Hexapawn

The game of Hexapawn and a method to learn a strategy for play-
ingthe game was described in "Mathematical Games" in the March
1962 issue of Scientific American. The method described in the
article was for a hypothetical learning machine composed of
match boxes and colored beads. This has been generalized in
the Program HEX.
The program learns by elimination of bad moves. All positions
encountered by the program and acceptable moves from them are
stored in the array P$(I). When the program encounters an. un-
familiar position, the position and all legal moves from it
are added to the list. If the program loses a game, it erases
the move that led to ·defeat. If it hits a position from which
all moves have been deleted (they all led to defeat), it erases
the move that got it there and resigns. Eventually, the program
learns to play extremely well and, indeed, is unbeatable. The
learning strategy could be adopted to other simple games with a
finite nwnber of moves (tic-tac-toe, small board checkers, or
other chess-based games).
For complete playing directions, respond YES or Y to the ques-
tion, INSTRUCTIONS?

Becoro # 2023-07-17 23:41:45 - source

hardware: Hewlett-Packard HP-65, Hewlett-Packard HP-67, Hewlett-Packard HP-97,

display: text

port of
Hexapawn (custom)