We have designed a backgammon program to play intelligent games. It can make good decisions of the moves of its checkers. It also can learn from its past experience. This thesis describes how this program works. This thesis is divided into seven chapters. Chapter one gives some background in Artificial Intelligence and the backgammon game. Chapter two summarizes the history of the last 30 years of computer game playing by some well-known researchers in A.I. field. Chapter three "briefly describes the overview of this backgammon pro gram, including the basic playing, input, move verification, output, game starting and terminating, and move generation routines. Chapter four describes implementation & performance of the multiple linear polynomials decision-making algorithm that was used in this project. This chapter also discusses different strategies and the scoring terms that were used in the strategy polynomials. Chapter five describes the fast learning algorithm that was used, and how it improved the performance of the program. Chapter six gives sample games that the program played with a few human competitors and also analyzes the program's performance. Chapter seven discusses some future developments, extensions of the program, and the conclusions that we have from the result the project.

Library of Congress Subject Headings

Computer games--Design; Backgammon--Data processing; Microcomputers--Programming

Publication Date


Document Type


Student Type


Degree Name

Computer Science (MS)

Department, Program, or Center

Computer Science (GCCIS)


John Biles

Advisor/Committee Member

Peter Lutz

Advisor/Committee Member

Warren Carithers


Physical copy available from RIT's Wallace Library at GV1469.2 .Y354 1983


RIT – Main Campus