Re: Introduction

[Darren Cook <darren@xxxxxxxxxxxxxx>]

>introduced to the game of Go only a month or so ago and immediately
>found it fascinating and started working on a program. 

I learnt go mainly to program it, and I feel the stronger a player you are
the more it will help you.

>amount of work that has been put into this game is staggering and the
>outcome quite humbling, to us all I think (IMHO).  This is a challenge
>worth working on !-)

Indeed!

>Checkers, Othello, Go, etc.).  The only good way to truly "read a
>position" seems to be to play it out to a reasonable depth.   How can
>one predict future positional characteristics accurately using only the
>current position?  I don't think it's possible.?  Anyone disagree?

Disagree, but only partially. Here's a simple example:

If you have a chain (aka string; solidly connected stones) and two of it's
liberties fufill the criteria for one point eyes (friendly stones on all 4
sides and at 3 of the 4 diagonals) then it is alive, and you've not needed
to search.

Not that useful in a real game, as you don't often get single point eyes
appear. But maybe the idea can be expanded. The program I entered into the
FOST cup last year was statically evaluating a lot of quite complex
life/death positions correctly. Life/death reading (as David Mechner has
also pointed out) is the biggest problem with today's programs. 

(So why did my program do so badly? 1.the positions it could evaluate are
like those in the problem books - in real games life/death has much more to
do with connectivity; 2.it wants to spend about 30 seconds analyzing each
chain on the board, so I had to remove lots of stuff to meet the time
controls; 3. I exagerrated - it was also getting a number of quite complex
life/death positions wrong :-).

BTW, I only partially disagree - your heuristics must be able to tell you
when they are not sure, and in that case you have to fall back on search.

>difficult, there are similarities.   Good chess programs do not play
>like humans play and I don't think a good (dan level) Go program will
>play like humans play.  The AI progress made to date is simply not up to

I think go will need a different kind of brute force from chess, and as with
chess it won't be the way humans play. I'm not too sure what that will be
though.

Darren