Rating non-elite tennis players using team doubles competition results

Abstract

Statistical methods can be useful in rating non-elite tennis players. This paper shows how clubs can use simple optimization techniques to rate their club's players in doubles tennis competitions. Even though clubs lack all relevant information, the effects of home advantage, position played, partner and strength of team opposition can be taken into account and evaluated. The results from all competing clubs, available to the body organizing the competition, are used to rate all players in the competition, and validate the results that were obtained with limited information. We show a home ground advantage exists in non-elite doubles tennis. A simple exponential smoothing method of rating players is then tested, and shown to produce reasonable results.

Keywords:

• sports
• tennis
• regression
• forecasting
• exponential smoothing
• home advantage

Acknowledgements

A short version of this paper was presented at the IMA Second International Conference on Mathematics in Sport at Groningen, 1990. Thanks to David Burn and The Eastern Melbourne Veterans Tennis Association for supplying scorecards.