[Computer-go] Scoring Bug on KGS

Francois van Niekerk flash.slash at gmail.com
Mon Jan 31 11:18:25 PST 2011


Hi All

My program is now playing 9x9 games on KGS (see "oakfoam" in the
Computer Go room) and I have noticed a possible problem:

My program assumes Tromp/Taylor rules, ie: any stones left after 2
passes are alive. I have set kgsGtp for Chinese rules which should be
identical in scoring once all dead stones are removed. The
documentation says the following:

"If the engine and the human disagree on the status of living and dead
  stones at the end of the game, one of two things will happen: If the game
  uses non-Japanese rules and the engine supports the
  kgs-genmove_cleanup command, then kgsGtp will recommend
  that the human press "undo", after which play will continue but the engine
  will get kgs-genmove_cleanup commands (in place of genmove
  commands). In all other cases, kgsGtp will simply refuse to finish the game
  unless the human agrees to the dead stones that the engine listed in its
  response to the final_status_list command."

My program already supports the "final_status_list" command; so what
I'd expect to happen is my program to insist that all stones are alive
and play continues until all dead stones are removed. Unfortunately
this is not what happens. After 2 passes, the opponent can mark *any*
stones as dead and this is accepted. This is obviously not fair to my
program. It is trivial to cheat by marking all the program's stones as
dead.

I have also tried adding the "kgs-genmove_cleanup" command, but it
doesn't seem to ever get called. Is there a way to stop fix this from
happening? Is this a known bug? Does the same happen with ranked bot
games?
--
Francois van Niekerk
Email: flash.slash at gmail.com | Twitter: @francoisvn
Cell: +2784 0350 214 | Website: http://leafcloud.com



More information about the Computer-go mailing list