*-Minimax Performance in Backgammon
- Thomas Hauk
- Michael Buro
- Jonathan Schaeffer, Department of Computing Science, University of Alberta
This paper presents the rst performance results for Ballard's *-Minimax algorithms applied to a real{world domain: backgammon. It is shown that with eective move ordering and probing the Star2 algorithm considerably outperforms Expectimax. Star2 allows strong backgammon programs to conduct depth 5 full-width searches (up from 3) under tournament conditions on regular hardware without using risky forward pruning techniques.We also present empirical evidence that with today's sophisticated evaluation functions good checker play in backgammon does not require deep searches.
Citation
T. Hauk, M. Buro, J. Schaeffer. "*-Minimax Performance in Backgammon". Computers and Games, pp 51-66, June 2004.Keywords: | |
Category: | In Conference |
BibTeX
@incollection{Hauk+al:04, author = {Thomas Hauk and Michael Buro and Jonathan Schaeffer}, title = {*-Minimax Performance in Backgammon}, Pages = {51-66}, booktitle = {Computers and Games}, year = 2004, }Last Updated: June 05, 2007
Submitted by Staurt H. Johnson