Video Game Story Analysis Using Layered Graphs and Eye Tracking System Iwona Grabska-Gradzińska Department of Games Technology Jagiellonian University Kraków, Poland e-mail: iwona.grabska@uj.edu.pl Abstract— This paper discusses the prospects of using layered graphs and eye-tracking system for describing and analyzing player activity and his/her decisions in video game stories. Understanding how the game interface affects the user’s decision is a very interesting problem, especially in the serious game field, because of real-life applications of acquired skills. Very often, the game winning strategies lead players to fall into bad habits. Reasoning based on the formal game system gives tools to game analysis and to collect information about players’ behavior for further analysis. Eye-tracking information, i.e., gaze plot and heat maps give us knowledge about user perception of the game screen and helps with answering the question: what has affected the user during the decision making process? The main goal of the paper is to show usefulness of the proposed model while answering the sample questions about players’ decisions. Keywords-game design; user decisions; narrations; eye tracking system; layered graph I. INTRODUCTION During the narrative game design process, especially for games with educational aim and real world dependencies and knowledge base, there are a few conditions that are extremely important. Scenarios and situations must be authentic and adequate and push the player to act. The scenarios should create the illusion of unlimited possibilities, but they should be precisely connected with the scope of knowledge and trained skills. It should be possible to replay and to explain the meaning of the winning strategies [1]. The comparison of user decisions in the game and in the real life is connected with both problems of the real life object representation in the game and the problem of the perception of the Heads-Up Display (HUD) elements. The challenge is how to prepare the “computable” model, i.e, the one which helps us compare different players’ activity as well as connect user physiological reactions (eye perception) with the decisions taken. The holistic model is based on the gameplay graph that represents the current state of the game. Different groups of elements can be identified: e.g., characters, locations, items and abstract narrative elements. These four categories reflect how players perceive games [2]. These elements could be identified in traditional narratives as well, but such formal models were not needed for their analysis. The usage of such model is connected with gameplay based on user’s decision while comparing user’s strategies. Our aim was to make all these elements uniform – so our model would process elements in the same way. Player actions and all other interactions in this world are denoted by connections between these elements [4]. Decision making process is correlated with elements available in the game, especially those noticed on the screen at the moment of decision making (others had to be seen prior) and player’s general knowledge. Especially in narrative and serious games, it is extremely important to examine why the user picks a certain game strategy and condition of decision process. Researchers can speculate on them using two elements of a given system: typical sequences of gameplay graph states and information from the eye-tracking system. A. State of the art Eye-tracking is more and more often used in research game studies, especially as a game controlling device. Many researches were conducted ease, immersion and player satisfaction while using eye-tracker as input device, e.g. as described in [16]. Eye-tracker as an evaluation tool was also used in research studies to investigate the visual search patterns and heap maps describing the main area of interest on the screen [18]. The most popular approach is to collect heat map data and take into consideration the player activity throughout the game. The approach presented in this paper is based on the dependence between the gaze direction immediately before decision. This approach is possible due to the graph of user strategy connected with the eye-tracker data. The gaze patterns and decision-making patterns can be associated in this approach. B. Research goals The research goal is to find a correlation between players eye-movement patterns and decision-making actions during the gameplay. During the work, the simple adventure game developed in the Department of Games Technology at Jagiellonian University was tested on a group of students and a question about their decision on choosing one of the three strategies was asked. The purpose of the research works is to show which elements of the user interface influenced the user’s decision and which are useless in this context. The idea of the graph representation of the game state is presented in Section II and formal definitions are explained 105 Copyright (c) IARIA, 2016. ISBN: 978-1-61208-468-8 ACHI 2016 : The Ninth International Conference on Advances in Computer-Human Interactions