This paper explores the evolution of machine learning (ML) applications in sports over the past decade, focusing on soccer, basketball, cricket, tennis, and Formula 1. By analyzing advancements in key technologies like the YOLO (You Only Look Once) object detection algorithm, the research highlights how ML has transitioned from basic predictive models to sophisticated, real-time analytics that optimize player performance, injury prevention, and strategic decision-making. The study tracks improvements in both mean Average Precision (mAP) and Frames Per Second (FPS) across YOLO versions, illustrating how these developments have increased the accuracy and efficiency of sports analysis. While the benefits of ML in sports are clear, the research also addresses limitations such as inconsistent data availability across sports, the lack of standardized benchmarks, and the challenges of real-time application. Nonetheless, the findings suggest that ML has become a crucial tool in modern sports and will continue to shape the future of sports analytics and performance monitoring .