Expected Goals Beyond the Basics: How to Read xG in Real Football Analysis

📅 Published: July 29, 2026 • 7 min read 📎 ← Back to all articles

Written by Fabiaana

Football statistics become useful when they help us ask better questions.

Expected Goals, or xG, is one of the clearest examples of this principle. In our previous guide, “Expected Goals (xG) in Football: A Complete Guide to Understanding Chance Quality,” written by Jose Sixpenze, we introduced the foundations of the metric, including how individual shot probabilities are assigned, how total xG is calculated, and why different providers can produce different values.

This article takes a different direction.

Rather than repeating how xG is calculated, we will focus on how to interpret xG when looking at real football performances.

A number such as 1.72 xG is useful, but the number alone does not tell the whole story. Two teams can produce exactly the same xG through completely different attacking patterns. A team can repeatedly generate strong chances without scoring, while another can produce a similar xG through a smaller number of opportunities.

Understanding those differences is where xG becomes much more interesting.

From Measuring Chances to Understanding Performance

As Jose Sixpenze explained in our introductory xG guide, expected goals gives analysts a way to distinguish between different qualities of scoring opportunities.

That distinction is important, but it raises a second question:

What should we actually do with the information?

Suppose Team A finishes a match with:

1.80 xG

Team B also finishes with:

1.80 xG

At first glance, the two attacking performances appear identical.

They are not necessarily identical at all.

Team A might have produced:

• 9 shots
• 4 shots inside the penalty area
• 3 major chances
• several dangerous attacks

Team B might have produced:

• 18 shots
• mostly from outside the penalty area
• very few central opportunities
• many attempts after losing possession

The total xG may be similar, but the route to that xG can be very different.

This is one of the most important ways to move beyond simply reading the number displayed beside a team's name.

xG Is Better Viewed as a Distribution, Not Just a Total

One of the easiest mistakes when reading xG is to focus exclusively on the final number.

A total such as 2.10 xG combines the probabilities of many individual opportunities.

But the distribution of those opportunities can tell us much more.

Imagine two hypothetical performances.

Team A

Team B

The totals could be similar, yet the attacking profiles are clearly different.

Team A generated several strong opportunities.

Team B relied heavily on one particularly valuable chance while producing many smaller opportunities.

This matters when analysts investigate how sustainable a team's chance creation may be.

High xG Does Not Always Mean the Same Thing

A high xG total is generally positive evidence of chance creation, but analysts should still investigate its source.

Consider three possible routes to 2.00 xG.

Scenario 1: Consistent chance creation

The team creates several high-quality opportunities throughout the match.

This suggests sustained attacking pressure.

Scenario 2: One dominant opportunity

The team creates one extremely valuable chance and several minor opportunities.

The total remains high, but the performance was less evenly distributed.

Scenario 3: A penalty dominates the number

A penalty can contribute a substantial amount of xG because penalties are historically converted at a relatively high rate.

Therefore, the team's overall xG can look significantly stronger even though its open-play chance creation was more modest.

This is one reason why analysts sometimes separate penalty and non-penalty contributions rather than treating every xG point as identical.

The Difference Between Chance Volume and Chance Quality

Football analysis often involves a tension between how many chances a team creates and how good those chances are.

A team taking 20 shots is not automatically creating a better attacking performance than a team taking eight.

The eight shots could contain several excellent opportunities.

This is where xG adds information that shot totals alone cannot provide.

Consider:

Team A:
20 shots → 1.10 xG

Team B:
8 shots → 1.80 xG

Team A produced much greater shot volume.

Team B produced the stronger expected scoring output.

Neither statistic should simply replace the other.

Instead, the combination can reveal something about the team's attacking approach.

A team with high shot volume but low xG may be generating many attempts without consistently reaching dangerous positions.

A team with fewer shots but high xG may be creating fewer but more valuable opportunities.

xG Can Help Identify Different Attacking Styles

This is one of the areas where xG becomes particularly useful for tactical analysis.

Two teams can have similar attacking output while reaching that output through very different methods.

For example, one team may prefer:

• controlled possession;
• patient build-up;
• combinations around the penalty area;
• cutbacks;
• central attacks.

