Modern sports broadcasting employs numerous venue cameras to gather diverse data, which can be applied in sports analytics, using data science and AI to enhance athlete performanceAthlete performance. This paper introduces a tennis analyticsTennis analytics framework that simulates matches using data from broadcast videos. We extract granular shot-level information through deep-learning video analytics. Then, we apply probabilistic model checking (PMC)Probabilistic Model Checking (PMC) of Markov Decision ProcessesMarkov Decision Process (MDP) (MDP) for strategy analysis, which predicts players’ winning probabilities and suggests game-changing strategies. The framework is intuitive and applicable for tennis players across various skill levels. Extensive analysis of professional matches from the past decade validates our approach and demonstrates its effectiveness inStrategy optimization strategy optimization.