Strokes Gained has revolutionized the way we analyze golfers. Gone are the days of looking at fairways hit, greens in regulation, and putts per round to determine a player's true skill. Today, SG off the tee, SG approach the green, SG around the green, and SG putting provides a much more accurate representation of a golfer's skill. Furthermore, strokes gained has transformed and revolutionized golf strategy. It has birthed the driving distance revolution embraced by players such as Bryson DeChambeau and Dylan Frittelli. It has also brought Scott Fawcett's Decade Golf which has provided course management concepts that have improved golfers of all different skill levels. For more information on what strokes gained is, visit this link: https://www.pgatour.com/news/2016/05/31/strokes-gained-defined.html. In this post, we will explore Strokes Gained on the PGA Tour and see the statistical significance of Strokes Gained.
Using Python's Beautiful Soup package, we can obtain all the strokes gained data posted by the PGA Tour since the start of the strokes gained era in 2004. Here are some graphs of each strokes gained category versus strokes gained total.
From this exploratory analysis, we can see that the long game components of driving and approach play appear to be more significant in a player's SG:Total than the short game components of pitching and putting. This is an idea that dominated the "old-school" golf mentality but has been challenged since the advent of strokes gained. Let's do some regression analysis to further investigate this claim.
Univariate Linear Regression: SG:OTT vs SG:Total
Univariate Linear Regression: SG:APR vs SG:Total
Univariate Linear Regression: SG:ARG vs SG:Total
Univariate Linear Regression: SG:Putting vs SG:Total
From our regressions, we have these 4 equations that describe the relationship between each strokes gained category and SG:Total.
SG:Total = 0.11 + 1.01*(SG:OTT)
SG:Total = 0.07 + 1.27*(SG:APR)
SG:Total = 0.12 + 1.26*(SG:ARG)
SG:Total = 0.13 + 0.76*(SG:Putting)
(i.e: For every 1 stroke increase in SG:OTT, there is a 1.01 stroke increase in SG:Total. Essentially, SG:OTT and SG:Total are directly correlated which is very interesting.)
The models tell us that all of the individual strokes gained components are significant to measuring SG:Total. However, further analyses and diagnostics tell us more interesting results. From the equations, it appears that SG:APR and SG:ARG are the most important variables to predicting SG:Total. However, a low multiple r-squared value of 0.1647 tells us that the around the green model is not as strong as the approach the green model with a multiple r-squared value of 0.5133.
Furthermore, contrary to "old-school" beliefs, putting appears to be the least significant component due to it's coefficient of 0.76. To further confirm this, our putting model did not pass R's Breusch-Pagan Test of heteroskedasticity. By rejecting the null hypothesis, we have evidence that heteroskedasticity is present in this model and can be skeptical about its predictive power.
Ultimately, there is statistical evidence to suggest that approach play is most significant to a golfer's score. However, we can also validate this claim intuitively. The margin of error in approach play is very small. For example, the PGA Tour one putt probability from 5 feet is 75%, 10 feet is 38%, and 30 feet is 7%. Just 25 feet of area (just over 8 yards) changes the probability of score by 68%. Now, these probability tapers off, and there is not much of a difference hitting it 30 feet compared to 50 feet (1.978 average score vs 2.135 average score). But, it cannot be denied the enormous effect that a "tight" approach shot brings compared to a worse approach shot.
Again, each part of golf is important in its own way. Longer drives lead to shorter approach shots which lead to shorter putts that players have a better chance of making. Golf is not a series of independent shots. Each shot has an impact on the success of the next shot. However, golfers must emphasize getting longer off the tee and their iron play. The long game is just as important if not more important than the short game, and golfers need to keep this in their minds.
To end this post, below are the top 20 and bottom 20 seasons of all time in the strokes gained era.
Comments