Another may rely more heavily on:

• counter-attacks;
• direct passes;
• transitions;
• crosses;
• second balls.

Their xG totals may eventually look similar.

But the underlying attacking processes are different.

This is why xG should not be viewed as a replacement for watching football.

It is better understood as another layer of evidence.

Why Match Context Matters

A raw xG number does not tell us everything about the circumstances in which the chances were created.

Match state can have a major influence on football behaviour.

For example, imagine a team takes an early lead.

The team may subsequently defend deeper.

The opponent may increase possession and take more shots.

The resulting statistics could show:

• higher possession for the losing team;
• more shots for the losing team;
• higher late-match xG for the losing team.

Yet the statistics need to be interpreted in the context of the scoreline.

The team that was behind had a reason to attack.

The team that was leading had a reason to protect its advantage.

Therefore, analysts should ask not only:

“How much xG did each team produce?”

but also:

“When and under what circumstances was that xG produced?”

First-Half and Second-Half xG Can Tell Different Stories

Breaking a match into periods can reveal changes that the final total hides.

Suppose:

First half

Team A: 0.20 xG
Team B: 1.10 xG

Second half

Team A: 1.40 xG
Team B: 0.30 xG

The full-match numbers might suggest a relatively balanced contest.

But the match itself changed dramatically.

Team B started strongly.

Team A then adjusted and became much more dangerous after the interval.

This type of analysis can be especially useful when investigating:

• tactical changes;
• substitutions;
• fatigue;
• game state;
• changes in defensive structure;
• changes in pressing intensity.

The final xG number is therefore only one part of the story.

xG Difference Is Often More Informative Than xG Alone

When evaluating a team, it is useful to consider the relationship between the chances it creates and the chances it allows.

A simple conceptual measure is:

xG Difference = xG For − xG Against

For example:

2.00 xG For − 0.80 xG Against = +1.20 xG Difference

This suggests that the team generated substantially more expected scoring value than its opponent.

A negative figure would indicate the opposite.

Over one match, however, this number should be treated cautiously.

A single match can contain unusual events.

Over a larger sample, the relationship between expected goals created and conceded can become much more informative about the underlying performance of a team.

Research has found xG to be a useful indicator when evaluating future team performance, although the quality and design of the underlying model remain important.

What Happens When Goals and xG Disagree?

This is perhaps one of the most interesting applications of xG.

Imagine a team has:

8 goals

from:

13.5 xG

There is a substantial difference between the two numbers.

Another team might have:

13 goals

from:

8.5 xG

These situations deserve investigation.

But they should not immediately be labelled as proof of “luck”.

The difference can arise from several factors, including:

• finishing outcomes;
• goalkeeper performance;
• shot placement;
• defensive actions;
• penalties;
• unusual match events;
• model limitations;
• small sample size.

xG describes expected outcomes based on the model's information. Actual goals remain the observed result.

The important analytical question is therefore not:

“Which number is correct?”

It is:

“Why are the numbers different?”

xG Overperformance Should Be Interpreted Carefully

Suppose a striker has:

12 goals

from:

7.5 xG

It is tempting to conclude immediately that the player is an elite finisher.

That conclusion may eventually be justified.

But the sample matters.

A player can temporarily outperform expected goals because of:

• unusually accurate finishing;
• exceptional shot placement;
• favourable rebounds;
• penalties;
• goalkeeper errors;
• deflections;
• a small number of unusually successful attempts.

The longer the sample becomes, the more useful the comparison can become.

This is why analysts should avoid making strong conclusions from a handful of matches.

xG Is Not a Player Rating

Another important distinction is often overlooked.

A player's xG is primarily related to the quality and quantity of scoring opportunities associated with that player.

It does not automatically measure:

• passing;
• ball progression;
• defensive contribution;
• pressing;
• chance creation;
• movement without the ball;
• decision-making;
• overall tactical value.

A midfielder can have low xG while making an enormous contribution to his team's attacking structure.

Likewise, a striker can have high xG because teammates consistently create excellent opportunities for him.

Therefore:

High xG does not automatically mean “better player.”

It means something much more specific.

Who Creates the xG Matters

This leads to another useful question.

