- 24-bit ANSI colored, console-based chess using Java.
- Either side can be human or a multi-threaded AI instance using the Minimax algorithm with alpha-beta pruning.
- Uses map/reduce and streaming techniques to speed up the parallelism using multiple cores during board valuations since those happen after every move millions of times on every thread.
- Includes high-speed mapping and caching of all moves seen by all threads in order to gain short-circuit efficiency on parallel threads evaluating the same board state for which the best move to be made has already been determined earlier by the popularity winner of previous evaluation threads that all ran to ply (or quiescent search) depth and were repeatedly determined to be the best moves. The required percentage of repeated wins as "best move" is configurable to control the required reliability of the cached best moves answers.
- Includes end game recognition and several map/reduce specialization approaches including walking the King into a corner if possible and intentionally advancing and sacrificing pawns if necessary in order to allow other pieces to make higher value moves or unblock pieces when it is recognized that they cannot advance.
- configurable choice of how many of the available physical processor cores to use
- Includes quiescent search support.
- Full support for reading and writing FEN chess notation strings and full game replay.
- Fully configurable properties text file including: ply depth, thread pool size, optional AI time limit, all colors, and more.
$ java -jar chess.jar
To play against the computer change the humanPlayer value to true in the editable chess.properties file and re-run the jar file. Or change any of the other options and re-run the jar file. Play around! Many interesting combinations are available there:
#
# Usage:
#
# General Game Settings:
# numDrawReps number of repeated moves to call game a draw
#
# AI Settings:
# humanPlayer: set to true to play as player 1 (white)
# maxThreads: maximum number of threads for AI to run simultaneously
# aiPlyDepth: maximum number of moves for AI to look ahead
# maxAISeconds: maximum number of seconds to allow AI to think (0 for no time limit)
#
# UI Settings:
# showVictims: color the board to show current players possible moves if true
# showTargets: color the board to show opponents possible moves if true
# showCapturesOnly: limit board coloring to only show possible captures if true
#
# player1 label to use for player 1
# player2 label to use for player 2
#
# ANSI Colors (24-bit):
# blkBack background color for dark squares
# whtBack background color for light squares
#
# blkForeB foreground color for black piece on dark square
# blkForeW foreground color for black piece on light square
#
# whtForeB foreground color for white piece on dark square
# whtForeW foreground color for white piece on light square
#
# blkMoved foreground color for last moved black piece
# whtMoved foreground color for last moved white piece
#
# blkCheck foreground color for black king in check
# whtCheck foreground color for white king in check
#
# targetsShade color influence for target path positions
# victimsShade color influence for victim path positions
#
#
#Wed Jul 03 01:13:30 CDT 2019
numDrawReps=3
humanPlayer=false
maxThreads=100
aiPlyDepth=6
maxAISeconds=75
showVictims=true
showTargets=true
showCapturesOnly=true
player1=Player 1
player2=Player 2
blkBack=142, 142, 142
whtBack=204, 204, 204
blkForeB=64, 64, 64
blkForeW=0, 0, 0
whtForeB=255, 255, 255
whtForeW=255, 255, 255
blkMoved=192, 0, 0
whtMoved=0, 0, 255
blkCheck=192, 128, 0
whtCheck=192, 128, 0
targetsShade=0, 0, 128
victimsShade=160, 0, 0