Estimating bowling workload across maximal and submaximal intensities in male cricket fast bowlers

Abstract

It is commonly accepted in the literature that there is an association between workload and injury in fast bowlers. While certain workload values have been more strongly linked to injury than others, the relationship between workload and injury is often only examined in terms of the number of deliveries bowled (i.e., external workload) with no consideration to the intensity at which the deliveries are bowled. The primary aim of this thesis was to examine ways to improve the estimation of workload in fast bowlers across a range of intensities. Initially, a systematic review found there was little strength to any association between external workload and lower-back injuries, primarily due to a high risk of bias and general low quality of evidence among the included studies. The findings demonstrate that it is not sufficient to only examine external workload, as all deliveries cannot be viewed the same in terms of the load they place on the body, particularly because bowlers work across a range of intensities. One important factor for estimating workload is the intra- and inter-individual variability in the bowling action, highlighted in Chapter Three, with the magnitude of variability affecting the cumulative loading that results from repetitive bowling. As well as the implications the results of this study had for the estimation of workload, the results also influenced the design and analysis of the subsequent studies. For instance, individual analyses of fast bowlers were used because of the high inter-individual variability across the group, and repeated sessions were used because of the intra- individual variability in bowlers across two bowling sessions. The data for all of the ensuing studies were collected at the same time and allowed two methods of workload estimation to be examined. The first method involved measuring an internal workload variable over a period of time, such as an over or bowling session, and multiplying this by an accepted external workload measure such as the number of deliveries bowled over the same time period. Heart rate and rate of perceived exertion (RPE) were investigated as potential internal workload measures, by examining their association with release speed across maximal and submaximal intensities. It was found that either heart rate or RPE could be used to help estimate workload in fast bowlers, but both require considerations prior to their use. Primarily, around how to deal with the different movement patterns that are likely to be displayed by bowlers in trainings, warm-ups, and games, and how the estimation of workload will be affected. The second workload-estimation method involved quantifying each delivery using a workload variable that is sensitive to intensity. The simplest way this can be done is by examining release speed, but the lack of automation and issues around data collection in large group sessions means that other approaches are more attractive, such as using IMUs. The most common workload variable measured using IMUs is PlayerLoad which was broadly investigated in terms of its association with release speed, again, across maximal and submaximal intensities. Two calculations for PlayerLoad were compared - the maximum and accumulated value from each delivery, as were two IMU locations, the upper-back and non-bowling wrist. The results indicated that accumulated PlayerLoad had a stronger association with release speed, with comparable findings between the two IMU locations. There is also the potential that the non-bowling wrist location could be more practical and help to differentiate between bowling and throwing events, which is an important consideration if IMUs are to be used for workload estimation in fast bowlers. Another consideration when using IMUs to estimate workload is how well the timing of key events in the bowling action can be determined from IMU data, this was also examined. The main finding was that the lack of consistency in the timing of these minimum/maximum values across maximal and submaximal intensities, indicating that different bowlers may alter their technique in alternate ways when bowling deliveries at submaximal intensities. In the final study, clustering using self-organising maps (SOM) allowed a visualisation of this technique change and a quantification of the change was shown by calculating the reliable change index (RCI). The results from the final study highlight the fact that not enough is known about how fast bowlers change their technique when bowling at submaximal intensities. Before workload can be hoped to be estimated across a range of intensities, it should first be understood how each bowler, or groups of similar bowlers, are changing their technique. Once this is understood, workload variables should be able to be determined that can accurately estimate the load that is placed on the body by each delivery.