Tuesday, December 19, 2023

Digital Ludeme Project: Modelling the Evolution of Traditional Games

 [First posted in AWOL 3 October 2019, updated 19 December 2023]

Digital Ludeme Project: Modelling the Evolution of Traditional Games

Modern Artificial Intelligence for Ancient Games
Cameron Browne

Thousands of games have left evidence of their existence scattered throughout history. They hide a wealth of information: for example, about the culture of the people that played them and about the cultural influences those players were subjected to. However, our knowledge about historical games is limited. What little we do know relies on interpretation. The Digital Ludeme Project aims to reconstruct the missing knowledge of ancient games using state-of-the-art artificial intelligence. The goal is to construct a ‘family tree’ of games, showing how games evolved and spread across the world.

Consider the Indian game of Pachisi and the Mexican game of Patolli. At first glance, they look remarkably similar: both game boards are shaped like a plus and some of the cells appear crossed out. Are they the same games? Do they share a common ancestor? While there is currently no way of knowing, the Asian and North-American games pre-date Columbus’ journey to the Americas in 1492. Columbus is typically credited with the discovery of a new continent, but the obvious similarities between Pachisi and Patolli raise some questions.

Pachisi (left) and Patolli (right).
Image credit: Daniel Schwen (right, CC BY-SA 4.0)

Of course, it could all be a coincidence. Pachisi and Patolli may not be the same game, and even if they are, they could have arisen independently. There is no way of telling without knowing the details of these games. For example, the boards used to play chess and checkers appear alike, but the games are far from identical. Conversely, two chess sets may look different while they are still used to play the same game. We can say this only because we know chess and checkers well.

Reconstructing Lost Knowledge with Ludii

This is not the case for the vast majority of ancient games. Scholars learned of them through artefacts hidden away in tombs, travel diaries, boards carved out in rocks or hieroglyphs, for example. Based on such limited information, it is impossible to determine whether games are related, and where and when those relations came about. The Digital Ludeme Project aims to fill these gaps in historical knowledge. To do this, we are developing an advanced platform that uses artificial intelligence to reconstruct approximately one thousand historical games.

   Project   Games   Ludii   Outputs   People

No comments:

Post a Comment