Robotics Reference
In-Depth Information
Figure 17.
The starting position in a
game of Chess
Figure 18.
Numerical representation of
the starting position
attacked.
1
The basic approach described in Shannon's paper has been
employed in almost all of the Chess programs written since then, as well
as in programs for many other classic two-person games such as Checkers
(Draughts) and Reversi (Othello
TM
).
How Computers Recognize a Chess Position
Computer programs know nothing of the different shapes and sizes of
the Chess pieces so they must be given some way to represent a Chess
position—a way that a program can understand. A simple representation
assigns different numerical values to the various Chess pieces, designat-
ing White's pieces by positive values and Black's by negative. Thus the
starting position in a game of Chess (see
Figure 17)
could be represented
numerically as in Figure 18.
In the right-hand diagram, the pawns are represented by 1 (White's
pawns are +1 and Black's are -1), the knights by 2 (and -2), the bishops
3 (and -3), rooks 4 (and -4), queens 5 (and -5) and kings 6 (and -6).
Empty squares are represented by 0.
The Moves of the Game
The first thing that a human beginner learns at Chess is how the pieces
move. Similarly, a program cannot begin to play the game until it can
work out which moves are possible from any Chess position. The detail
of how this is accomplished varies from one program to another but the
essence is the same. As a simple example, a program could be given a
11
Shannon's paper was originally presented at a conference in New York, on 9 March 1949—
Bobby Fischer's sixth birthday.
