AutomatedPlay Testing

Video game developers currently rely on manual playtesting to identify bugs, uncover game design flaws, and enhance player experience. Although highly efficacious, manual playtesting is time-consuming, labor-intensive, and variably reproducible. Consequently, automated playtesting is a research topic of growing interest, yet remains widely regarded as impractical due to the inherent complexity and nondeterminism of video games. We introduce Aplib.NET, an agent-based testing library for games written in C#, inspired by ongoing research on automated playtesting with the Belief-Desire-Intention (BDI) model. The Aplib.NET library is designed to make automated playtesting accessible to game developers of all skill levels. We demonstrate the usability of Aplib.NET through a video game we created and present empirical evidence that automated playtesting could address some of the limitations of manual playtesting. Most notably, we achieved a promising test coverage of over 80% using automated playtests alone.