Goals scored is the stat most bettors look at first, and it's one of the noisiest. Expected Goals (xG) shows what goals can't: which teams are actually creating quality chances and which are getting lucky.
Our own data shows how noisy goals are: in our study of 372,650 player starts, goals were among the least repeatable player stats (split-half correlation 0.66, vs 0.89 for shots).
Here's why xG is one of the most important metrics for football analysis.
What is xG? (Expected Goals Definition)
xG (Expected Goals) measures the quality of a scoring chance in football. Each shot gets a value between 0 and 1 based on how likely it is to be scored.
Think of it as: xG = Probability that shot results in a goal
Example: a penalty is usually rated at around 0.76-0.79 xG (roughly a 76-79% chance, depending on the model), while a speculative long-range shot might be 0.03 xG (3% chance)
xG is calculated from historical data analyzing:
- Shot location (distance & angle to goal)
- Type of assist (through ball, cross, corner, cutback)
- Body part used (foot, head, other)
- Game situation (open play, counter-attack, set piece)
- Defensive pressure (number of defenders, goalkeeper position)
Master statistical analysis for football betting →
xG Quality Scale (What xG Values Mean)
| xG Value | Quality | Example Situations | Conversion Rate |
|---|
| 0.01 - 0.05 | Very Low | Long-range shots (25+ yards) | 1-5% |
| 0.05 - 0.15 | Below Average | Edge of box, tight angle | 5-15% |
| 0.15 - 0.35 | Average | Inside box, contested | 15-35% |
| 0.35 - 0.60 | Good | Central position, space | 35-60% |
| 0.60+ | Excellent | One-on-one, tap-ins, penalties | 60-90% |
How to Read Match xG (Real Example)
Example Match: Team A 2.1 xG - Team B 0.4 xG (Actual Score: 1-1)
This tells you:
- Team A created much better quality chances
- Team A "deserved" to win based on underlying performance
- Team A was unlucky (underperformed xG)
- Team B was lucky (overperformed xG)
- Betting Insight: if this keeps happening over several matches, Team A may be undervalued for upcoming games
Key Rule: over a season, xG is a better predictor of future results than actual goals scored, because it filters out finishing luck.
1. Find Unlucky/Lucky Teams (Value Betting Opportunity)
Unlucky Team = Value Opportunity:
- Creating high xG per game
- Scoring fewer goals than xG (negative variance)
- Can be undervalued if the market leans on recent results
- Expected to regress to mean (improve results)
Lucky Team = Be Careful Backing Them:
- Low xG per game
- Scoring more goals than xG (positive variance)
- Can be overvalued if the market leans on recent results
- Expected to regress to mean (worsen results)
Note: bookmakers use xG too, so the gap is usually smaller than it looks. The opportunity is in prices that still follow the scoreline.
Learn how to find value bets systematically →
2. Compare Home vs Away xG
Check xG separately by venue:
- Home games xG
- Away games xG
Example (hypothetical): a team averaging 2.1 xG at home but 0.9 xG away is two different teams depending on the venue.
League context: in Europe's big five leagues in 2025/26, home teams scored 1.53 goals per match against 1.23 for away teams. Some teams have a much bigger home/away gap than that.
Deep dive into home vs away analysis →
Look at recent xG trends:
- Average xG created (attacking strength)
- Average xG conceded (defensive strength)
- xG difference (xG for - xG against = overall quality)
Sample size matters: Use minimum 5 games, ideally 8-15 games for reliable patterns.
