1 Sofia University (BULGARIA)
2 FMI at Sofia University (BULGARIA)
About this paper:
Appears in: INTED2019 Proceedings
Publication year: 2019
Pages: 9845-9853
ISBN: 978-84-09-08619-1
ISSN: 2340-1079
doi: 10.21125/inted.2019.2448
Conference name: 13th International Technology, Education and Development Conference
Dates: 11-13 March, 2019
Location: Valencia, Spain
Nowadays, video games have become an integral part of our culture and everyday life and are one of the most popular forms not only for entertainment but to a great extent for training and other applied purposes. As in both fun games and serious games, one of the important and still unresolved issues is the effective adaptation of different game features to the individual player's model. This model may include indicators of cognitive abilities and skills, the psycho-emotional state and the player's style of behavior, such as learning or playing styles. Within the APOGEE ( research project, adaptive video games are developed to target these three types of player metrics.

The presented article aims to determine the effect of adapting dynamically different video game features solely to the player's cognitive abilities and skills measured at a given level of play. For an experimental study exploring the methods of adapting video games to the outcomes achieved by the player, two genres of three-dimensional video games were selected - a simulator (car driving game) and a first-person shooter on moving objects. Both types of video games were developed in three versions:
1. The first version of each game does not use any adaptation methods;
2. The second version uses the classic method of dynamic adaptation at achieved levels (thresholds) of the result, which changes the dynamics of the game and hence the difficulty of the task – e.g., by changing the environment features like fog, rain, darkness, sleepiness, etc.;
3. The latest version uses the Dynamic Adaptation Method to detect patterns in the learning curve. In order to automatically recognize a learner's learning curve by detecting a behavioral pattern of player's performance change, that adaptive video game uses a Player-Centric Rule-and-Pattern-Based Adaptation Component created in the scope of the European project RAGE ( Its integration into the game helps to dynamically detect specific learning curves of the player, representing his/her overall performance for a certain time.

A practical experiment was conducted which included a preliminary study of the characteristics of patterns in the individual learning curves, after which they were defined in the component. Next, the experiment continued with a dynamic recognition of these patterns during play.

To validate the qualities of the three versions of the same video game, the games were distributed to randomly selected individuals for the purpose of organizing individual game sessions. Three sets of experiments were conducted - with non-adaptive versions of the games and with two types of adaptation: based on threshold values and on learning curves. Each player who completed a game session was interviewed online to determine how effective and efficient was the dynamic adaptation. The quiz included 46 questions and was mainly based on the Game Experience Questionnaire (GEQ) questionnaire, which assessed the level of skill, immersion, flow, tension, challenge, and positive or negative impact felt by the player. The article presents a part of the results of the survey, as well as the effects of the dynamic adaptation of a specific game to the player's achievements. The results conclude that such dynamic adaptation leads to a statistically significant improvement in both the players' achievements and the gaming experience, learning, efficiency, immersion, motivation, and emotion.
Video games, adaptation, player model, pattern, dynamic.