A stochastic approach to predicting competition squash match-play.

Modelling sport competition is an informative analytic technique for the investigation of athletic behaviour because it focuses attention on the critical aspects of data which characterize sporting success. A stochastic (Markov) model of championship squash match-play is reported which infers prospective behaviour from previous sport performance through forecasting shot response and associated outcome from the preceding shot. The results presently restrict the model to the competing players already analysed because, while a consistent athletic response was established when competing against the same opponent (P > 0.25), an inconsistent response was found when competing against different opponents (P < 0.25). These results may impact the information value traditionally drawn from previous sport analysis. The hypothesis that future squash match-play can be successfully predicted from previous sport performance requires further investigation, since it remains unclear whether the present analysis failed to identify the preceding condition(s) which elicit a reasonably stable behavioural response, or indeed whether athletic behaviour in championship squash is more variant than initially believed. Alternative modelling considerations are discussed and a refinement to the present model is proposed in an effort to improve future model utility and application.