Predicting Pseudo-Random Behaviour in Professional Sports

Truly random behaviour is difficult for humans to generate, and they often unconsciously fall back to repetitive sequences that only appear to be random. In sports, discovering these regularities could yield a substantial benefit, as athletes often try to exhibit random behaviour in order to prevent their opponents from foreseeing the planned strategy. This paper presents a case study where statistical and machine learning methods are employed in order to predict the serve placement of professional tennis players. Surprisingly, results show that professional tennis players manage to mix up their actions such that it is difficult to accurately predict the location of the next serve.