Re: [computer-go] Can Go be solved???... PLEASE help!

[Chrilly]

Besides the technical question if it is possible, there is the 
ethical/philosophical one if it should be done. I think solving a game is 
killing a game. It can be played further, because the solution is too 
complex to comprehend, but it looses its charm/mysterium. Nobody would also 
want to play against such a programm. Its no game anymore to play against a 
perfect progamm.
In chess the GMs insist, that once a database position is reached, the game 
is stopped and the result of the database is taken as the final result. Its 
exactly for this reason, playing against a perfect database is no game 
anymore.
I think it is also not worth the computing time. There is no fun playing 
against such a programn, but what is gained by knowing that Black wins 19x19 
in case of perfect play by X points?
The only interesting thing are the computation methods which must be 
developed for such a solution. E.g. Ken Thompson (the Unix-Thompson) 
developed compression methods for endgame databases. He used this 
knowledge/experience to develop a compression method for music which is much 
better than MP3. I have such a CD with this method at home.
Such methods could be also developed by solving a slightly different game 
E.g. to solve chess-endgames where passing is allowed. This would avoid the 
"killing the (end)game" problem. It would be even interesting to compare the 
solution of this modified game witht the original one. One could 
invastigate, how much passing (Zugzwang) is worth.

Chrilly
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/