Common xG Mistakes to Avoid
❌ Don't: Judge team quality based on one game's xG (variance too high)
✅ Do: Use at least 5-8 games for reliable xG trends
❌ Don't: Ignore the actual score (context matters)
✅ Do: Use xG + match context together (motivation, weather, tactics)
❌ Don't: Think a 0.5 xG chance "should" produce half a goal
✅ Do: Understand a single 0.5 xG shot is a 50% probability (variance exists)
❌ Don't: Compare xG across different data providers (different models)
✅ Do: Stick to one consistent xG source for analysis
Avoid the 5 biggest analysis mistakes →
Quick Reference: 4 xG Golden Rules
1. Higher xG = Better Chances
Team with higher xG usually deserved to win (regardless of actual score)
2. xG Predicts Future Better Than Goals
A team with good xG but poor results will likely improve (regression to mean)
3. Track Over Multiple Games
Need 5-8 games minimum for reliable xG patterns (avoid recency bias)
4. Compare Like with Like
Home xG vs home xG, away vs away (venue makes huge difference)
Faster xG Analysis
Tracking xG in spreadsheets means collecting data from several sources for every match. A faster routine:
- xG source: Understat for the big five leagues, FotMob or SofaScore for wider coverage
- Over/underperformance: our free xG difference calculator compares xG with actual goals for both teams
- Shot volume and home/away splits: StatPair shows shots and shots on target for every team and 35,000+ players, with home/away filters (StatPair doesn't provide xG itself)
→ Try StatPair free for 7 days
Real xG Analysis Example
Hypothetical example: Team A (Last 8 Games)
- Average xG created: 2.0 per game
- Actual goals scored: 8 goals (1.0 per game)
- Variance: Underperforming xG by 8 goals (-50%)
Betting Insight:
Team is creating quality chances but unlucky with finishing. That's a potential value opportunity if the odds are still based on the poor results rather than the underlying performance.
What usually happens next? Teams this far below their xG tend to score more over the following matches. The edge only exists if the odds still reflect the poor results.
Frequently Asked Questions About xG
Q: What does xG mean in football?
A: xG (Expected Goals) measures the quality of scoring chances. It represents the probability (0 to 1) that a shot results in a goal based on historical data. xG helps identify underlying team performance beyond actual goals.
Q: Is xG accurate for predicting football matches?
A: Over a season, xG predicts future results better than actual goals. While no metric is 100% accurate, xG over 8-15 games reveals true team quality better than results (which include luck/variance). Use xG as one key input, not sole predictor.
Q: Where can I find xG stats for free?
A: Free xG sources include Understat (big five leagues plus Russia) and FotMob or SofaScore (wider coverage). FBref removed its advanced data, including xG, in January 2026. Premium sources: Opta, StatsBomb (expensive for casual users). StatPair doesn't provide xG, but our free xG difference calculator helps you compare xG with actual goals.
Q: What's a good xG per game for top teams?
A: Over a season, league-average xG per team is close to league-average goals per team: Europe's big five leagues averaged 2.76 goals per match in 2025/26, or about 1.38 per team. The best attacking sides sit well above that and the best defences well below it. Context matters: in 2025/26, home teams in the big five scored 1.53 goals per match against 1.23 for away teams, so always compare home with home and away with away.
Q: How do I use xG for betting?
A: Compare team's xG to actual goals over 8+ games. Teams underperforming xG = value opportunity (expect improvement). Teams overperforming xG = avoid (expect regression). Combine with home/away splits and H2H data for best results.
Q: Can xG be negative?
A: No, xG is always 0 or positive (probability can't be negative). If a team has 0.0 xG, they created no quality chances. If confused by "xG difference" (xG for minus xG against), that can be negative.
xG helps you see beyond the scoreline to understand underlying team performance. Use it to find:
- Teams that are better than their results show (value bets)
- Teams that are worse than their results show (avoid traps)
- Real form vs lucky/unlucky variance runs
- Tactical strengths (high xG creation) and weaknesses (high xG conceded)
Remember: goals tell you what happened. xG tells you what was likely to happen. Bet on the second.
Want to go deeper? Bookmark our complete Football Betting Statistics Glossary covering the metrics every analyst should know, and learn to build your own betting model in Excel using xG and Poisson distribution.
Combine xG with StatPair - shots, shots on target, home/away splits and form for 35,000+ players. 7-day free trial.
P.S. xG is one of the most useful stats in football, but only when you use it with context: home/away splits, sample size and opponent strength. Use our free xG difference calculator to spot over- and underperformance.