Handicap System Update – Course & Slope Ratings
Over the last couple of weeks Mitchell and I have been exploring a plan to revise and improve the Handicap System in The Golf Club. Starting today the first step of our plan will go into effect: The Course & Slope Rating Update.
We have changed the method by which we calculate course & slope ratings. We have also changed how many, and what kind of plays are needed for a course to get a course & slope rating, as well as how we keep ratings up to date. These changes are live now.
New Calculation Method
The primary update here is how we decide what the course & slope ratings are for a given course under the hood. Here’s a run-down.
The Old Ways
Up until now the calculation method was making an educated guess at what the course & slope ratings should be based solely on the first two-hundred scores per platform. What it didn’t take into account was the relative skill of the players submitting each of those scores. This would be perfectly fine if every course had the same spread of skilled players playing on it, but that’s not the case. Players tend to gravitate toward courses based on their personal skill levels. This generally meant easier courses were often rated harder than they should’ve been, and the more challenging courses, rated easier.
Now that we have a base-line handicap for our active players we can use that to measure relative skill, and get a better picture of how skill maps to scores on a course. Here’s an example from the course Magnolia National, Georgia (Sunday) by TGC Member TheCLV24.
Reading the Rounds
As you can see there’s a pretty good correlation between the Handicap Index (X-Axis) and Score (Y-Axis). This is a phenomenon that’s consistent across all of the courses that we’ve tested. With data like this it’s very easy to visualize a straight line through the centre of the data. That straight line is what we want, as it’s a literal representation of slope rating. Luckily for us, computers are pretty good at finding the line of best fit for data like this.
Here’s Magnolia National again with the line of best fit on top.
Cream of the Crop
The Handicap System is meant to represent the expected scores on an ‘average best’ round of a golfer of a given skill. This is why your Scoring Record only uses the best 10 of your last 20 scorecards when determining your Handicap Index. Course & slope ratings need to work the same way.
As you can see in the graph above, the line of best fit represents the middle of the data. What we want instead is the middle of the best half of the data. To get this we ignore any scores that fall above the first line of best fit from the calculation and get a new line of best fit with the left over data.
Take a look at the Magnolia National graph with the ignored data cut away (red) and the new line of best fit for the remaining data (blue).
We’ve Got Linear Equations for Days!
If you can remember back to your grade-school math class you’ll know that a straight line on a graph can be represented by the equation y = mx+ b. That’s exactly what we need to get our course & slope ratings. You can see the linear equation we want for the graph above on the right in blue.
To get the slope from these equations is straight-forward, we simply multiply the ‘m’ value by 113—the ‘standard’ slope rating in the real world. So in the case of Magnolia National, we’d get 1.2525 * 113 = 142.
Looking at these graphs and the linear equations that come with them, one would expect that the course rating would be the Y-intercept (the value of ‘b’) of the line of best fit. This is what we thought at first too, but then we remembered why we were doing this in the first place: to improve handicap accuracy! We’re calculating values based on existing Handicap Indices, which are not as accurate as we would like, so we need to compensate.
We found that if we used the Y-intercept for the course ratings, the majority of courses would receive much higher (read: easier) ratings than before. This will in turn lower handicaps over time. To address this we decided to adjust the X value used to determine the course rating through the linear equation to -3. This means that the average best score of a +3 golfer will be taken as the course rating. Using Magnolia nation as an example again we get the following:
- Course Rating = mX + b
- Course Rating = 1.2525(-3) + 78.721
- Course Rating = 75
The value of -3 is not arbitrary. It is the best balance we found to equalize the spread of handicap indices in the community over time, without making too big a change too quickly. As players’ handicaps begin adjusting to the new course and slope ratings—and in turn course ratings begin to balance out—the X value will be normalized back to zero.
As time goes on course and slope ratings on courses will continue to self-correct with this new system. You will notice the ratings changing periodically, even on well-established courses. As we continue to observe the performance of the course rating system we will explore options of adjusting the minimum requirements for getting a course rated. The goal is to get an accurate value at as quick a pace as possible.
One of the big benefits of this new system is consistent course & slope ratings across platforms. Scorecards are now used from all three platforms and used to determine global course & slope ratings.
The number and type of scorecards needed to get rated has also changed. With the new system only 300 global plays are required to get your course rated on all three platforms—it was 200 plays per platform (for a total of 600) before. One caveat is that those 300 plays must be posted by users with handicaps; this ensures we have all the information needed to make an accurate calculation.
You may notice that there are still quite a few courses that do not have cross-platform course & slope ratings. The reason for this is likely because the course had 200 plays on the respective platform, but not 300 handicapped plays across all platforms. In these cases we have kept the legacy ratings intact until such time as the course has enough cross-platform handicapped plays to re-evaluate under the new system.
Effect on Handicaps
Some folks may wonder what effect these changes will have on their current handicap, and the answer is “none”. This change only affects the ratings of courses, and therefore will only affect your handicap if you post new rounds to your scoring record. Scorecards already in your scoring record have their corresponding course and slope ratings locked in with them, and will not change. Any unfinished scorecards you have will have the course & slope ratings synchronized to the most up to date values when you resume.
Final Notes & Looking Forward
There were a few other notes I wanted to mention before finishing up that didn’t fit in any of the above sections:
- Course Plays are now cross-platform but these are all plays, handicapped or not. Remember that to be rated, a course must have 300 handicapped plays.
- Some folks on the forums have asked for a view that breaks up the plays on their courses between Xbox One, PS4, and Steam platforms. This is not possible due to platform restrictions. We cannot acknowledge Xbox One on the PS4 version or vice versa.
- Player Ratings for courses have not changed to cross-platform. This is something we are interested in doing, but we need to explore the possible repercussions of doing so first.
We hope you’ll share your experiences with us as this on-going improvement in the handicap system continues, and we hope you reap the benefits of these changes soon. See you on the forums, and on the course!