Dynamic Game Balancing (DGB) is the process of real-time adjustment of game parameters so that the faced challenges fit the player’s ability, therefore keeping him/her in the flow. This way the player will not be bored (if the game is too easy) or anxious (if it is too hard) and will remain motivated to play the game. DGB provides an individualized approach to a game and replaces a single, common approach for all the players. Andrade et al. (2005) present two dimensions of DGB parameter application: competence vs. performance, that is, the understanding and mastering of the game vs the capacity to efficiently tackle its challenges.
Game analytics (GA), the set of methods designed to collect and interpret game data, is mostly related to game play metrics, that is, information about the actual behavior of the user as a player inside the game: object interaction, object trade, navigation in the environment, actions and position of the player’s character, results in each level, time spent, interactions with the game interface and menus, etc. GA feeds information to Procedural Content Generation (PCG) processes (presented in this blog series by A. Coelho), that is, the possibility of real-time generation of the game challenges based on a set of parameters instead of a complete, predetermined and fixed progress route. Finally, DGB requires the definition of a Player Model, constructed by the collection of the pregame and real-time game data through predefined rules.
A few technical approaches for DGB have been used: Andrade (2005) presented an approach where agents were trained to play against the human player at his/her skill level; Demasi (2002) used genetic algorithms techniques to define the behavior of agents that best fit the user level; Yannakakis (2004) used artificial neural networks (ANN) and fuzzy neural networks to estimate the parameters that provide engaging game play.
For Serious Games, DGB must include a component related to the serious objectives of the game (have learning outcomes been achieved in an educational game? has the advertising message caused an impact in an advergame?). The entertainment aspect of the game cannot hide the skill or competence development objective and therefore must be tuned to include this concern. Game analytics should, at the same time, provide data that allows assessing how the player is progressing towards those serious goals.
A Serious Player Model (SPM) based on the User Model of Adaptive Hypermedia Systems has to be created by incorporating all the relevant parameters of use. “Player modeling is, primarily, the study and use of artificial and computational intelligence (AI and CI) techniques for the construction of computational models of player behavior, cognition and emotion, as well as other aspects beyond their interaction with a game (such as their personality and cultural background)” (Yannakakis, 2013). The SPM extends this model into the player characteristics adequate for the serious game purpose. For instance, for an exergame the player model should incorporate physical parameters of the user, for educational games the SPM has to include parameters related to the knowledge, skills and competences pre and post serious game usage.
SPM-based DSGB with PCG can create highly motivating, adaptive and personalized skill and competence development environments (but still games) that keep users involved for a long time therefore ensuring that they are focused on the “serious” objective. However, further research is required to establish a taxonomic approach to Serious Games that delimits the DSGB methods and parameters adequate to each area of application. Serious games with educational purposes are quite different from serious games for marketing purposes, for instance. So, although it is possible to create a methodology that addresses the use of DSGB for serious games in general, there is a need to design and develop specific GA, SPM and PCG methods and tools for each area of application. Therefore finding these specificities is a state of the art research.
References
- Andrade G., Ramalho G., Santana H., Corruble V.: Challenge Sensitive Action Selection: an Application to Game Balancing. Proceedings of the IEEE/WIC/ACM International Conference on Intelligent Agent Technology (IAT05). Compiègne, France: IEEE Computer Society. pp. 194–200 (2005)
- Demasi P., Cruz A.: Online Coevolution for Action Games. Proceedings of The 3rd International Conference on Intelligent Games And Simulation. London. pp. 113–120 (2002)
- Yannakakis G.M., Hallam J.: Evolving Opponents for Interesting Interactive Computer Games. Proceedings of the 8th International Conference on the Simulation of Adaptive Behavior (SAB’04); From Animals to Animats 8. Los Angeles, California, United States: The MIT Press. pp. 499–508 (2004)
- Yannakakis G., Spronck P., Loiacono D. and Andre E.: Player Modelling, Dagstuhl Publishing, Schloss Dagstuhl – LeibnizZentrum für Informatik, Germany (2013)
