Novel experimental protocol to capture movement data and predict shot execution in cricket batting

Shot execution in cricket batting is reliant on intricate movement patterns of crucial body segments. When there is a substantial amount of batting movement data available, supervised machine learning can be used to classify when a batting shot execution takes place in a cricket batting cycle. An automated approach to identify and assess cricket batting could be useful for the applications including performance evaluation, talent identification and injury prevention. Current evaluation of movements and shot execution are generally undertaken in an artificial environment with camera-based, motion tracking systems to collect batting movement data, which require careful preparation, data collection and post-processing, and risk changing the natural gameplay of a batsman. By training a model based on data obtained from a close representation of a cricket batting innings, supervised machine learning was found to be capable of reliably predicting cricket batting shot execution.