Evaluating AI-Based Games through Retellings


  • Max Kreminski University of California, Santa Cruz
  • Ben Samuel University of New Orleans
  • Edward Melcer University of California, Santa Cruz
  • Noah Wardrip-Fruin University of California, Santa Cruz


We propose a new approach to the human-centered evaluation of AI-based games, grounded in the analysis of player retellings of their play experiences. Retellings offer unique insight into dimensions of player experience that can be hard to get at through existing evaluation methods, such as the typical narrative structures that tend to emerge in the player’s mind when they play a particular game; the variety of subjectively experienced narratives that are possible and probable within a particular game; and the ways in which a game supports, or fails to support, the player’s process of narrativization. We used a grounded theory methodology to analyze retellings of play experiences in Civilization VI, Stellaris, and two distinct versions of the research game Prom Week. We also interviewed the creators of several retellings to gain insight into the subjective experience of story construction in collaboration with these games.




How to Cite

Kreminski, M., Samuel, B., Melcer, E., & Wardrip-Fruin, N. (2019). Evaluating AI-Based Games through Retellings. Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, 15(1), 45-51. Retrieved from https://ojs.aaai.org/index.php/AIIDE/article/view/5223