Researchers Use AI to Decode Rules of Ancient Roman Board Game

Summary

An international research team has successfully used artificial intelligence to determine the rules of an ancient board game from Roman Netherlands, marking the first time AI-driven simulated play has been combined with archaeological methods to identify a historical game. By analyzing wear patterns on a limestone game board discovered in Heerlen, researchers used the Ludii AI system to simulate hundreds of possible rule sets, determining that the artifact was likely a strategy game known as a "blocking game"—where players trap opponent pieces rather than capture them. The findings, published in the Antiquity journal, suggest that blocking games may have a much deeper history than previously documented, with evidence pointing to their existence in Roman times rather than just the Middle Ages. The successful approach demonstrates AI's transformative potential for archaeology, offering a new tool to unlock the secrets of mysterious artifacts that lack documentation in surviving texts or artworks.

Editorial Opinion

This groundbreaking application of AI to archaeology demonstrates the technology's value beyond commercial applications—using machine learning to bridge gaps in human understanding of history is a compelling use case. By systematically testing hundreds of rule combinations against physical evidence, the researchers have shown how AI can reveal insights that would be virtually impossible to determine through traditional methods alone. This work could inspire similar interdisciplinary approaches across archaeology and cultural heritage studies.