Robotics Reference
In-Depth Information
game tournament, scoring two wins and two losses for a performance
rating of 1640, corresponding to Class B. In recognition of this achieve-
ment Greenblatt's program was made an honorary member of both the
U.S.C.F. and the Massachusetts Chess Association, and for the first time
computer Chess attracted some attention in the media.
The following year, 1968, saw the birth of serious human versus com-
puter challenges at Chess when I started what became known as “the Levy
bet”. During one of the annual Machine Intelligence Workshops, orga-
nized in Edinburgh by Donald Michie, I was invited to a cocktail party
along with several luminaries of the AI world, including John McCarthy.
During the evening McCarthy invited me to play a game of Chess, which
I won. He then said: “You may be able to beat me, David, but within
ten years there'll be a program that can beat you.” I was incredulous. As
the reigning Scottish Champion I had no doubt at all that McCarthy was
wrong, and offered to make a bet to that effect. McCarthy asked me how
much and, brashly, I suggested
£
500 (around $1,200 at that time). I was
in my first job, teaching programming at Glasgow University and earning
the magnificent sum of
£
895 per year, so
£
500 represented a significant
wager. McCarthy called over to Michie, who was sitting on the floor
a few feet away from us, and asked Donald what he thought about my
challenge. Michie's reaction was to ask McCarthy if he could take half of
the bet, and so it started, with the two of them wagering
£
250 each that
I would lose a match to a computer program before the end of August
1978. The following year Seymour Papert of MIT came in for another
£
250 and in 1971 the total stake rose to
£
1,000 when Ed Kozdrowicki,
a Computer Science professor at the University of California at Davis
and co-author of COKO, one of the early competitive Chess programs,
joined the fray. As time went on my confidence in the outcome of the bet
remained undimmed, so I was delighted when, in 1974, Donald Michie
offered to up the stakes, making the total
£
1,250, and that is how the bet
remained.
The start of the 1970s saw another important event in the history of
computer Chess. The Association for Computing Machinery, at its an-
nual convention in New York, organized the first ever Chess tournament
in which all of the participants were computer programs. There were six
entries and the event was won by a program called Chess 3.0, written at
Northwestern University by David Slate, Larry Atkin and Keith Gorlen.
Chess 3.0 evaluated approximately 100 positions per second and played
at the 1400 level on the U.S. Chess Federation rating scale.
