SimuExplorer: Visual Exploration of Game Simulation in Table Tennis.

We propose SimuExplorer, a visualization system to help analysts explore how player behavior impacts scoring rates in table tennis. Such analysis is indispensable for analysts and coaches, who aim to formulate training plans that can help players improve. However, it is challenging to identify the impacts of individual behaviors, as well as to understand how these impacts are generated and accumulated gradually over the course of a game. To address these challenges, we worked closely with domain analysts who used to work for a top national table tennis team to design SimuExplorer. The SimuExplorer system integrates a Markov chain model to simulate individual and cumulative impacts of particular behaviors. It then provides flow and matrix views to help users visualize and interpret these impacts. We demonstrate the usefulness of the system with three case studies. The domain analysts think highly of the system and have identified insights using it.