a group is a set of stones that die or live all together in playouts.
That is if P(stone A lives and stone B lives) is almost equal to
P(A lives or B lives), they are in the same group.
a semeai is then two groups of stones with anti-correlation
P((white stones live and black stones live) OR (white stones
die and black stones die)) almost 0, whereas P(B lives) and
P(W lives) both nonzero and non-one.
I really like these suggestions. Thanks!
Glad to see my random comments might be of some more use than just
sparking the discussion!
Group identification using static code is hellish. Pebbles devotes hundreds
of lines of code to it, and still has problems. I really want to find ways
to eliminate such code.
I suggest one additional condition: semeai groups have to be in contact.
Otherwise it is possible to have two groups with "1.5 eyes" that have
anti-correlated life because MCTS will make sure that each color wins one of
them.
I'm afraid I do not picture this counter-example. Do you mean two black
groups with 1.5 eyes, so that if B saves one, W kills the other? Then
the two groups are of the same colour, instead of being of different
colours.
If on the other hand it's a white group and a black group, if B saves
its group, W saves its, too, so the two groups are correlated instead of
being anticorrelated.
What am I missing?
Jonas
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
