Re: Using partial plys

[Joan Pons Semelis <joan@xxxxxxxxxxxxxxxx>]

From: David Fotland <fotland@xxxxxxxxxxxxxxxxx>
> For example:  Here is a group that is undecided, and unclear if the
> answer is correct.  But I am way ahead, this is a small group, and
> it not a strategic group (cutting, etc).  So I will assume it is dead.
> Now I have one less uncertain group.  I am certain it will die, since I
> won't try to save it.  If the situation becomes clear and I can confidently
> save it in the future, I can change my mind, but I want to keep the evaluation
> consistent as long as possible.
> 
> I don't do this in Many Faces yet, but I think it has promise.
> 
> I think a consistent evaluation is as imortant or more important than
> a correct evaluation.  Inconsistent evaluation is one of the causes of
> the many bad tenuki moves most go programs make.
> 
> David Fotland

I though those tenukis were because of incorrect evaluations, at least in 
my program they are. But I agree that taking a pessimistic point of view
in an unclear position is safest. 

Joan Pons i Semelis joan@xxxxxxxxxxxxxxxxx