Abstract:

Nowadays, video games became part of the culture of people of all ages and their popularity grows every day. They are amusing and engaging, and besides entertainment, they are also used in various fields such as medicine, engineering, education, etc. to acquire new knowledge and skills. The main goal of video games in these areas is to create gameplay scenarios for learning new knowledge and training skills based on tracking player progress, i.e. level of performance, acquired knowledge, or skills. One of the methods for player progress monitoring is using of learning curves. A learning curve presents the course of progress made in learning new skills and knowledge over time. In video games, the learning curve illustrates the specific performance to an individual player in acquiring of new playing skills, cognitive abilities, and knowledge to solve the challenges in the game, usually in a restricted time. Detection of the player learning curve is very important for the implementation of an effective learning process through playing a game because it enables to identify the current state of a player and, hence, what activities (gameplay scenarios with different difficulties, learning exercises, etc.) are appropriate for him/her. A learning curve is different for each player. This fact implies developing an adaptive gameplay scenario depending on it that adjusting of game difficulty and challenges in order to retain player motivation, engagement, and efficiency of the learning process.

In this study, we present a method for dynamic game adaptation based on detection of behavior patterns in the player learning curve and, as well, an experimental study that evaluates its effectiveness. The adaptation according to this approach was accomplished using a game component named “Player-centric rule-and-pattern-based adaptation asset” and developed in the scope of the RAGE (Realising and Applied Gaming Ecosystem) H2020 project. The component is integrated within a car driving video game and allows dynamic registration and detection of behaviour patterns of specific player learning curves representing overall player performance over time (OPP). When a pattern behavior in the player learning curve is identified a pre-defined event is triggered and weather conditions (level of illumination, fog, rain, etc.) in the game are changed. The goal of weather changes in the game is to provide the most appropriate gameplay scenarios in order to improve driving skills of the player depending on his/her current OPP.

A practical experiment was conducted with students from Sofia University, Bulgaria, in order to evaluate the effect of a dynamic adaptation of the car driving game according to pre-defined behaviour patterns that are detected at runtime by the RAGE game component. The experimental study consisted of two game sessions. In the first game session students played a car driving video game without adaptation, but in the second game session they played the same game but using the RAGE asset for adaptation. After the two game sessions, participants completed a post-game questionnaire about their gaming experience. Results of conducted experiment demonstrate preferences of students to the car driving game using the RAGE game asset for detection of behavioral patterns in the player learning curve. According to students’ opinion, the adapted version of the game is more immersive, engaging and effective.

Keywords:

Game adaptation, player learning curve, detection, behavior pattern, game component, RAGE.