What Are Expected Goals in Football? A Practical Introduction to xG
August 12, 2026 16 Views
Expected goals (xG) estimates the probability that a shot will become a goal. A model…
What Is xG per Shot, and When Is It Useful?
Beyond shot totals 
Load More - Link
A busy shot chart can flatter an attack that never truly threatened.
Contents show
Picture a goalless match: one team takes 18 shots, mostly hurried efforts from 25 yards; the other takes six, including two close-range cutbacks. The score says they were equally successful, while the shot count makes the first team look dominant. Neither reveals which side created the better average opening.
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.
Goals are too rare and volatile to answer that question reliably—finishing, goalkeeping, and luck can decide a small sample. Raw shot totals have the opposite problem: they treat a hopeful blast and an unmarked tap-in as equivalent. xG per shot fills that gap by dividing total expected goals by total attempts, showing the estimated scoring probability of a typical chance.
Key terms
Two views of shooting danger
- Expected goals (xG)
A model-based estimate of how often a shot with similar characteristics would be expected to score. An xG value of 0.20 describes a class of chances—not a promise that this particular attempt has a fixed fate.
- Shot xG
The estimated scoring probability assigned to one attempt. The broader guide to expected goals from first principles explains the inputs models commonly consider.
- Total xG
The xG values of all shots added together. It measures cumulative chance creation, so many modest attempts can produce a high total.
- xG per shot
Total xG divided by the number of shots. It estimates average chance quality and helps distinguish frequent low-value shooting from fewer, clearer opportunities.
The same total can hide different attacks
Ten shots worth 0.10 xG each produce 1.0 total xG and 0.10 xG per shot. Four shots worth 0.25 each also produce 1.0 total xG, but average 0.25 per attempt.
The cumulative threat is equal; the route to it is not.
Worked example
Calculating xG per shot
- List the five attempts
Suppose their xG values are 0.04, 0.06, 0.08, 0.12, and 0.20.
- Add the xG values
The attempts total 0.50 xG: 0.04 + 0.06 + 0.08 + 0.12 + 0.20 = 0.50.
- Count the shots
There are five attempts in the sample.
- Divide total xG by shot count
The calculation is 0.50 ÷ 5 = 0.10 xG per shot.
- Interpret the result
The average attempt was valued at 0.10 xG—roughly the model’s expected return of one goal per ten similar shots, not a guarantee that every tenth shot will score.
Check what counts as a shot
The formula is simple, but its inputs are not always standardized. Providers may differ in whether they include penalties, blocked attempts, or shootout kicks—and in how those events receive xG values.
These choices can change both total xG and the shot denominator. Comparisons are safest when every figure comes from the same provider and uses the same competition and event rules.
Reading the 0-to-1 scale
From hopeful efforts to clear scoring chances
An xG value of 0.05 means that 100 comparable attempts would be expected to produce about five goals. It does not mean the next attempt is “five percent of a goal”; it is a probability estimate based on similar situations.
| Approximate xG | Typical example |
|---|---|
| 0.01–0.05 | Long-range efforts or shots from very tight angles |
| 0.08–0.15 | Promising box shots with pressure or an imperfect angle |
| 0.25+ | Close-range chances, clear cutbacks, or lightly contested one-on-ones |
| Around 0.75 | A penalty in many models |
Location provides much of the context. The effect of shot angle and distance explains why a central attempt near goal usually scores higher than one from the edge of the box. Defensive pressure, the type of assist, body part, and whether the goalkeeper is positioned can also affect the estimate.
For a team, 0.12 xG per shot suggests an average scoring probability of roughly 12% across its attempts. That average can still hide a mixture of speculative shots and excellent chances.
These ranges are guides, not universal benchmarks. Providers use different data and assumptions, while leagues can differ in shot selection and style. Comparisons are safest within the same model, competition, and time period.
Team diagnosis
What the average reveals
Chance quality makes sense only alongside chance quantity.
xG per shot works best as a diagnostic of shot selection. A high average may reflect patient buildup, cutbacks, close-range finishes, or a reluctance to shoot until space opens. A low average can point to frequent long-range attempts, rushed attacks, or difficulty entering dangerous areas.
Neither profile is automatically better. Consider two teams:
| Shots | xG per shot | Total xG |
|---|---|---|
| 12 | 0.08 | 0.96 |
| 6 | 0.16 | 0.96 |
The first creates twice as many attempts but from weaker positions; the second creates fewer, clearer chances. Their total expected threat is identical, reached through different attacking styles.
That trade-off is why xG per shot should always be read beside shot volume and total xG. A team averaging 0.20 xG per shot sounds dangerous, but two shots produce only 0.40 xG. Meanwhile, 15 attempts at 0.09 generate 1.35 xG despite the less impressive average.
Match context also matters. Scoreline, red cards, opponent strength, and small samples can reshape the number. High xG per shot therefore indicates efficient chance selection, not territorial control or dominance; those claims require broader evidence from volume, total xG, possession, and match flow.
