CHESS AND COMPUTERS –
THEY GO TOGETHER LIKE CHOCOLATE
AND PEANUT BUTTER
Chess and computers seem like a natural match. Chess players like to
talk about the “calculations” they had to make when thinking about a
move; since a computer is simply a calculating machine, chess and computers go together like a hand and glove.
Claude Shannon thought so too. Shannon was a seriously brainy guy who
worked as a researcher at Bell Labs in New Jersey. In 1949, Shannon began to give public talks about how a computer might be programmed to
play chess. He wrote two versions of a paper on the subject, a “crunchy”
version for other scientists, which was published in Philosophical Magazine, and a less math-intensive “mass market” (in other words, simplified)
version for public consumption, which was published in Scientific American; both articles appeared at about the same time.
Shannon speculated that a computer might be programmed to play a
passable game of chess, not that there was any intrinsic value to such a
pursuit in itself (we have to remember that computers at the time were still
room-sized machines which cost a fortune to operate), but that research
into this area might be transferable to a number of other, more useful
purposes, such as automated telephone switching (still performed by
human operators in 1949 – Shannon worked for Bell Labs, remember?),
language translation, even as an aid to making military decisions (later
explored in-depth by Trevor Dupuy in a book I own which can make your
head explode from all the math in it). Shannon explored the basics of how
a computer might be made to “think ahead” in chess, evaluate positions,
and come to a logical conclusion as to what to play in a given position.
Although some of his ideas have been shot down or replaced over time,
many others are still used in programming chess computers today; his
papers are still considered basic required reading for anyone who wishes to write a chess program. (You can easily find Shannon's “Programming a Computer for Playing Chess” on the Internet if you'd like to read
it. And although Shannon's paper didn't furnish the algorithm [formula]
for an actual chess playing program, you can find at least one downloadable program named “Shannon” [in a proprietary format] which is based
primarily on his original ideas.)
The first stab at a working chess program came from across the pond
from Shannon. England's legendary Alan Turing (hailed today as “the father of computer science”) wrote an actual working chess program he
called Turochamp around the same time as Claude Shannon's theories
saw print. As mentioned a paragraph ago, computers were obscenely
18
chessking.com