Proposing a Model to Catch the Momentum of Games: Visualization of Momentum in Japanese Professional Baseball

Abstract

There are some researches about positive streaks (points scored by one team) lead participants to predict the streak’s continuation (belief in the hot hand or momentum effect), but negative streaks lead to prediction of its end (gambler’s fallacy). In basketball, the validity of “hot-hand fallacy” from successive shoot success was examined [1], and in volleyball, factors that influence momentum on players were explored by using survey data [2]. However, both papers have concluded that the existence of momentum is based on psychological or subjective randomness.

A purpose of this study is to examine within-game momentum in one particular game of Japanese professional baseball quantitatively. We assume that “streaks” in Japanese professional baseball based on superiority or inferiority of the game and the status of play itself which affect the result. Then, we propose a “momentum” model by using the 2015 Japanese professional baseball game data. Lastly, the created model is fitted against the game data of 2016, and the “momentum” of each team, competing against each inning, was calculated and verified. As a result, we could successfully predict a change of streaks of one particular game by the “momentum” model of the game.