[Computer-go] Quiescence Search for Go?

Rémi Coulom Remi.Coulom at free.fr
Mon Mar 29 01:23:53 PDT 2010


Stefan Kaitschick wrote:
>> Stefan Kaitschick wrote:
>>> Maybe a way to do quiescence in go is to trade moves in result-critical
>>> local fights before doing the actual search.
>> "Local fights" are not entirely local always. It might be better to do
>> something at the ends of the tree before simulations.
>>
>> --
>> Yamato
> 
> Yes, trading moves ahead of time would create new problems, even if a fight 
> really is local.
> For example, a move for a semeai might be ladder breaker.
> And ofcourse the ko threat balance would be upset.
> It's a question of is it more poison or more cure. :-)
> There might be some way to create a retraction mechanism for provisional 
> moves, but I agree it's thorny.
> That's why I think that the less radical approach of reducing frequencies of 
> potential thrashing moves in the move generator might be more tractable.
> 
> "ends of the tree before simulations"?
> I don't get that. Isn't the simulation allready underway when you arrive at 
> end nodes?
> 
> Stefan 

Yamato probably meant at the frontier between random playouts and tree 
search. I do something like that in Crazy Stone, too. For instance, if 
the last move in the tree was an atari, and a quick tactical analysis 
shows that the opponent cannot survive by extending, then the first 
random move of the playout should not extend.

Rémi



More information about the Computer-go mailing list