Using Evolutionary Algorithms to Target Complexity Levels in Game Economies

Abstract

Game economies (GEs) describe how resources in games are created, transformed, or exchanged: They underpin most games and exist in different complexities. Their complexity may directly impact player difficulty. Nevertheless, neither difficulty nor complexity adjustment has been explored for GEs. Moreover, there is a lack of knowledge about complexity in GEs, how to define or assess it, and how it can be employed by automated adjustment approaches in game development to target specific complexity. We present a proof-of-concept for using evolutionary algorithms to craft targeted complexity graphs to model GEs. In a technical evaluation, we tested our first working definition of complexity in GEs. We then evaluated player-perceived complexity in a city-building game prototype through a user study and confirmed the generated GEs' complexity in an online survey. Our approach toward reliably creating GEs of specific complexity can facilitate game development and player testing but also inform and ground research on player perception of GE complexity.