Year: 2025
Type: conference-paper
Abstract: Emergent Narratives (EN) are a popular approach in game design where stories naturally emerge from players' interactions with the game world.Story sifting is an EN technique where the player's attention is drawn to particularly interesting narrative patterns, but these story sifting patterns can be complex and difficult to create.To address this, we have developed an event-based rules system that supports composable patterns: building blocks that allow higher-level stories to be constructed from lower level ones.We demonstrate our approach through a simulation (of an alien invasion scenario) alongside a powerful incremental story sifter that curates compound events.We conduct agent ratio and stability testing to demonstrate the system's robustness, and a complexity test to show how the performance of the sifter scales with agent numbers and map size.Our findings show that the system remains stable throughout.Our work demonstrates that an event-based rules approach can effectively support composite patterns, ultimately enabling the curation of more complex EN stories.
Authors: Qianwen Lyu, David Millard, Nicholas Gibbins
Institution: University of Southampton
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4410237378