If a striker records 10 xG, how did those chances arrive?

Were they created through:

• through balls;
• crosses;
• cutbacks;
• individual dribbles;
• rebounds;
• set pieces;
• defensive errors;
• transitions?

The same xG total can emerge from very different attacking environments.

For player evaluation, therefore, xG becomes much more useful when combined with information about chance creation and involvement in the attacking sequence.

This is one reason modern football analytics increasingly looks beyond isolated events.

The Difference Between Pre-Shot and Post-Shot Information

Another important development in football analytics is the distinction between evaluating a chance before and after the shot.

Traditional xG is generally concerned with the probability that a shot becomes a goal based on characteristics of the opportunity.

But once the shot has been taken, additional information becomes available.

For example:

• where the shot was placed;
• how difficult it was for the goalkeeper;
• whether it was on target;
• whether it was saved;
• whether it was blocked.

This is the territory of post-shot expected goals, commonly known as PSxG.

PSxG can therefore provide another perspective on finishing and goalkeeping that standard xG alone does not fully capture.

This is a useful example of how football analytics has continued to evolve beyond the original xG concept.

xG and Goalkeepers

The same principle applies to goalkeepers.

Suppose two goalkeepers face shots with similar xG values.

Their actual outcomes can still differ considerably.

One goalkeeper may concede more goals than the expected value of the shots faced.

Another may concede fewer.

Post-shot models can help analysts investigate the quality of the shots actually directed toward the goal rather than simply evaluating the chances before the shot was struck.

This can help separate:

• chance quality
• from shot execution and goalkeeping outcome.

The distinction is important because a goalkeeper cannot control the quality of every chance his defence allows, but he can influence what happens after the shot is taken.

Why xG Should Be Combined With Other Metrics

Expected Goals is powerful precisely because it answers a specific question.

But football contains many questions.

For example:

xG: How valuable were the scoring opportunities?

Shots: How frequently did the team attempt to shoot?

Possession: How much of the ball did the team control?

Progression metrics: How effectively did the team move the ball toward dangerous areas?

Expected Assists: How valuable were the opportunities created for teammates?

Defensive metrics: How effectively did the team prevent or disrupt opposition attacks?

The strongest analysis combines relevant measurements rather than expecting one statistic to explain the entire game.

A Team Can Improve Without Immediately Scoring More

One of the most useful ways to interpret xG is to distinguish between process and outcome.

Imagine a team previously averaged:

0.95 xG per match

and now averages:

1.65 xG per match

but its actual goals have remained almost unchanged.

It would be easy to conclude that the team has not improved.

That may be premature.

The team may now be creating significantly better opportunities.

Finishing outcomes can fluctuate in the short term.

This does not guarantee that goals will increase later. It simply means that the underlying attacking process deserves further investigation.

This is one of the reasons analysts use xG to look beyond the scoreboard.

But xG Should Not Be Used to Rewrite History

There is an important limit here.

If a team loses 3–0, the result remains 3–0.

xG does not change the score.

It can provide additional information about how the match unfolded, but it should not be used as a way of pretending that the actual result did not happen.

A team that repeatedly creates excellent chances but fails to score has a real finishing problem if the pattern persists.

Likewise, a team that repeatedly wins despite generating weak chances should not automatically be declared fraudulent.

The job of analysis is to investigate the evidence, not replace reality with a preferred statistic.

Why One Match Is Usually Not Enough

This is one of the most important practical rules when interpreting xG.

A single match contains too much randomness to support strong conclusions about long-term team quality.

A team can produce:

2.5 xG

and score:

0 goals

in one match.

Another can produce:

0.6 xG

and score:

2 goals.

Neither result is impossible.

The analytical value increases when observations are accumulated over many matches.

Instead of asking:

“Why did this team have 2.5 xG today?”

analysts can ask:

“Has this team consistently generated high-quality chances over the last 10, 15 or 20 matches?”

That is a much stronger question.

Look for Trends, Not Isolated Numbers

Suppose a team records the following xG totals:

The progression is more interesting than any individual number.

It may suggest that the team is increasingly generating dangerous chances.

But even here, context matters.

The opponents may have varied considerably.

