist type: Whether the shot followed a cross, a through-ball, a corner, a free kick, or a direct dribble
• Body part used: Headed shots have a lower xG than shots with the dominant foot, which are in turn generally higher than shots with the weaker foot
• Whether it was a direct free kick: These typically carry a low xG due to the wall and goalkeeper positioning
• Speed of attack and number of defenders between shooter and goal (in advanced models)

Each model provider (StatsBomb, Opta, Understat etc.) uses slightly different methodologies and inputs, which is why xG figures for the same shot can differ between platforms.

Penalty xG

A penalty kick carries an xG of approximately 0.76 — meaning that, based on historical data, a penalty is converted into a goal roughly 76% of the time. This figure varies slightly between models but is broadly consistent. The value reflects the fact that while penalties are high-quality opportunities, they are not certain goals — goalkeepers save roughly one in four.

xGA: Expected Goals Against

xGA (Expected Goals Against) is the same metric applied defensively. It represents the total quality of chances a team or goalkeeper has conceded, calculated by summing the xG of every shot they faced. A goalkeeper who concedes fewer goals than their xGA is outperforming expectations — they are saving shots they would statistically be expected to concede. Conversely, a goalkeeper allowing more goals than their xGA is underperforming relative to the quality of chances they have faced.

This makes xGA a useful way of separating goalkeeper performance from defensive quality. A goalkeeper can have a high xGA not because they are poor, but because their defence is generating many high-quality chances against them.

xGOT: Expected Goals on Target

xGOT (Expected Goals on Target) is a refinement that considers only shots on target and attempts to model goalkeeper performance more precisely. Rather than asking “what is the probability of this shot being a goal regardless of where it goes?,” xGOT asks “given that the shot was on target, what was the probability of it being a goal based on where it was placed within the goal?” This attempts to account for the goalkeeper's positioning and reaction as well as the quality of the shot placement.

Limitations of xG

xG is a powerful descriptive statistic but it has genuine limitations:

• Goalkeeper quality: Standard xG models do not fully account for individual goalkeeper quality — the same shot will have a different likelihood of resulting in a goal depending on whether it is faced by one of the Premier League's elite goalkeepers or a reserve player.
• Shooter quality: Similarly, a 0.15 xG chance for an elite finisher may have a real-world conversion rate significantly higher than 15%, while the same chance for a less technically gifted player might be lower. xG treats all shooters as equal in its base models.
• It describes, not prescribes: A team that consistently underperforms their xG (scoring fewer goals than their xG suggests) may be doing so because they have poor finishers, or simply because they are experiencing an unusual run of bad luck. xG can tell you which of these is more likely over a large sample, but cannot definitively separate the two in the short term.

How Clubs Use xG

Football clubs — particularly those with sophisticated analytics departments — use xG as part of recruitment analysis (is a striker outperforming or underperforming their xG over multiple seasons?), match preparation (which zones of the pitch do opponents allow high-xG chances from?), and post-match evaluation (did we create high-quality chances, or just a high volume of low-quality shots?). Tools used by analysts and fans alike include FBref.com (which uses StatsBomb data), Understat, and Sofascore, all of which provide xG data for Premier League and other top-flight matches.