How it Works
Correlation Analysis
It all begins with an idea probably learned by most in school: correlation analysis. If two variables are correlated, it means that one is influenced by the other. The age-old example is that ice cream sales are correlated with summer months. This is because people tend to buy more ice creams in nice weather.
Some relationships are stronger than others. The stronger the relationship, the easier it is to predict an outcome.
At Golfgeek, we have identified the variables correlated with your strokes gained in a category. For example, your one-putt percentage can predict your strokes gained in short putt distance buckets with almost 100% accuracy.
However, looking at Off The Tee and Approach Play, it quickly becomes clear that more than one variable significantly influences the strokes gained. This is where regression analysis comes in.
Regression Analysis
Regression is similar to correlation analysis, except it allows you to add more than two variables to a model. Now, we can predict strokes gained using a multitude of different inputs.
For example, we can accurately predict your strokes gained off the tee using three variables. These are driving distance, fairway accuracy and poor tee shot percentage. These three variables explain 96% of the variation between strokes gained in this category.
The difficulty comes in building these models. At Golfgeek, we have built 26 different models that can accurately predict your skill.
This allows you to simulate improvement, set goals and find out which variable will drive the most gains. For example: ‘‘Should I focus on hitting more fairways or hitting it further off the tee?”. We answer all these questions