How To Calculate Expected Goals

Calculate the probability of a shot resulting in a goal based on key match factors

Comprehensive Guide: How to Calculate Expected Goals (xG)

Expected Goals (xG) is an advanced soccer metric that measures the quality of a shot based on several factors to determine the probability that the shot will result in a goal. This statistical model has revolutionized how analysts, coaches, and fans evaluate player performance and team strategies.

The Science Behind Expected Goals

xG models use historical data from thousands of shots to calculate the probability of a goal being scored from a particular situation. The core principle is that not all shots are equal – a shot from 6 yards out has a much higher probability of scoring than one from 30 yards.

Key Factors in xG Calculation

1. Shot Location: The distance from goal and angle to the goalposts are the most significant factors. Shots from central positions closer to goal have higher xG values.
2. Shot Type: Headers generally have lower xG than foot shots, while volleys can vary based on the situation.
3. Assist Type: Through balls and crosses create different quality chances compared to simple passes.
4. Defensive Pressure: The number and proximity of defenders can significantly impact the shooter’s ability to place the ball accurately.
5. Body Part Used: Shots with the preferred foot typically have higher xG than weak foot or header attempts.
6. Player Position: Strikers generally convert chances better than defenders, which is factored into some xG models.

Mathematical Foundation of xG Models

Most xG models use logistic regression, a statistical method that predicts binary outcomes (goal or no goal). The general formula is:

xG = 1 / (1 + e^-z)

Where z is a linear combination of the various factors:

z = b₀ + b₁×distance + b₂×angle + b₃×shot_type + … + b_n×factor_n

The coefficients (b₁, b₂, etc.) are determined through machine learning algorithms trained on historical shot data.

Comparison of xG Models Across Leagues

League	Average xG per Shot	Conversion Rate	Top Team xG/Game	Bottom Team xG/Game
English Premier League	0.10	10.5%	2.1	0.8
Spanish La Liga	0.09	10.1%	1.9	0.7
German Bundesliga	0.11	11.2%	2.3	0.9
Italian Serie A	0.08	9.8%	1.8	0.6
French Ligue 1	0.10	10.3%	2.0	0.8

Practical Applications of Expected Goals

• Player Evaluation: xG helps identify players who consistently create high-quality chances (high xG) or finish them well (score more than their xG).
• Tactical Analysis: Coaches use xG to evaluate which attacking patterns create the best chances and which defensive systems concede the fewest high-quality opportunities.
• Transfer Market: Clubs use xG metrics to identify undervalued players who may be performing better than traditional statistics suggest.
• Betting Markets: Sophisticated bettors incorporate xG data to find value in betting markets where odds don’t accurately reflect true scoring probabilities.
• Youth Development: Academies use xG to track player development and identify which young players show promise in creating or converting chances.

Limitations of Expected Goals

While xG is a powerful tool, it has some limitations that analysts should be aware of:

1. Context Missing: xG models don’t account for game state (score, time remaining) which can significantly affect player decision-making.
2. Player Quality: Most models don’t account for the specific shooter’s ability, treating all players equally for a given shot.
3. Goalkeeper Quality: The model assumes an average goalkeeper, though in reality, some keepers are significantly better at saving certain types of shots.
4. Defensive Organization: While pressure is factored in, the specific defensive formation and organization isn’t always captured.
5. Data Quality: The accuracy depends on the quality and granularity of the underlying data collection.

Advanced xG Metrics

Beyond basic xG, analysts have developed several related metrics:

Metric	Description	Typical Use Case
xG Chain	Measures a player’s involvement in buildup play leading to shots	Evaluating creative midfielders and playmakers
xG Buildup	Similar to xG Chain but excludes the final pass/shot	Assessing deep-lying playmakers
xG Assisted	The xG value of shots a player assisted	Evaluating creative players who set up chances
xGOT (xG on Target)	Expected goals based only on shots on target	Assessing goalkeeper performance
xG Difference	Team’s xG for minus xG against	Evaluating overall team performance

How to Implement xG in Your Analysis

For coaches and analysts looking to implement xG in their workflow:

1. Data Collection: Use tracking data providers like Opta, StatsBomb, or WyScout which provide the raw data needed for xG calculations.
2. Model Selection: Choose between open-source models or develop your own based on your specific league/data.
3. Visualization: Present xG data through heatmaps, shot charts, and comparative analyses to make it actionable.
4. Contextual Analysis: Combine xG with other metrics like possession, pressing intensity, and player positioning for deeper insights.
5. Longitudinal Tracking: Monitor xG trends over time to identify improvements or declines in performance.

Authoritative Resources on Expected Goals

For those interested in the academic foundations of expected goals, these resources provide valuable insights:

• MIT Sloan Sports Analytics Conference Paper on xG (2017) – A foundational academic paper presenting xG methodology
• National Library of Medicine – Application of Machine Learning in Soccer Analytics – Discusses advanced statistical methods in soccer including xG
• U.S. Soccer Coaching Resources – Includes modern analytical approaches used by national team coaches

The Future of Expected Goals

As technology advances, xG models are becoming more sophisticated:

• Player-Specific Models: Future models may incorporate individual player abilities to provide more personalized xG values.
• Real-Time Tracking: Integration with player tracking data (like STATSports or Catapult) could add factors like player speed and acceleration at the moment of shooting.
• AI Enhancement: Machine learning techniques may identify non-linear relationships between factors that current models miss.
• Tactical Context: More advanced models may incorporate tactical formations and opponent quality in real-time.
• Biomechanical Data: Wearable technology could provide data on shooting technique to refine xG calculations.

Expected Goals has fundamentally changed how we understand soccer, moving beyond simple counting statistics to a more nuanced appreciation of chance quality. As the sport continues to embrace analytics, xG will remain at the forefront of performance evaluation and tactical analysis.