Predicting a Starting Pitcher in Baseball by Heuristic Rules

Baseball is one of the best popular sports in Japan. A large number of baseball spectators are much interested in various predictions related to the game, such as starting players and outcome of the games. In this paper, we propose a heuristic method for predicting a starting pitcher. Predicting a starting pitcher using computers is difficult as as various aspects need to be considered and the volume of information is limited. The accuracy of prediction is low even by ardent followers of baseball. Our proposed method is modeled on the human method of prediction. The preliminary evaluation results show that the proposed method attains a prediction ratio which is equal to or higher than that obtained by the human method.