Bridge Bidding Website

Background

This is a research project concerned with applying Artificial Intelligence (AI) techniques for search and planning under incomplete information. The example we will use in this research is the game of bridge. In contrast to games like chess where all the information is known, in games like bridge the players operate under conditions of incomplete information. For example, in the bidding phase of the game, the players have incomplete information of the cards in the hands of the other three players; and in the playing phase, of the cards in the hands of the two defending players. This project should be of interest to game players, and especially to bridge players.

Bridge bidding strategies and optimization will be investigated through a bridge bidding simulator. The input to the simulator is a set of thirteen cards. The simulator will evaluate this set of cards and interact with other independent simulator processes or human players to make bids that will produce an optimal bridge contract. Bridge is an interesting game to analyze because it is played by four separate entities. Each entity plays with another entity in a partnership. Since each entity only knows 13 cards (a hand) of the total set of cards (52), it is required to communicate with the other entities to find out information about the other 39 cards. This communication is done via an artificial language called a bidding system.

Project

The project team will further develop a Bridge Bidding Website that will be used by Bridge players to practice, build partnerships, and develop their bidding systems. At present there are four major screens for the website:
  1. Bridge Filter Rules - this screen enters in filter rules that are used to select a specified type of hand
  2. Bridge Setup - this screen allows the user to setup a set of hands. The number of hands in a set is a parameter with default value of 10. The setup includes naming the set, the players who will play (West, North, East, and South), and which filter rules will be used to select the hands. Possible players include members of a bridge team and two special types of players, PASS and COMP. PASS passes all the time regardless of the type of hand. COMP will use a set of computer rules but is not yet implemented.
  3. Bridge Bidding - this screen allows the player to make all the bids for a set of hands when it is the player's turn to bid. The player can make an alternate bid and a comment on the bids. The set of hands is randomized when passed to the next player to prevent memorization of the hands.
  4. Bridge Display Results - after the set of the hands have been bid, the hands will be displayed. Also, the hands will go through two scoring routines to determine the score of the hands: genetic algorithm and comparison with hands played over the internet on OKBridge. (For research purposes, 7-card hands will also be evaluated by genetic algorithms and a depth-first analysis that is too slow for 13-card hands.)

The project will proceed as follows:

  1. A copy of the website will be created on a server that is under the control of the project team. The Current Website is on the customer's machine.
  2. The site will first be enhanced as it is, making it more user-friendly. The site uses php to interact with a backend mysql database. Java script needs to be added to perform better validations. For example adding bidding boxes when the user enters the bidding field on the bridge bidding screen.
  3. Further enhancements can be added as determined by the progress of the team. For example, if time permits, the application could be rewritten as a servlet.

Finally, it might be interesting to develop a business plan for marketing such a Web application. This will entail determining whether similar software is available for bridge players. This project is especially recommended for those who play the game of bridge.