A novel mechanical examination of the tennis serve using expert qualitative analysis and machine learning

Access to quality tennis coaching, often limited by cost or location, poses a significant challenge for players worldwide. Recent advancements in machine learning offer scalable and objective alternatives to traditional coaching methods yet remain largely unexplored in tennis. This thesis investigates the potential of machine learning as a supplementary tool for assessing the mechanics of tennis’ most critical shot, the serve. By analysing expert agreement in serve assessments and leveraging machine learning to evaluate serving mechanics from video footage, this research lays the groundwork for a virtual coaching tool, marking an important early step toward enhancing player development through technology.