Masters Thesis

Generating an End Game Tablebase for the Game of Breakthrough using Quasi-Retrograde Analysis

End game tablebases are frequently used in chess to store game winning positions where few pieces remain on the board. Breakthrough usually has many more pieces than chess remaining on the board at game conclusion, making a standard style end game tablebase infeasible. However, due to the way the pieces in Breakthrough move, a table base entry for Breakthrough can instead be limited to a certain number of rows. By using quasi-retrograde analysis, forced wins can be found through reverse play from a winning position that are several moves away from game conclusion and may not be easily identifiable by a computer player. In theory, a computer player using Monte Carlo Tree Search can complete more simulations ending in a game outcome that can be known with certainty and therefore increase the likelihood of finding game winning moves if random play out can be terminated early via a query to an end game tablebase of proven forced wins.

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.