Therefore, the strongest interpretation combines:

xG trend + opponent quality + match context + actual performance.

Home and Away Context Can Change Interpretation

A team's xG profile may also change depending on venue.

Some teams are much more aggressive at home.

Others become more conservative away from home.

Therefore, comparing a team's overall xG without considering location can hide meaningful differences.

A useful analysis might compare:

• home xG;
• away xG;
• home xGA;
• away xGA;
• home shot quality;
• away shot quality.

This can reveal whether a team's attacking identity changes depending on where the match is played.

Opponent Strength Matters

Suppose Team A produces 1.80 xG against a league leader.

Team B produces 2.10 xG against the team at the bottom of the table.

It would be simplistic to conclude that Team B necessarily produced the better performance.

The opposition matters.

This is one reason why xG should be interpreted within the broader competitive context.

A team's underlying numbers become more informative when analysts consider who those numbers were produced against.

Why Different xG Providers Can Disagree

Jose Sixpenze's original article already explained that different providers can assign different xG values to the same shot.

There is an additional point worth understanding.

The disagreement is not necessarily evidence that one provider is “wrong”.

Different models may use different:

• datasets;
• definitions;
• variables;
• training periods;
• modelling techniques;
• contextual information.

Academic research has demonstrated that xG models can be built using a range of statistical and machine-learning techniques, and that adding additional information can affect model performance.

Therefore, when comparing xG from two providers, the correct approach is not simply to ask:

“Which number is higher?”

The better question is:

“How was each number produced?”

Model Quality Matters

An xG number looks precise.

For example:

0.37 xG

But the apparent precision can be misleading.

The number is an output from a model.

Behind that output are assumptions about the relationship between historical shots and future outcomes.

A sophisticated model can incorporate more contextual information than a simple model, but greater complexity does not automatically guarantee better performance.

Researchers have tested different approaches including logistic regression and machine-learning methods, with model performance varying according to the data and features used.

This is another reason why xG should be treated as model-based evidence, not objective truth.

A Better Way to Read an xG Match Report

Instead of looking at only:

Team A 1.65 xG
Team B 0.82 xG

try asking a sequence of questions.

Question 1: Who created more?
Look at total xG.

Question 2: Who created better individual opportunities?
Look at the distribution of chances.

Question 3: How many shots produced that xG?
Compare xG with shot volume.

Question 4: Where did the chances come from?
Look at shot locations and attacking sequences.

Question 5: When were the chances created?
Consider match periods and game state.

Question 6: Were penalties involved?
Separate penalty and non-penalty contributions when relevant.

Question 7: Is this performance consistent?
Compare multiple matches.

Question 8: Who were the opponents?
Consider opposition quality.

Question 9: What happened compared with actual goals?
Investigate the difference rather than automatically calling it luck.

This produces a much more useful interpretation than simply declaring the team with the higher xG “deserved to win”.

xG and Football Prediction Are Not the Same Thing

Another distinction worth making is between analysis and prediction.

A team with a strong long-term xG profile may provide useful evidence when assessing its future performance.

But xG itself does not automatically produce a final score prediction.

A football match remains uncertain.

Even a team consistently producing more xG than its opponents can lose individual matches.

Research has found xG to be useful for forecasting future team performance compared with simpler traditional measures, but that does not eliminate uncertainty from individual matches.

This distinction is especially important when statistical information is used in football prediction.

The Most Useful xG Questions

Instead of asking:

“What is the xG?”

try asking:

“Where did the xG come from?”
This reveals the attacking mechanism.

“How consistent is it?”
This helps distinguish a trend from an isolated match.

“Against whom was it produced?”
This adds competitive context.

“How much came from penalties?”
This helps separate different types of scoring opportunities.

“How does it compare with actual goals?”
This identifies areas requiring investigation.

“What happened before and after?”
This introduces time and tactical context.

“Does another metric support the same conclusion?”
This prevents over-reliance on a single statistic.

These questions transform xG from a number into an analytical framework.

A Practical Example

Imagine Team Red has played five matches.

Across those matches:

Total xG = 9.80

Total goals = 5

The difference is substantial.

