[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [computer-go] An [open] question on game tree search theory
Don Dailey wrote:
Actually, this is not true. There has been some research that shows
even randomly generated leaf values produce play significantly
stronger than just playing random moves. This may not be as much the
case in GO, because the experiments were done on other games. The
reason seems to be based on the idea of pushing the game towards
positions where the computer has more options (which is good in most
games) and thus has a better opportunities to "find" the random nodes
with high scores.
- Don
I understand this result but in go it seems to be mostly irrelevant. In
go almost any vacant intersection can be played (with suicide rule,
*every*), so both players have the same number of options available. If
black can play to N positions, white can play to N-1 next, or N+K if
black captures. But then white should in general be worse, not better off...
--
Antti Huima (Mr.)
Director, Conformiq Tools
mobile: +358 40 528 8667
email: antti.huima@xxxxxxxxxxxxxxxxx
Conformiq Software Ltd.
Stella Terra, Lars Sonckin kaari 16
FIN-02600 Espoo, Finland
tel: +358 10 286 6300
fax: +358 10 286 6309
http://www.conformiq.com/
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/