Authors:
Eric Piette, Matthew Stephenson, Dennis J. N. J. Soemers, Cameron Browne

Venue:
IEEE Conference on Games (CoG), 2019

Pages:
626–629

DOI:
10.1109/CIG.2019.8847994

Topics:
General Game Playing, Ludii, RBG, game description languages, efficiency, readability

Links: PDF · IEEE / ACM entry · arXiv

Although General Game Playing systems are useful for AI research, they are often computationally inefficient and specialised to narrow classes of games.

This paper presents an experimental evaluation of Ludii, focusing mainly on a comparison with the Regular Boardgames language (RBG) in terms of two key properties for general game systems: simplicity and clarity of game descriptions, and reasoning efficiency.

The results show that Ludii generally uses fewer tokens to describe games and often achieves higher playout rates than RBG, especially compared with the RBG interpreter, while also offering strong human-readability through its ludemic approach.

This paper is important because it provides one of the first direct empirical comparisons between two modern alternatives to GDL: Ludii and RBG.

Beyond raw performance, it also addresses a crucial issue for general game systems: whether game descriptions are simple and readable enough to be useful outside a narrow expert community.

In that sense, the paper helps position Ludii not only as an efficient platform for AI research, but also as a more accessible framework for game design, historical modelling, and broader interdisciplinary work.

Piette, E., Stephenson, M., Soemers, D. J. N. J., Browne, C. (2019). An Empirical Evaluation of Two General Game Systems: Ludii and RBG. In Proceedings of the IEEE Conference on Games (CoG), pp. 626–629. DOI: 10.1109/CIG.2019.8847994

@inproceedings{piette2019ludii_rbg,
  author    = {Piette, Eric and Stephenson, Matthew and Soemers, Dennis J. N. J. and Browne, Cameron},
  title     = {An Empirical Evaluation of Two General Game Systems: Ludii and RBG},
  booktitle = {Proceedings of the IEEE Conference on Games (CoG)},
  pages     = {626--629},
  year      = {2019},
  doi       = {10.1109/CIG.2019.8847994},
  url       = {https://arxiv.org/abs/1907.00244}
}