Information Technology Reference
In-Depth Information
Much more of the history of computer chess can be found in the topic by Levy and
Newborn [50].
The reason for this fascination with chess is clear: it is intelligent behavior that can
be exhibited in a very clear-cut way. Nobody disputes the achievement of a program
that is able to defeat a world chess champion.
Since the time of the Deep Blue computer, there have been a number of progres-
sively stronger chess programs, with names like Deep Fritz and Rybka. Although
these programs appear to be outdistancing their human competitors, chess federa-
tions like FIDE do not actually rate them, chess competitions mostly exclude them,
and chess champions often use them only to help improve their games against other
humans. Chess commentators might mention what “the computer” would do in their
analyses, but leave it at that.
Outside of chess, the games of Go and checkers are perhaps the most studied. As
noted earlier, the game of checkers was solved by Jonathan Schaeffer [52], a researcher
who had been working on a checkers program called Chinook for some time. (The late
world champion, Marion Tinsley, lost a game to this program in 1992; he had not lost
a game in forty years.) The game of Go, on the other hand, is interesting because it
proved to be so difficult. Until the 1990s there were simply no computer programs
that played Go competently [51]. Currently, there are much better programs, but they
use statistical techniques beyond the minimax presented here. Statistical techniques
are also used quite successfully in other games like bridge and backgammon that are
outside the category of games considered here.
One interesting development in the area of strategic games is the recent attention
to general game-playing, where the players do not know ahead of time the rules of
the game they will be playing [49]. At the time of a competition, the game-playing
programs are given the rules (specified in a certain language) as input and then have
to figure out from there how to play and how to play well.
Exercises
1.
Consider a game of tic-tac-toe where X begins by choosing position 2.
a.
Show that the general game player in gameplayer.pl finds no winning move
for O from here.
b.
The usual move for player O here would be a center or a corner. Show that
the general game player finds an equally good move for O .
 
Search WWH ::




Custom Search