[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gnugo-devel] Gnu go at the computer olympiad

From: Gunnar Farneback
Subject: Re: [gnugo-devel] Gnu go at the computer olympiad
Date: Wed, 03 Jul 2002 22:36:46 +0200
User-agent: EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/20.7 (sparc-sun-solaris2.7) (with unibyte mode)

Inge wrote some weeks ago:
> It is now one month left to the computer olympiad. [...]

The web page at

has some interesting info. First we can see the list of participants
in 19x19 go:

F. Lu              Ether          
I. Wallin          GNU Go         
K-H. Chen          Go Intellect   
M. Reiss           Go4++          
G. Szymborski      Shape          
S-J. Yen           Star of DongHwa
I. Tonkes          Tsgo           

and 9x9 go:

I. Wallin          GNU Go      
K-H. Chen          Go Intellect
M. Reiss           Go4++       
E. van der Werf    Magog       

(Notice that the listed people are the "corresponding authors" and may
or may not be a main developer of each program.)

Completely unknown to me, as is the author. The fact that the author
has travelled all the way from Taiwan may indicate that it is likely
to be relatively strong, but this is far from certain.

Go Intellect
Well established program which GNU Go has met at 21st Century Cup
2001, losing a rather close game by 10.5. They have also met in Stefan
Mertin's 13x13 tournament with a disappointing 5-1 game win for Go
Intellect. Hopefully we can do better at this tournament.

Clearly the favourite of the tournament. GNU Go has played one game
against Go4++ at 21st Century Cup 2001 and lost by 43.5, having been
behind almost from the start. We might be closer this time, but I
don't think we can hope to win yet.

Also completely unknown to me. The author is a Polish 2 dan.

Star of DongHwa
Same comment as for Ether.

GNU Go won clearly when they met last summer at the European Go
Congress. We should win easily now too.

Unknown to me. The author is apparently doing machine learning
research at the University organizing the Olympiad. My guess is that
the program is fairly weak.

One can also see that the allowed time is only 30 minutes per program.
This is fairly fast but shouldn't be a problem on the 1.3 GHz Duron
Inge is planning to use. It may be worth running a couple of test
games before the tournament to make sure of this though. There may
also be some time lost for operator overhead in non-GMP games.

Unfortunately I can find no other information about rules or scoring.

I look forward to progress reports and game records.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]