Re: [gnugo-devel] Strange game.

[Gunnar Farneback]

Kynio wrote:
> Were 32-43 moves nessesery???

That depends, but it's not a GNU Go issue as much as a KGS issue.

If a robot on KGS doesn't agree with the opponent about which stones
are dead at the end of the game and the rules are other than Japanese,
the opponent is asked to resume the game. Then the engine is given a
special command intended to generate moves which clean up the position
so that the status will become clear. GNU Go does this by playing on
until all opponent stones it considers dead are removed and all own
alive stones (except ones living in seki) are unconditionally alive in
the sense that they cannot be removed even if the opponent is allowed
an arbitrary number of consecutive moves. That is what happened in
moves 32-43.

If you don't want to see such moves you should agree with the robot
about dead stones. If you don't think you did disagree there's a bug
in the KGS server and/or the program connecting robots to KGS. That
bug should be reported to the person running the robot and/or the KGS
admins. GNU Go worked as intended.

/Gunnar