Expected goals against (xGA)
The total xG value of the chances a team concedes. It estimates the quality and volume of opportunities allowed, not the number of goals actually conceded.
What Are Expected Goals in Football? A Practical Introduction to xG
Two versions of a match 
Load More - Link
Sometimes the losing side creates every chance that lingers in the memory.
Contents show
It is the 89th minute. One team has hit the post, missed a one-on-one and forced three sharp saves. The other has scored with a deflected shot. The whistle goes: 0–1.
Our Top Sportsbook Bonuses for August 2026
18+ only. Full terms apply.
250% Up To $1,000 Deposit Bonus
18+ only. Full terms apply.
18+ only. Full terms apply.
Then the match statistics appear: expected goals, perhaps 2.1–0.4, favour the defeated side. That can feel irritating. The result is indisputable—goals decide matches—yet the xG figure suggests the more threatening team “won” in another sense. It is not an attempt to rewrite the score. It offers a separate account of the game, based on the quality of the shooting chances rather than the number that happened to go in. The scoreboard records the outcome; xG helps describe the opportunities behind it.
Core concept
What xG actually means
A probability attached to each attempt
Expected goals (xG) estimates the probability that a shot will become a goal. A model compares the attempt with thousands of previous shots, considering details such as distance, angle, body part, assist type, and whether the chance was a header or one-on-one. Different models use slightly different inputs, so their figures may vary.
A shot worth 0.30 xG has an estimated 30% chance of being scored. Put another way, roughly 30 goals would be expected from 100 similar attempts. It does not mean 0.30 of a goal was scored, nor that the player was required to finish the chance.
Shot values are added together to create totals. If a team takes attempts worth 0.30, 0.20, and 0.05 xG, its total is 0.55 xG. The same addition produces a player’s match, season, or career xG.
These totals describe the quality and volume of chances, not an alternative result. Goals remain whole, decisive events, while xG expresses likelihood across repeated situations. That distinction explains why xG can differ sharply from the final score: unlikely shots sometimes go in, and excellent chances are sometimes missed.
Inside the model
How an xG model judges a shot
Historical attempts turn shot context into a probability.
An xG model learns from a large archive of shots whose outcomes are already known. It identifies which combinations of circumstances have historically produced goals, then assigns a new attempt a probability based on comparable situations.
Important inputs may include:
- Distance and angle: Central, close-range shots generally score more often than distant or narrow-angle attempts.
- Body part: Headers and weaker-foot attempts can have different conversion patterns from strong-foot shots.
- Assist type: A cut-back, through ball or cross changes the kind of chance being taken.
- Phase of play: Open play, corners, free kicks, counterattacks and penalties create distinct shot conditions.
- Goalkeeper position and defensive pressure: Where the goalkeeper and defenders are can determine how much of the goal is realistically available.
Not every provider captures all these details. Event-data models mainly use recorded actions and locations: where the shot occurred, how it was assisted and what happened immediately beforehand. They are practical across many competitions but may simplify pressure or positioning.
Tracking-based models use player and ball coordinates sampled throughout the move. They can account more directly for nearby defenders, goalkeeper location and open shooting lanes, but require richer data and different modelling choices.
Consequently, two reputable providers may give the same shot slightly different xG values. The difference usually reflects available data, definitions and model design—not necessarily an error.
Worked example
How three shots become 0.80 xG
- First shot: 0.05 xG
A speculative effort from a difficult position contributes only a small amount to the total.
- Second shot: 0.15 xG
A more promising attempt raises the running total to 0.20 xG.
- Third shot: 0.60 xG
This is a strong chance, though still generally less valuable than a penalty. For comparison, the xG value assigned to a penalty is usually around 0.76–0.79, with small differences between providers.
- Total: 0.80 xG
Adding the attempts gives 0.05 + 0.15 + 0.60 = 0.80 xG. That represents the combined quality of the three chances, not a fraction of a goal actually scored.
The total does not determine the score
The same three attempts could produce very different outcomes:
Zero goals: all three shots are missed or saved. One goal: one attempt is converted and the other two fail. Two or three goals: several chances are finished successfully.In every case, the shot total remains 0.80 xG because xG rates the opportunities rather than their eventual outcomes. Over many similar sets of chances, scoring should average closer to the expected figure; in a single match, finishing and chance can create a much wider result.
Reading the numbers
What each xG view reveals
- Team totals
The match total shows the overall balance of chances, but not how they were created. The same 1.5 xG could come from one excellent chance or many weak attempts.
- Individual shots
Shot values reveal which attempts drove the total. A single 0.70 chance tells a different story from seven 0.10 efforts.
- Average xG per shot
Dividing total xG by shot count separates quality from volume. Twenty low-value shots may produce more xG overall than eight strong shots while still having a lower average.
- Shot maps
Locations and marker sizes expose patterns hidden by totals, such as repeated close-range attempts or hopeful shooting from distance.
- Timelines
A cumulative line shows when xG was added. Sharp jumps mark valuable chances; long flat sections indicate no shots, not necessarily no pressure.
Avoid two common shortcuts
A rising cumulative line does not prove continuous dominance: it only rises when shots occur. “Big chance” is also a provider-defined label, not a universal xG threshold, so the two measures are not interchangeable.
Patterns matter more than one result
xG becomes more informative when the same tendencies keep appearing.
A single match can be distorted by a deflection, a remarkable save, an early red card, or one excellent finish. Even a convincing xG advantage may therefore say more about the chances created that day than the teams’ underlying quality.
Larger samples reduce the influence of these one-off events. Over several matches, attention can shift to repeated questions:
- Is the team consistently creating more xG than it concedes?
- Does its attack produce clear central chances or mostly difficult shots?
- Are defensive problems recurring against particular opponents or game states?
- Does the pattern survive changes in venue, lineup, and schedule strength?
There is no universal match total at which xG suddenly becomes reliable. More data generally helps, but squad changes, injuries, tactical adjustments, and unequal fixtures can make an older sample less relevant. Rolling windows are useful, provided the matches behind the total are still examined.
Persistent overperformance deserves investigation rather than an automatic verdict. A striker repeatedly scoring above xG might possess unusual finishing skill, receive chances that the model describes imperfectly, or simply be enjoying a run that will fade. Shot placement, penalties, minutes played, and the size of the gap all add context. The sensible conclusion is not “luck” or “proof of skill,” but a pattern worth testing against more evidence.
Common misconceptions
What xG can—and cannot—prove
Myth
The team with higher xG deserved to win.
Reality
Higher xG indicates better estimated chances, not entitlement to the result.
Why
Finishing, goalkeeping, game state and events outside shots still matter. xG describes chance creation more reliably than sporting merit.
Myth
Equal xG totals mean two matches were alike.
Reality
The same total can come from many weak attempts or a few excellent chances.
Why
Providers may also value shots differently; model disagreement reflects different data and assumptions, not necessarily an error.
Myth
One xG source gives the objective answer—and a betting edge.
Reality
Every source is an estimate, while betting markets already absorb extensive information.
Why
The question of whether xG adds useful betting insight requires testing beyond headline totals. Source choice should instead match the task: consistency, transparent methods, update speed, cost and coverage all count.
Step List
- Start with the score
Record the result and scoring sequence first.
- Compare xG from one source
Note the gap; never mix providers.
- Inspect the largest chances
Check whether one penalty or opening dominates.
- Separate volume from quality
Compare shot count, average value, and locations.
- Add match context
Allow for red cards, score effects, and late chasing.
Conclusion
Use model-qualified phrasing: “this model suggests,” not “the match proves.” xG gains meaning when one source is followed across many matches and its totals are checked against chance shape and context.
