Capstone - Complete Analytics System
11.1 The Modern Goalkeeper Revolution
Goalkeeper analysis has undergone a revolution in recent years. Traditional metrics like save percentage failed to account for shot quality, leading to the development of sophisticated models that evaluate goalkeepers on shot-stopping, distribution, and sweeping ability.
Learning Objectives
- Understand post-shot expected goals (PSxG) and goals saved above expected (GSAx)
- Analyze goalkeeper distribution and passing patterns
- Evaluate sweeper keeper actions and effectiveness
- Build comprehensive goalkeeper evaluation frameworks
- Create goalkeeper comparison visualizations
Why Traditional Metrics Fail
Save percentage treats all shots equally, but a goalkeeper facing mostly long-range efforts will have a higher save rate than one facing point-blank chances. Consider:
- GK A: 75% save rate (easy shots)
- GK B: 68% save rate (difficult shots)
- Conclusion: GK A is better?
- GK A: -2.5 GSAx (underperforming)
- GK B: +4.3 GSAx (overperforming)
- Conclusion: GK B is actually better!
11.2 Post-Shot Expected Goals (PSxG)
Post-shot expected goals (PSxG) is a refinement of xG that incorporates information about the shot itself - specifically where the ball is heading. PSxG is calculated after the shot is taken and considers:
PSxG Components
Shot Placement
- Distance from goal center
- Height of the shot
- Top corners are hardest to save
Shot Velocity
- Harder shots are harder to save
- Reaction time decreases
- Combined with distance
Shot Trajectory
- Swerve and dip
- Shot technique
- Deflections handled separately
Understanding PSxG Values
| Shot Type | Typical xG | PSxG (Top Corner) | PSxG (Center) |
|---|---|---|---|
| 18-yard shot | 0.08 | 0.75-0.90 | 0.02-0.05 |
| Inside box | 0.15 | 0.80-0.95 | 0.08-0.15 |
| Six-yard box | 0.40 | 0.85-0.98 | 0.25-0.40 |
| Penalty | 0.76 | 0.95+ | 0.50-0.70 |
11.3 Goals Saved Above Expected (GSAx)
Goals Saved Above Expected (GSAx) is the premier metric for evaluating goalkeeper shot-stopping. It compares actual goals conceded to the PSxG of shots faced:
GSAx Formula
GSAx = PSxG Faced - Goals Conceded
- Positive GSAx: Goalkeeper saved more than expected (good)
- Negative GSAx: Goalkeeper conceded more than expected (poor)
- Zero GSAx: Goalkeeper performed exactly as expected
GSAx Interpretation Guide
| > +8 | Elite |
| +4 to +8 | Above Average |
| -2 to +4 | Average |
| -6 to -2 | Below Average |
| < -6 | Poor |
| > +0.20 | Elite |
| +0.10 to +0.20 | Above Average |
| -0.05 to +0.10 | Average |
| -0.15 to -0.05 | Below Average |
| < -0.15 | Poor |
11.4 Distribution Analytics
Modern goalkeepers are expected to be excellent distributors. Distribution analytics evaluate a goalkeeper's passing ability, including their role in build-up play and ability to launch attacks.
Key Distribution Metrics
Pass Completion %
- Short: <15m
- Medium: 15-30m
- Long: >30m
Launch %
- % of passes that are long
- Playing style indicator
- High = more direct
Progressive Distance
- Yards moved upfield
- Passing and carrying
- Build-up contribution
Goal Kicks
- Short vs Long %
- Average length
- Success rates
Distribution Styles
| Style | Launch % | Short GK % | Avg Pass Length | Example |
|---|---|---|---|---|
| Ball-Playing | <25% | >70% | <25m | Ederson, Alisson |
| Hybrid | 25-40% | 50-70% | 25-35m | Lloris, Martinez |
| Traditional | >40% | <50% | >35m | Direct play keepers |
11.5 Sweeper Keeper Analytics
The "sweeper keeper" role involves goalkeepers acting as an extra defender, coming off their line to intercept through balls and clear danger. This is measured through several key metrics:
Defensive Actions Outside Box
Clearances, interceptions, and tackles made outside the penalty area
Average Distance from Goal
How high up the pitch the keeper positions during defensive actions
#OPA (Outside Penalty Area)
Count of actions performed outside the 18-yard box
Sweeper Keeper Performance
11.6 Cross Handling and Claiming
A goalkeeper's ability to claim crosses and deal with aerial threats is crucial, especially in leagues where crossing is prevalent. Key metrics include:
Cross Claiming Metrics
- Crosses Faced: Total crosses into the box
- Crosses Stopped %: Percentage of crosses claimed or punched
- Claim Success Rate: % of claims without dropping
- Punch Rate: % of stops that were punches vs catches
- Aerial Duels Won: Contested aerial challenges won
- Drop Rate: Catches spilled under pressure
- Corners Claimed: Set pieces handled
- Command of Box: Composite metric
11.7 Comprehensive Goalkeeper Evaluation
Elite goalkeeper analysis combines shot-stopping, distribution, and sweeping into a comprehensive evaluation framework. Here we build a complete goalkeeper profile:
Important Considerations
- Team quality affects workload and shot quality faced
- Different systems require different goalkeeper types
- Sample size matters - at least 1000+ minutes for reliable metrics
- League differences impact cross-comparison
Building a Goalkeeper Radar Chart
Complete Goalkeeper Report
11.8 Practice Exercises
Exercise 11.1: Calculate GSAx Season Rankings
Task: Using FBref data, calculate GSAx for all Premier League goalkeepers with 1500+ minutes. Create a bar chart showing the top 10 and bottom 5 performers.
Exercise 11.2: Distribution Style Classification
Task: Create a scatter plot categorizing goalkeepers by their distribution style. Use Launch % on the x-axis and Pass Completion % on the y-axis. Add quadrant labels to identify ball-playing vs traditional keepers.
Exercise 11.3: Build a Sweeper Keeper Index
Task: Build a sweeper keeper index combining OPA/90, average distance from goal, and cross stop percentage. Rank the top 15 goalkeepers and visualize with a lollipop chart.
11.9 Chapter Summary
Key Takeaways
- PSxG accounts for shot placement and quality, providing better evaluation than raw save percentage
- GSAx is the gold standard for shot-stopping evaluation, comparing actual performance to expected
- Distribution analytics reveal a goalkeeper's passing style and build-up contribution
- Sweeper metrics quantify a goalkeeper's ability to act as an extra defender
- Comprehensive evaluation requires balancing multiple skill dimensions based on team needs
Next Steps
In Chapter 12, we'll explore positional analytics for defenders, examining metrics for center-backs and full-backs including defensive actions, progression, and aerial dominance.