Determinants of Scoring in the Open Championship
This analysis is based on scores and stats from individual rounds in the last 10 Open Championships: 4,673 rounds in total.
Section 1: Absolute Correlation Coefficients with Score
This graph presents the absolute correlation coefficients between the 18-hole score and traditional golf metrics, including Driving Distance, Driving Accuracy, Greens in Regulation, Scrambling, and PPGIR, for all available years.
Interpretation and Discussion:
- Driving Distance: Although highest at Royal Birkdale in 2017, the correlation with score is relatively low, suggesting that sheer driving distance has not been a strong determinant of success in The Open Championship across the years.
- Driving Accuracy: Similar to driving distance but more consistent, accuracy off the tee has had a limited impact on score.
- Greens in Regulation: This metric shows higher correlation, suggesting that hitting greens in regulation was generally influential in determining score in the Open Championship.
- Scrambling: The correlation for scrambling is generally moderate-to-high, with the exception of the 2022 Open Championship at St Andrews.
- PPGIR (Putts Per Green in Regulation): This metric exhibits a relatively higher and more consistent correlation, particularly in the last two years at St Andrews and Royal Liverpool.
This graph shows the absolute correlation coefficients between the 18-hole score and Par metrics (Par3, Par4, and Par5) from 2013 to 2023.
Interpretation and Discussion:
- Par3: The correlation for Par3 is relatively low and stable, suggesting that performance on par-3 holes had a consistent but minor impact on scores.
- Par4: This metric shows much higher correlation, indicating that performance on par-4 holes was a significant determinant of the overall score, particularly at the Open Championships held at St Andrews (2015 and 2022).
- Par5: The correlation for Par5 is moderate, with some variability, implying that performance on par-5 holes had an influence on the 18-hole score across different years and Open venues.
Section 2: Partial Dependence Plots against Score
Partial dependence plots (PDPs) are a tool used in machine learning and statistical modeling to illustrate the relationship between a target variable and one or more feature (e.g. SGApp, SGATG, DrivingDistance, GreensInRegulation). They show the marginal effect of a feature on the predicted outcome of a model. PDPs are particularly useful for understanding how individual features impact the target variable, allowing for better interpretation and insights from the model.
In determining the value of Score, PDPs can help visualize how changes in each feature impact the predicted score, holding other features constant. This can provide insights into which features are most influential and how they affect the score.
This figure presents the partial dependence plots for traditional golf metrics, including Driving Distance, Driving Accuracy, Greens in Regulation, Scrambling, and PPGIR.
- Driving Distance: The plot shows how changes in driving distance impact the predicted score. Variations in driving distance influence player performance in The Open Championship, where course length and weather conditions vary significantly.
- Driving Accuracy: The partial dependence plot for driving accuracy highlights its influence on the predicted score. Accurate drives are essential at The Open Championship due to the often challenging rough and course layout.
- Greens in Regulation: This plot illustrates the impact of greens in regulation on the predicted score. Hitting greens in regulation is vital for scoring opportunities, especially on the generally large greens found in The Open Championship.
- Scrambling: The scrambling plot demonstrates the effect of a player's ability to recover from missed greens on the predicted score. Effective scrambling is important in The Open Championship, where greens and surrounding areas can be particularly challenging.
- PPGIR (Putts Per Green in Regulation): This plot displays the impact of putting efficiency on score. Putting performance is critical at The Open Championship, where greens can be undulating and putting impacted by wind.
This figure shows the partial dependence plots for Par metrics, including Par3, Par4, and Par5.
- Par3: The partial dependence plot for Par3 indicates how performance on par-3 holes affects score. The effect is significant, but not as marked as for Par-4 holes.
- Par4: The plot for Par4 reveals its impact on the predicted score. Par-4 holes often represent the majority of holes on each course on the Open rotation, and consistent performance on these holes is critical for success at The Open Championship.
- Par5: The Par5 plot shows the effect of performance on par-5 holes on score. The effect is significant, but not as marked as for Par-4 holes.
Section 3: Importance of Each Metric in Determining Score
Random Forest Regressor and Feature Importance
Random Forest Regressor is an ensemble learning method that constructs multiple decision trees during training and outputs the average prediction. It combines the predictions of several models to improve accuracy and robustness.
Feature importance is a technique used to interpret a machine learning model. It refers to the score that quantifies the contribution of each feature to the prediction made by the model.
In a Random Forest, the importance of a feature is computed by looking at how much the feature decreases the impurity (e.g., variance for regression tasks) across all the trees in the forest. The more a feature decreases the impurity, the more important it is considered.
The calculated importance scores for all features are then normalized to give relative importance as a percentage. This shows the relative contribution of each feature to the prediction task.
Interpreting Feature Importance
Features with high relative importance percentages have a strong impact on the model's predictions. They are crucial for accurate predictions and indicate key areas where performance matters most.
Features with low relative importance have a minimal impact on the model's predictions. While they can still contribute, they are less critical.
Using Random Forest Regressor, the relative importance of each factor on Score is quantified as follows:
- DrivingDistance: 6.98%
- DrivingAccuracy: 4.48%
- GreensInRegulation: 31.42%
- Scrambling: 28.51%
- PPGIR (Putts per GIR): 28.62%
The findings indicate that the key to achieving low scores in the Open Championship lies significantly in the ability to hit greens in regulation, recover effectively (scrambling), and excel in putting (PPGIR). While driving distance and accuracy do play a role, they are less critical compared to short game skills and putting.
At The Open Championship, played on links courses with challenging conditions, these factors become even more pronounced. Links courses typically feature deep bunkers, undulating greens, and unpredictable weather, making scrambling and putting paramount. Players who can navigate these challenges by hitting greens, recovering effectively when they miss, and putting well are more likely to succeed.
For comparison, here is the relative importance of each factor across all DP World Tour events over the last 10 years:
- DrivingDistance: 9.56%
- DrivingAccuracy: 5.77%
- GreensInRegulation: 28.39%
- Scrambling: 26.15%
- PPGIR (Putts per GIR): 30.12%
Using Random Forest Regressor, the relative importance of each factor on Score is quantified as follows:
- Par3: 12.84%
- Par4: 72.63%
- Par5: 14.53%
The importance of par 4 performance highlighted in the analysis aligns with the typical course setup of The Open, where par 4s are the most common hole type. Additionally, the ability to take advantage of scoring opportunities on par 5s and avoid significant errors on par 3s are important in this event.
For comparison, here is the relative importance of each factor across all DP World Tour events over the last 10 years:
- Par3: 15.36%
- Par4: 64.77%
- Par5: 19.87%
Top 5 Ranked Players - 2024 Open Championship
The table below shows the top-5 ranked players and their average estimated scores from the three different Random Forest models above.
| Player | Score |
|---|---|
| Scottie Scheffler | 68.42 |
| Joaquin Niemann | 68.97 |
| Ludvig Aberg | 69.14 |
| Xander Schauffele | 69.14 |
| Jon Rahm | 69.18 |
Estimated scores for all players can be found here.