[computer-go] 19x19 CGOS

Jacques Basaldúa jacques at dybot.com
Wed Oct 31 04:49:30 PDT 2007


Don Dailey wrote:

 > There is no simple way to fake it with combinations of pass
 > moves with programs cooperating. Probably most programs
 > won't pass on the first 20 moves or so, but we can't count on
 > that behavior because it's incorrect. You should always pass
 > immediately if it wins the game outright.

Probably right. There is probably no simple idea that works
with arbitrary handicap and any player, but what I suggested
does work for white giving 3 stones (or less). The server
should still understand that the program is special in the sense
that it should always play white. I guess that shouldn't require
a lot of work.

What John Tromp suggests works for higher handicaps, but
not for 3. And it is a bad idea for black. Just imagine:

1 c4 ..
2 .. f3
# Now B encloses W as John Tromp suggested:
3 e3 ..
4 .. pass
5 f4 ..
6 .. pass
7 g3 ..
8 .. pass
9 f2 ..
# The white stone is captured
10 .. q16

And now black's advantage is small for a handicap 3 game.
Black could lead in 3 corners, one of them with shimari after
move 9 and what did he get attacking? One corner.

Jacques.



More information about the computer-go mailing list