Translating Elementary Cellular Automata into Tangible Gameplay: Design of Shellmate

Abstract

Computational thinking is increasingly being recognized as a critical cognitive skill beyond computer science education (Wing, 2006) (Shute et al., 2017). While game-based approaches to fostering computational thinking have been explored, much of the existing work focuses on instructional tools (Wu et al., 2025) (Wang et al., 2023). There is limited research on how abstract computational models can inform the design of non-digital strategy games. The aspects of Elementary Cellular Automata were analysed to identify transferable principles such as local rules, generational growth, and emergent pattern formation. This paper presents a case study of the design and development of Shellmate, a two-player strategy-based board game derived from the principles of Elementary Cellular Automata (ECA). The study documents how key ECA concepts, including neighbourhood interactions, rule-based state transitions, and generational progression, were translated into game elements through an iterative design process. Design decisions were guided by repeated prototyping and playtesting, allowing mechanics to be refined while preserving the underlying computational principles. Rule combinations and edge cases were explored through repeated rapid prototyping before arriving at the final design. The final playtesting was conducted with 10 participants in pairs over two rounds each, where participants showed progressive improvement in rule card selection and strategic planning, with blocking and diagonal anticipation emerging as dominant behaviours. A central finding from the design process is that strict adherence to ECA simulation rules proved incompatible with playable game design. Leaving some neighbourhood combinations undefined within each round, while departing from what a complete ECA would require, was a necessary decision that introduced uncertainty and created scope for player agency. Playtesting showed that this structure naturally prompted computational thinking behaviours such as pattern anticipation, forward planning, and adaptive reasoning, without explicitly instructing players in the underlying model. Participants gradually shifted to anticipating longer-term consequences of their actions, suggesting that the evolving game state encouraged strategic reasoning through repeated play. The paper demonstrates that ECA as a computational model can serve as a structural foundation for game design, and that the translation process explored here could be helpful for gamifying simple computational models. The resulting design process highlights how abstraction and selective implementation can preserve the underlying logic of a computational model while leaving space for player agency and playability. This approach has potential implications for computational thinking education, where physical game systems can engage learners with computational concepts without direct instruction. Notably, leaving rule combinations partially undefined proved not to be a design compromise but a mechanism that actively drives engagement and reasoning.