|
Hello,
Jens Lieberum wrote:
>There
remains the (maybe unlikely) case of an unsettled block where a
search
>with W to move first, B wins all kos and W must answer nearby threats would >return "dead". In this case the status should be "special unsettled" as >well. Does the additional time to determine this status exactly not >sufficiently pay off in playing strength? David Fotland wrote:
>I assume that if a W group can be captured with W moving first, it will also be >captured with B moving first. Sorry, I still did not get it and I hope that I do
not write complete nonsense now.
I agree with your assumption that if W moves first
and the block is dead, then it will also be dead if B moves first (given the
same ko winner and the same defender of nearby threats). But still, after the
three block searches you do, the question remains open whether or not the block
is alive or dead under the conditions I mentioned (W moves first, B wins kos, W
must answer nearby threats). If it turns out to be dead, then the status of the
block depends on ko or nearby threats if W moves first and it is dead if B
moves first. This situation is possible, and it might be better to treat
it by global evaluations in a quiescence search. For example, this case might be important, if W is behind and wants to
initiate a fight (or if B wants to turn down this option).
The fourth block search would lead to four types of
unsettled blocks, three of them being special-unsettled (and each special case
might require a different treatment). The four types have
a partial ordering. One type is better for white,
one better for black, and two uncomparable cases in the middle. One of the
middle cases is unsettled (not special), the other middle case is the most
unlikely one, but I guess that it is not impossible (e.g. it might appear
in situations involving a two-step-ko). Of course, some good examples would
be necessary to see, if it is really helpful to distinguish all these cases.
Anyway, these considerations are far too
advanced for what I am trying to do at the moment and I am very satisfied with
the overview on possible solutions I got now.
Jens |
_______________________________________________ computer-go mailing list computer-go@xxxxxxxxxxxxxxxxx http://computer-go.org/mailman/listinfo/computer-go