Training System for Learning Tactics from E-sports Playing Video Based on Explanations

Abstract

In e-sports, the play videos of skilled players are good learning materials for beginners who want to acquire and learn tactics. Tactics are a series of actions that are taken to accomplish various objectives in response to a situation. When learning tactics from videos, players need to select meaningful actions. However, discriminating whether actions are meaningful for forming tactics is difficult. Our research supports players who want to acquire the skills of identifying meaningful actions from playing videos because such actions have causal relations with each other. If players can explain the causal relations of the extracted actions, this ability suggests that they have indeed extracted appropriate actions. This paper constructs a system that gives a playing video that contains a scene with important tactics and provides an environment in which players can explain them in the video. It also gives feedback about the appropriateness of the extracted actions.