We quantify the nature and frequency of anticipation behaviours in professional tennis using video coding of incidents where the time delay between the opponent's stroke and the reaction of the player were recorded. We argue that anticipation is based on uncertain information and should lead in some situations to erroneous decisions. We identified the transition between reaction (with 100% accuracy in the selection of where the ball is played on the court) and anticipation (with less than 100% accuracy) as being 140-160 ms after ball contact. Anticipation behaviours occurred on between 6.14% and 13.42% of the coded situations. These anticipation behaviours appeared almost exclusively in 'unfavourable' situations, where the opponent had a significant tactical advantage, with the type of playing surface having only a limited effect. Moreover, the decrease in accuracy with shorter response times is not monotonic, with an increase in response accuracy being observed for times shorter than 120 ms before ball contact. We propose that very early anticipation behaviours occur when players use significant context-specific information before the opponent's stroke. When such information is not available, players produce anticipation behaviours that are closer to the moment of ball-racket contact using information that is more likely to be based on the opponent's preparation of the stroke. This study opens new directions for research focusing on the testing and training of anticipation in fast ball sports.

No abstract

We compared the performance of tennis experts and non-experts using a simulated interceptive task, in which the ball could be unexpectedly deviated 400 ms before contact. The results showed that experts were more accurate than non-experts when intercepting balls that deviated in their trajectory and that this could be explained by their shorter visuo-motor delay in adapting their interceptive movement. In addition, multiple regression analyses revealed that visuo-motor delay was a good predictor of accuracy in this task. Finally, accuracy in the simulated interceptive task was shown to be a reasonable predictor of expertise in tennis assessed by national ranking. In combination, the present results suggest that an important component of expertise in interceptive skills is fast information-movement coupling, which corresponds to a reduced delay in integrating vision and action. Our findings highlight the potential of the virtual interceptive task used here to predict performance in tennis.