Player analysis
What it says about a player
Chance quality often reflects role before finishing ability
A player’s xG per shot usually says more about where and how he shoots than how well he finishes. A penalty-box poacher receiving cutbacks and rebounds may post a high average because many attempts come from close range. A winger cutting inside or a midfielder shooting from 25 metres will often record a lower figure—even with excellent technique.
Role is therefore essential to interpretation. Tap-in specialists, set-piece targets, penalty takers, and long-range shooters should not be ranked as finishers from this number alone. Penalties can also distort small samples, so non-penalty xG per shot is often cleaner.
The average becomes more revealing when read alongside:
- Shots: shows whether the pattern rests on five attempts or fifty.
- Total xG: captures how much threat the player accumulates.
- Goals: provides the actual scoring return.
- Conversion rate: shows the share of attempts scored; football shot conversion rates add context but also fluctuate sharply.
A player scoring above xG may be finishing well, benefiting from variance, or both. Only a sufficiently large shot sample makes persistent differences more persuasive. Even then, xG per shot remains primarily a description of shot profile and attacking role, not a finishing grade.
Fair comparisons
Context before comparison
- Shot type and roleHeaders, cutbacks, through balls, and long-range attempts produce different averages. A poacher and a set-piece specialist should not be judged as though they receive the same chances.Look forSimilar roles and shot profilesAvoidDirect comparisons between unlike shooters
- Game stateScorelines change behaviour: trailing teams often shoot more urgently, while leading teams may attack open space against stretched opponents.Look forMinutes played at level, ahead, or behindAvoidTreating every match situation as equivalent
- Competition and providerLeague strength, playing style, and model design can shift recorded values. Providers may classify shots differently or assign slightly different probabilities.Look forOne competition, period, and data sourceAvoidMixing datasets without qualification
- Penalties and sample sizePenalties are unusually high-value chances, so excluding them better reflects repeatable open-play and set-piece creation. This makes non-penalty xG in team analysis especially useful.Look forNon-penalty figures across a season or rolling windowAvoidPenalty-inflated or one-match conclusions
Sample warning
One match can mislead
A single game may contain only a handful of shots, so one tap-in or speculative effort can swing the average sharply. Rolling windows or full-season samples smooth that noise; the shot count should always accompany the rate.
Common misconceptions
What xG per shot does not prove
Myth
A higher xG per shot is automatically better.
Reality
A high average can accompany too few attempts.
Why
Selective shooting may improve the average while reducing total threat. A lower figure can still belong to a productive attack that creates and takes many reasonable shots.
Myth
It measures finishing ability.
Reality
It describes opportunities before the outcome is known.
Why
Finishing concerns whether goals exceed the quality of chances over a substantial sample. xG per shot mainly shows the typical difficulty of the attempts taken.
Myth
The best tactics produce the highest average.
Reality
Different effective styles can produce different averages.
Why
Counterattacks, crosses, rebounds, long shots and sustained pressure create distinct shot profiles. The average alone cannot prove control, creativity or tactical superiority.
Myth
An xG value is simply a big-chance label in decimal form.
Reality
They are separate provider-defined measures.
Why
xG assigns every attempt a probability; big chances are categorised using a provider’s criteria. The distinction matters when examining how big chances compare with xG.
Blind spots
Some attacking value never reaches the metric
xG per shot omits possessions that end without an attempt. It cannot credit line-breaking passes, carries, decoy runs, pressing, space creation or dangerous deliveries that narrowly miss a teammate. Those contributions need event data, video and broader team context.
Practical checklist
A repeatable way to compare xG per shot
- Use one data source
Take shot counts and xG values from the same provider. Mixing sources can combine different models, event logs, and rules, making the result unreliable.
- Confirm what counts as a shot
Check whether the denominator includes blocked attempts, penalties, shootout kicks, or other unusual events. Apply exactly the same definition to every player or team.
- Make an explicit penalty choice
Either include penalties throughout or remove both penalty attempts and their xG. Excluding them is often more informative for open-play comparisons, but consistency matters more than the choice itself.
- Build like-for-like samples
Compare the same competitions, time periods, and match types where possible. For players, consider similar roles and require enough attempts to prevent one or two chances from dominating the average.
- Calculate and cross-check
Divide total xG by the number of included shots, then inspect total xG and shot volume beside the result. A team producing 0.18 xG per shot from five attempts created less accumulated threat than one averaging 0.12 from fifteen.
- Record the method
Keep the provider, date range, shot rules, penalty treatment, and minimum sample with the result. That short note makes later updates and comparisons reproducible.
If the underlying event data changes, the calculation should be rerun rather than patched with figures from another source.
Conclusion
- Report the denominator and penalty treatment alongside the figure.
- Treat striking differences from tiny samples as prompts for investigation, not firm conclusions.
xG per shot answers a narrow question: how good was the average attempt? It does not describe the full quality of an attack. Shot volume and total xG are still needed to show how often danger was created and how much threat accumulated.