But the correct conclusion is not automatically:

“Team Red is unlucky.”

A deeper analysis would ask:

• Were the chances genuinely high quality?
• Were many opportunities concentrated in difficult shooting situations?
• Were there penalties?
• Did the same players take most of the chances?
• Were opposing goalkeepers performing unusually well?
• Was the team repeatedly missing clear opportunities?
• Did the team's attacking structure change during the five matches?

Only after investigating these questions can the discrepancy be interpreted responsibly.

What xG Can Tell You — and What It Cannot

xG can help describe:

• quality of scoring opportunities;
• attacking chance creation;
• defensive chance prevention;
• differences between goals and underlying chance quality;
• trends across multiple matches;
• player opportunity profiles;
• team attacking profiles.

xG cannot independently tell you:

• who is definitely going to win;
• exactly how many goals will be scored;
• whether a player is universally better than another;
• whether a team “deserved” victory in every possible sense;
• why every chance was created;
• the complete tactical quality of a performance.

This distinction keeps the statistic useful rather than turning it into an answer to every football question.

How Fabiaana Recommends Using xG

For readers who already understand the basics from Jose Sixpenze's earlier guide, I would suggest a simple progression.

Start with the total
Understand the overall expected scoring output.

Then inspect the distribution
Determine whether the total came from many chances or a few major opportunities.

Add context
Consider opponent strength, venue and match state.

Compare with actual goals
Investigate significant differences.

Examine the trend
Look at several matches instead of one.

Combine statistics
Use xG alongside shots, possession, chance creation and other relevant indicators.

Finally, watch the football
Statistics can tell you what happened statistically.

Video and tactical analysis can help explain why it happened.

That combination is much stronger than either method on its own.

Final Thoughts

Expected Goals is often introduced as a number assigned to shots.

That explanation is correct, but it understates what the metric can become when used properly.

As Jose Sixpenze explained in his original Expected Goals guide, xG provides a way of measuring the quality of scoring opportunities rather than simply counting attempts or goals.

The next step is learning how to interpret that information.

A total xG figure is only the beginning.

The distribution of chances, match context, opponent strength, game state, penalty contribution, historical trends and the difference between expected and actual outcomes can all change how that number should be understood.

This is why xG should not be treated as a magic number that tells analysts what happened or what will happen.

It is better viewed as evidence.

Good football analysis asks what that evidence means, what it does not explain and whether other information supports the same conclusion.

That is ultimately where xG becomes most valuable.

The question is no longer simply “How much xG did the team produce?”

It becomes:

“What does the pattern behind that xG tell us about the way the team is playing?”

And that is a much more interesting question.

Frequently Asked Questions

Is a higher xG always better?

Not necessarily in isolation. Higher xG generally indicates greater expected scoring output, but the circumstances producing that xG and the quality of the opposition should also be considered.

Can a team have high xG and still play badly?

Yes. xG measures scoring opportunities, not every aspect of football performance. A team could create good chances while performing poorly in possession, pressing, defending or ball progression.

Why should I look at several matches instead of one?

Football has substantial short-term variation. A larger sample provides more information about whether an observed pattern is persistent.

Does xG predict the final score?

No. xG represents expected scoring based on the model's assessment of chances. It does not determine the actual score.

Should xG be used alone?

No. It is generally more informative when combined with other relevant statistics and qualitative football analysis.

Why can two websites show different xG?

Different providers may use different data and modelling methods. As a result, the same event can receive different values.

Is xG useful for analysing goalkeepers?

Yes, particularly when combined with post-shot information. This can help analysts distinguish the quality of chances faced from what happened to shots after they were taken.

Sources and Further Reading

This article was independently written by Fabiaana. It builds on the earlier xG guide by Jose Sixpenze published on Yowtips, while taking a different analytical approach rather than reproducing its explanations.

• PLOS ONE — Expected goals in football: Improving model performance and demonstrating value — academic research examining xG modelling approaches and predictive value.
• StatsBomb — The Dual Life of Expected Goals — useful discussion of why xG should be interpreted differently across individual matches and larger samples.
• StatsBomb — Mythbusting: Is Long Range Shooting a Bad Option? — historical discussion of xG and shot quality.