 |
|
|
|
|
The first international computer tournament
was held in Stockholm in 1974, where the world's strongest Chess
programs competed in the first World Computer
Chess Championship. The first title went to the Soviet Union
with a program called Kaissa, spurring the
West to come back with Chess 4.6 the next
year, the start of a 5 year reign by the USA.
|
|
|
Since then we have seen many more such events,
including the widely publicised contest in 1997 between the supercomputer
Deep Blue and the world's strongest player
at the time, Kasparov. Other competitions
have appeared with the games Go (Weichi
from China), Shogi (Japanese Chess) and
Bridge hosting the biggest events.
|
|
|
Recent competitions:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
 |
|
The
Contribution to Engine Development |
|
|
These events are important in
many ways. First they provide an opportunity to test your program. These
are also good opportunities to meet and discuss ideas with competing teams.
These events are often linked to academic conferences, so the evenings become
brainstorming sessions for new ideas. Finally, success in these world championships
has an important commercial impact: It is much easier to sell a program
that is the current World Champion, or in the top few competing programs,
than a program that has not performed well. This generally results in most
teams putting huge amounts of effort preparing for these events. You can
be sure that everyone you face has been burning the midnight oil tuning
their program! |
|
 |
|
|
|
|
|
Our in-house Shogi program " Shotest",
authored by Jeff Rollason, competed for
the first time in the 1997 World Computer Shogi Championships in
Tokyo (http://www.computer-shogi.org/index_e.html).
This is a field completely dominated by Japan, now attracting some
60 competing programs. Shotest made some waves, coming 13th at its
first attempt, 10 places higher than any other non-Japanese program.
These early competitions were held in Tokyo bay, close to the huge
Tokyo Disneyland complex.
|
|
 |
|
|
 |
|
The following year really attracted attention
as Shotest qualified for the final round of the top 8-programs:
the first Western program to do so. When the final round came, the
final game between Shotest and IS Shogi would decide who took the
title. Since Shotest had already beaten IS Shogi earlier, having
a non-Japanese World Champion was a strong possibility! Shotest
unfortunately lost, coming 3rd overall, a position it defended the
following year. Shotest has competed in all World Championships
since, excepting 2003 and 2004, where the creation and consolidation
of AI Factory took precedence!
|
|
|
|
Shotest currently ranks 7th in the world, having
successfully always been the strongest Western entry in all championships.
It has competed in 8 World Championships, 2 Invitation ISF tournaments
(Tokyo) and 3 Olympiads, winning one of the latter. We will continue
to compete in the World Championship and expect to enter with a
substantially stronger program at the next event.
|
|
 |
|
|
|
|
|
|
|
Dr Mick
Reiss's program Go4++ (now simply Go++), has been competing since 1987,
where he has a solid track record of always being in the top few ranked
programs, having competed in some 25 events in Japan, Korea, China, the
USA and Europe. These events are entirely run using Microcomputers in modern
times, but in early years Mick's Go program was also used to test experimental
hardware, and in one Hamburg tournament ran on a 32 processor transputer.
|
|
 |
|
|
|
Go++ is now ranked the overall strongest Go
program, from the last few tournaments, and iscurrently the top
selling Go program in Japan. His arch-rival for many years has not
been from Japan, but from China, where the venerable Chen Zhixing,
author of Handtalk (now called Goemate) dominated the early tournaments,
but who has now dropped below Go++ in the rankings.
|
|
Mick's program has featured in many publications,
including a long article in New Scientist: See New Scientist vol 174 issue
2338 - 13 April 2002, page 38 |
|
|
|
|
|
|
Andrew Bracher started
work on Oxford Bridge as early as 1980, then with a modest Commodore
VIC20. Unlike Go, Shogi and Chess, Bridge is not a game of complete
information, which adds another substantial dimension of complexity
to solving the game. From a modest start, a solid Bridge Program
was developed that now competes in Computer Tournaments. In 1998
his program (Omar Bridge) competed in a tournament with six other
Bridge programs and the World Bridge Champion Zia Mahmood. Omar
Bridge did not rank above the champion, but led the field jointly
with Zia for several rounds and finished joint top among the computer
opponents. Oxford Bridge is still being very actively developed.
|
|
|
|
|
|
|
Chess accounts for the bulk of Computer tournaments
and attracts the most substantial effort. The top programs now play
at a level that would allow them to reasonably compete in the human
Grandmaster events, even when running on modest microcomputer hardware.
One of our Chess engines is based on the Chess System Tal program,
authored by Chris Whittington. This exceptional program is renowned
for its strong human style, resembling that of the former World
Champion Mikhail Tal. His great strength was in the creation of
positions with great tactical complexities, as the World Champion
Petrosian remarked "I know what to do in the positions that
Tal creates, but I do not know how to create them". Chess Tal
plays in this attacking style and has been shown to perform well
in human contests. In Computer Chess it came 10th among an exceptionally
strong field of 34 programs at the 15th World Computer Championships
in Paris 1997. The Chess Tal engine is not currently being developed,
but already plays at an exceptionally high level, beyond the vast
majority of human players.
|
|
|
|