There are a number of definitions of solved, ranging from "a program
exists that can beat any human" to "we can quickly determine, for any
position, the best move and the result under optimal play". In the
latter strong sense, I believe Go has only been solved up to 5x5,
maybe 6x6.
There are some games, such as Hex, for which we know who wins from
the starting position given optimal play, but we don't know how to
figure out the best move.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Jan 12, 2007, at 8:45 AM, terry mcintyre wrote:
From http://senseis.xmp.net/?7x7BestPlay it looks like 7x7 Go
may already have been solved. 5x5 was solved in 2002, according
to http://erikvanderwerf.tengen.nl/5x5/5x5solved.html
AFAIK, 9x9 Go has not been solved yet. 19x19 Go will surely exceed
the capabilities of computers in my lifetime, I suspect.
-- Terry McIntyre
----- Original Message ----
From: Chris Fant <[EMAIL PROTECTED]>
To: computer-go <computer-go@computer-go.org>
Sent: Friday, January 12, 2007 8:16:35 AM
Subject: Re: [computer-go] Can Go be solved???... PLEASE help!
Seems like a silly title. Any game of perfect information that has a
clear rule set can be solved. Plus, some would argue that any Go
already is solved (write simple algorithm and wait 1 billion years
while it runs). A better question is, "Can Computer Go Surpass Human
Go?" But again, clearly it will. It's just a question of how long
until it occurs.
On 1/12/07, Mehdi Ahmadi <[EMAIL PROTECTED]> wrote:
> Hello & thank in advance for any interests/ responses.
>
> I'm unfortunately (or not) doing a dissertation as part of my
final year
> project (undergraduate) on the game of Go. The exact title is:
"Can the game
> of go be solved? Analysis of computational methodologies for go."
And I have
> included my overall objectives below.
>
> I have many works from different people on different aspects of
Computer Go
> which would make for great inclusion at different parts - but
overall I am
> still gravely struggling. In reviewing some of these my greatest
difficulty
> is in understanding exactly how say Monte-Carlo-UCT or even Alpha-
Beta
> testing (pruning, etc) occur so as to be able to give a
simplified depiction
> (illustrated or otherwise) of the process. Can this be done
without having
> to go through the source code of say something like GNU Go?
>
> Also another difficulty I've had is in trying to get further
information on
> the commonly referred top ranking packages, Handtalk, Go++, Many
Faces of
> Go, etc due to their commercial nature? (the only thing I've been
able to
> find which is a bit outdated:
> http://www.inventivity.com/OpenGo/Papers/EditedGoPapers.html).
>
> Lastly can any general categorisation - distinction be made of
current
> approach/ implementations in trying to 'solve' Go. in comparison
to say
> traditional disciplines used in trying to solve games (complex or
otherwise)
> via computer? To put simply I am trying to have some core root
comparison in
> current methodologies (if there is any?).
>
> If anyone has any suggestions/ guidance on anything mentioned - I
would be
> eternally indebted.
>
> ==================================
> 5.1 OBJECTIVES
> . To concisely review all game playing aspects of Go (rules,
openings,
> middle game, etc) and its relevance to the complication of
meaningful
> measurements of interest.
> . To evaluate, gain and develop further understanding of specific
game
> aspects including (eg):
> - Representation:
> . Eyes
> . life-and-death
> . territory estimates and weakness
> - Move Evaluation:
> . Territorial and strategic affluence.
> . Address specific and current implementation methodologies
including:
> - Search algorithms (Alpha-Beta - local/global, Monte-Carlo -UCT)
> - Move Generation
> - Positional Evaluation (Patterns, Neural Networks)
> . To detail inadequacies in research and reasons for shortfalls
where
> applicable.
>
>
>
> _______________________________________________
> computer-go mailing list
> computer-go@computer-go.org
> http://www.computer-go.org/mailman/listinfo/computer-go/
>
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
We won't tell. Get more on shows you hate to love
(and love to hate): Yahoo! TV's Guilty Pleasures list.
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/