What is xG (Expected Goals)?
The metric that changed how we understand football. Here's what xG actually means and why it matters for understanding the beautiful game.
The 30-Second Explanation
Expected Goals (xG) measures the quality of a scoring chance. It answers the question: "Based on thousands of similar shots, how often does this type of chance result in a goal?"
A penalty kick has an xG of about 0.76 — roughly 76% of penalties are scored. A shot from 30 yards out might have an xG of 0.03 — only about 3% of those go in. A tap-in from 3 yards? That's closer to 0.90.
Add up all the xG from a team's chances in a match, and you get their total xG — a measure of how many goals their performance deserved, regardless of how many they actually scored.
Why xG Matters
Football is a low-scoring game where luck plays a huge role. A team can dominate a match, hit the woodwork three times, and lose 1-0 to a lucky deflection. The scoreline says one thing; the actual performance tells another story entirely.
Real Example: The Misleading Scoreline
The 0-0 suggests an even game. The xG reveals Brighton dominated and were desperately unlucky. If you're trying to predict Brighton's next match, which number tells you more?
xG strips away the noise of finishing luck and reveals the underlying performance. Over time, a team's actual goals tend to converge with their xG — making it one of the most predictive metrics in football analytics.
How is xG Calculated?
Every shot in professional football is logged with detailed information. Machine learning models analyze millions of historical shots to determine which factors most influence whether a shot becomes a goal.
Distance & Angle
Closer shots and central positions have higher xG. A shot from the penalty spot is worth more than one from a tight angle near the byline.
Body Part
Headers are generally converted at lower rates than foot shots. The model accounts for how the shot was taken.
Assist Type
Through balls and cutbacks create higher-quality chances than crosses or long balls. The build-up matters.
Game State
Whether it's open play, a fast break, or a set piece affects the probability. Counter-attacks typically yield higher xG chances.
Different providers (Opta, StatsBomb, etc.) use slightly different models with varying inputs, which is why xG values can differ between sources. The concept remains the same — measuring chance quality based on historical data.
When xG and Goals Don't Match
Sometimes a player or team consistently outperforms or underperforms their xG. This tells us something interesting:
Outperforming xG (Scoring more than expected)
Could indicate elite finishing ability (think prime Messi or Lewandowski) or could be a sign of luck that's likely to regress. Sustainable over-performance is rare — only the very best finishers consistently beat their xG.
Underperforming xG (Scoring less than expected)
Often a sign of bad luck or poor finishing that's likely to improve. Teams significantly underperforming their xG are often good candidates for a turnaround in results.
Over a season (30+ matches), xG and actual goals tend to align closely. Over 5 matches, they can diverge wildly. This is why xG is particularly useful for predicting future performance — it smooths out the variance.
What Can You Use xG For?
Understanding Match Performance
Look beyond the scoreline. A 2-0 win where you had 0.8 xG was fortunate. A 0-0 where you had 3.2 xG was unlucky. xG tells the story the scoreboard can't.
Evaluating Teams
League tables show results; xG tables show underlying quality. A team sitting mid-table with excellent xG numbers might be due for a run of wins.
Assessing Players
Is a striker in great form or just getting lucky? Compare their goals to their xG over time to understand their true finishing ability.
Making Predictions
xG is more predictive of future results than past goals. This is why modern prediction models — including ours — are built on xG foundations.
The Limitations of xG
xG is powerful, but it's not perfect. Understanding its limitations helps you use it wisely:
- •It doesn't account for who's shooting. A Haaland chance is worth more than the same chance falling to a defender, but basic xG treats them equally.
- •Goalkeeper quality isn't factored in. The same shot has the same xG whether it's against a world-class keeper or a struggling one.
- •Defensive pressure varies. Some models account for this, but many don't capture how much space the shooter had.
- •Small samples are noisy. One match of xG data tells you less than a full season. Don't overreact to single-game numbers.
The Bottom Line
xG is the best tool we have for measuring what actually happened in a football match, stripped of the randomness inherent in whether shots hit the net or not.
It won't tell you everything — football is too complex for any single number. But combined with other metrics and your own football knowledge, xG gives you a much clearer picture than goals alone ever could.