Logo image
Back
Assessing the effect of player ability on scoring performance for holes of different par score in PGA golf
Conference paper   Peer reviewed

Assessing the effect of player ability on scoring performance for holes of different par score in PGA golf

Bradley O'Bree, James Baglin and Anthony Bedford
Proceedings of the 5th International Conference on Mathematics in Sport, pp.133-138
MathSport International Conference, 2015 (Loughborough, United Kingdom, 29-Jun-2015–01-Jul-2015)
MathSport International
2015
url
http://www.mathsportinternational.com/MathSport2015Proceedings.pdfView
Webpage

Abstract

Human Movement and Sports Science
This work looks to serve as foundation research into the effect scoring performance on holes of different pars has on corresponding final round scores in PGA golf. When conducting predictive modelling to estimate round scores it can be argued that each of the different hole structures, par threes, fours and fives, should be player dependent because players can have different strengths and weaknesses. The purpose of this study was to investigate the degree to which performance is player dependent by analysing the variation of par totals (the total score for all hole scores for each par type) using multilevel regression modelling. An intercept only multilevel regression model allows the decomposition of the variance of scores into two parts, a par total based and a player based. The separate components are used to calculate the Intraclass Correlation Coefficient (ICC), which describes the proportion of variation in par totals that can be attributed to the player. Smaller ICC values indicate smaller player components and imply performance does not vary across players. The complete 2014 US PGA tour was analysed, with results indicating that when all players were considered together that each ICC was significant (p<.00), however each relatively small at .012, .046 and .055 for par threes, fours and fives respectively. Controlling for player strength showed clear differences in ICC estimates and levels of significance, with a consistent trend of decreasing player strength associated with increasing ICC values and improved statistical significance. The latter result indicates that the proportion of variation in scores explained using the player component increases as player skill decreases, suggesting it is more important to consider player ability at the bottom of the leaderboard than at the top. These results suggest that when modelling par totals, and by extension round scores, some level of player-specific modelling should be incorporated.

Details

Metrics

174 Record Views
Logo image