Automated Storytelling in Sports: A Rich Domain to Be Explored

Abstract

Sports broadcasting typically involves a play-by-play commentator and a color commentator. The color commentator’s function is to entertain the viewer during the game. The prevalent way of doing so is by telling brief stories relevant to the game in progress. In this paper, we propose that storytelling in sports is a challenging and rich problem for Artificial Intelligence (AI) research for the following reasons. First, storytelling is considered to be a very “human” activity as it requires engaging the audience in a story and appealing to their factual as well as emotional sides. Thus, automating it will be an advance in AI. Second, while automated storytelling in general has received ample attention, storytelling in sports has seen much less research. Third, an AI storyteller can be used as an assistant to a human color commentator in a real game, or autonomously in the context of a sports video game. Fourth, the task is non-trivial as we demonstrate by applying off-the-shelf machine learning methods in an attempt to match baseball stories to baseball game states.