Chess is a game that requires a lot of experience to correctly evaluate a position. This can be simplified with visualization, in which the squares are colored depending on the direct or indirect influence of the pieces. The thesis expands the concept by including future moves along with current ones. For this purpose, we model the chessboard as a matrix, in which we calculate the reach of each piece up to the desired number of moves. Each field of the matrix contains a value, dependant on the pieces that can reach it. The resulting matrix of values is then converted into a raster image by referencing a linear colour scale. We apply the algorithm on a large sample with a varying amount of allowed moves to assess its functionality and adequacy. For this we prepare attributes on the basis of the matrices, for which we calculate a correlation coefficient against a board evaluation, given by a chess engine.
|
