Hi Robert,
Thanks for your explanations.
I don't really understand what is the relation between what you tried to
explain and the equivalence in normal cases of Chinese, Japanese scoring
that I was trying to show in my approach.
I understand that in combinatorial game theory, we need to seperate the
position into "independent" local parts for analyzing (by mean and
temperature?). Indeed, maybe my approach causes some inaccuracy since passes
and Margin are considered only globally, but I think maybe it is already
enough for disposing most normal cases, except that seki and superKo.. etc
such special situations need to be handled.
What I was trying to do is to consider the effect of each PLAYED STONE and
compare the Chinese scoring with the Japanese. So, I was not considering
from the view of the move values.
I agree with Don that "correct Japanese scoring by computer is almost
impossible to get right in every case". But I am content that my program can
face to most cases (I believe over 99%) correctly.
I will be very glad to see if there will be strict mathematical proofs and
conclusions from Berlekamp and you. I think many Go players will be very
interested as well. :-)
Aja
For which purpose...?
If all local boundaries are fixed and life and death is settled locally in
each region and ko does not play a role, then the local behaviour of each
region considered together for all regions should equal the global
behaviour. Otherwise the same cannot be said. So why would you want to
consider things "only globally"?
. # O . O # . # O .
. # O O O # # # O .
. # O . O # . # O .
. O . O # . # .
# O O O # # # O
# O . O # . # O
. # O . O # . # O .
O # O O O # # # O #
O # O . O # . # O #
You might create problems for yourself... Globally terminal (for human
players) under perfect play does not(!) imply locally terminal in general.
If your aim is to approach rather than to guarantee perfect play, you
might ignore the global / local difference nevertheless. Maybe such a
strategic cheating by simplification saves computational time. You should
just be aware that your program might sometimes make strategic mistakes as
a consequence.
--
robert jasiek
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
__________ NOD32 5064 (20100427) Information __________
This message was checked by NOD32 antivirus system.
http://www.eset.com
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
