Player Ratings in PGA Golf

There are a small number of applied ratings systems currently in operation in PGA golf. The most well-known would have to be the Official World Golf Rankings (OWGR), which is an example of an accumulative player rating system. Each ratings system has been developed to measure one particular aspect of player performance, be it expected difference in scores, such as in Sagarin Ratings, or the accumulation of tournament success, such as in OWGR. This research had two aims when developing a ratings system; the first was to measure the variation in player performance between rounds, and the second to determine a recommended sample size for performance analysis. A moving average based rating system was developed using all tournaments from the 2012 through 2015 PGA Tour seasons. The system used a relative measure of performance taken to be the ratio of player round score to average field round score. Model fits were very encouraging with adjusted R2 measures consistently above .98. An algebraic manipulation of ratings equations showed that roughly 94% of variation in performance could be explained using a long-term measure of player strength, which can be defined as predominantly comprising scores greater than five rounds previous. The remaining 6% was described using results from more recent rounds, no more previous than four rounds back. Varying the total number of samples showed that a balance of predictive accuracy and minimisation of included samples occurred when a total of 20 rounds were used. Accuracy is seen to increase with the inclusion of more samples but plateau once 20 have been included. Results from this research provide a foundation for longitudinal analysis of performance in golf tournaments.