All according to plan : generating mystery game narratives

Abstract

Plot generation in video games is considered quite challenging as although an algorithm would be able to piece several plotpoints and plot elements together, it might not necessarily create a story with a legible and understandable flow and structure. The purpose of this thesis is to develop a narrative generating algorithm that would specialize in the creation of mystery narratives, with an emphasis on consistency as well as the surprise element. A focus on mysteries has been taken due to the genre’s extra difficulty in generating a believable narrative, as the story would have to have a logical flow of events as well as be populated with characters that have realistic motivations for them to commit the crime determined in the mystery. Since there are many different types of possible mysteries, the generator will specify further and only attempt to generate murder mysteries, or ’whodunnit’ stories. In order to demonstrate the algorithm, it shall be developed and implemented into a first person 3D mystery game in Unity where the player is tasked with discovering and solving the mystery that had been generated by the algorithm. The algorithm and game would be evaluated through experimentation as well as with surveys.