ABSTRACT
Service has been recognised as an important part of Grand Slam tennis, especially at Wimbledon; O’Donoghue and Ingram (2001) found that there were more aces and serve winners played at Wimbledon than at any other Grand Slam tournament. Furlong (1995) found that more sets required tie breaks at Wimbledon than at the French Open indicating a greater importance of serve on grass courts. A mathematical model (Gale, 1971) has shown that the probability of the serving player winning a point in tennis, P, is given by
where p1 and p2 are the probabilities of the first and second serves being in respectively and ql and q2 are the conditional probabilities of the point being won given that the first and second serves are in respectively. The model considers serving strategy as a pair of serves to be used as the first and second serve respectively. The model still holds when pl and p2 retrospectively represent the proportion of first and second serves that were in, and ql and q2 retrospectively represent the proportion of points won when the first serve and second serve were in. Since Gale produced his model, it has been possible to measure the speed of the player’s first serve, V1, and second serve, V2, during matches. This is done on selected courts at Grand Slam tournaments using the IBM radar gun (Wimbledon, 2003). Two specially designed radar sensors are positioned behind the baseline of each end of the court. Once the serving player strikes the ball during service, the radar guns detect its speed almost instantaneously utilising Doppler radar technology. The purpose of the current investigation was to analyse the relationship between service speed and Gale’s model.
