Decision-making under pressure: a study of tennis professionals

Abstract

This thesis examines the effect of pressure on a sample of highly-trained individuals in a simple strategic setting. By modelling the strategic interaction between servers and receivers in professional tennis matches, the main contribution of this research is to present an environment where pressure can be introduced in a practical and reasonable manner. Using this environment, I am able to investigate the effect that pressure has on players' behaviour and its consequences for their payoffs. Thus, the analysis in this thesis is two-fold. The first component of the analysis examines the effect that pressure has on the decision-making ability of players. The second component examines the effect that pressure has on their point outcomes.

The results indicate that pressure does have a marked effect on the service decisions of many players in the sample. I find that a significant number of servers have a particular strategy that they choose to play more often under pressure than they do otherwise. I also investigate the effect of pressure on the level of correlation between past and present choices and find similar results. The implication of this is that servers' behaviour changes when they are faced with high pressure situations.

Since players are also found to behave in accordance with theoretical predictions in the absence of pressure, this leads to the possibility that players could be performing sub-optimally in such situations. Indeed, the results confirm that pressure has a generally negative effect on servers' chances of winning the point. Interestingly, there is less evidence for a link between the two effects. The correlation between the players whose choices change under pressure and those players whose outcomes are affected is small. This leads me to believe that there is a possibility that either player could benefit from exploiting the sub-optimal choices of the other in high pressure scenarios. These results suggest that the theoretical models that have been previously used to describe the serve-return interaction may be insufficient in the presence of pressure. Consequently, researchers may be led to incorrect predictions about behaviour in high-stakes environments by excluding pressure from their models. Therefore, I also provide some theoretical basis to rationalise the observed results by proposing several approaches to incorporate pressure into existing models.