Skip to content

Alternate method of training #165

Description

@MarkZH

Instead of playing against each other, the genetic chess engines read a PGN file of a large number of games played by professional human players or high-level engines. As the engines read the file, they step through the games while trying to predict the next move. The engines that get the most moves correct get to create the next generation.

Or, the engine plays moves of the winning side and the opponent plays the moves of the losing player, resetting the board on each move.

Since there's an upper limit to fitness (100% correct predictions), this may only be useful as an initial bootstrap to a normally run gene pool.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions