next up previous contents
Next: Thesis Contributions Up: Introduction Previous: Introduction

Why Poker?

In recent years, the primary focus of games researchers has been placed on algorithms to solve games with perfect information. As a result, high-performance systems have been developed for games such as chess, Othello, and checkers. In many of these games, high performance can be achieved by brute-force search. Many advances in computer science have resulted from this research, including developments in problem specification, exhaustive search techniques, parallel algorithms, and others [Marsland and SchaefferMarsland and Schaeffer1990].

More interesting are games of imperfect information, where certain important information is not available or is unreliable. Computer science has been based on the processing of perfect information and the study of games like poker and bridge, where searching does not appear to be the key to success, could be highly valuable to all areas of computer science, particularly the field of artificial intelligence. Since these games offer different algorithmic and conceptual challenges, the successful development of a program capable of playing them well may provide solutions for problems that exist in computer science today. Poker can also be used to design and analyse situations where multiple interacting agents having competing goals. Since real life contains many similar situations, a method to solve a game of poker may be applied to problems in other areas.

Specific features of poker that are not as prominent in other games include risk management, the necessity to bluff and to predict deception, the implications of multiple opponents, and the importance of inference and prediction of the styles of other players in order to exploit their weaknesses. These characteristics are also present in many real-world applications that require rational behaviour.

Poker is ideal for testing automated reasoning under uncertainty. It is an example of a game of incomplete information in which chance plays a role. There is uncertainty introduced both by the physical randomness caused by shuffling, and through hidden information (the opponent's cards). Another source of uncertainty is the limited information available to construct psychological models of opponents, their tendencies to bluff, play conservatively, reveal weakness, etc. and the relation between their hand strength and betting behaviour.

These points form the focus of this thesis. To maximise practical results, it is essential to perform well in all aspects of the game. All of these uncertainties must be assessed accurately and combined effectively for any reasonable level of skill in the game to be achieved, since good decision making is highly sensitive to those tasks. Opponent modeling increases the ability to exploit weaknesses in an opponent's play and employ deception techniques (bluffing), while good risk-management techniques form the basis of an effective betting strategy.


next up previous contents
Next: Thesis Contributions Up: Introduction Previous: Introduction
Jason R Carlton
2000-11-13