Modelling Progression of Competitive Sport Performance

• Main Author: Malcata, Rita Maria (Author)
• Other Authors: Hopkins, Will G (Contributor), Pearson, Simon (Contributor), Vandenbogaerde, Tom (Contributor)
• Format: Others
• Published: Auckland University of Technology, 2014-07-18T03:47:30Z.
• Subjects: Tracking; Monitoring; Linear mixed model; Competition results; Variability; Career trajectories; Country performance; Olympic ranking; Triathlon; Swimming; Football; Olympic sports; Thesis
• Online Access: Get fulltext

 Athletes, coaches, sport scientists and managers need objective assessment of changes in competitive performance to provide evidence for guiding athletes' development, for assessing programme effectiveness, and for supporting decisions regarding allocation of funds in sports campaigns. This PhD is focused on the development of analytical tools using the Statistical Analysis System (SAS) software for assessing changes in competitive performance. The topic of variability of competitive performance is reviewed first, because estimates of variability provide thresholds of magnitude for assessing important changes in performance. Five original-research studies are then presented for assessing performance changes in five levels of performance: athlete, sport, team, country squad and all Olympic sports of a country. First, mixed linear modelling was used to develop individual career trajectories of triathletes while accounting for environmental and other external factors. This analytical tool allows evaluation and comparison of athletes against the typical performance progression of successful elite triathletes. Secondly, linear performance trends with calendar year were evaluated using a mixed modelling approach to investigate progression of mean performance times for the sport of triathlon providing coaches and support staff with the current state of the sport. Thirdly, improvement of a football team's performance was quantified using generalised mixed linear modelling to assess the effectiveness of a youth-talent development programme. The focus of the fourth investigation was the development of a country score to provide a more comprehensive measure of performance than measures based on medal counts. These scores were derived by properly combining each country's athletes' world rankings. Finally, performance progression of individual athletes and teams over an Olympic quadrennium was assessed using linear regression of athletes' placings at annual main competitions. The analysis also provided a measure to evaluate under- and over-achievement at Olympics. In this thesis, general and generalized linear and mixed linear models proved to be appropriate for modelling changes in sport competitive performance. Further investigation is required to extend the models presented here to other sports and to explore non-linear models for analysis of competitive performance.