Loading ...
Sorry, an error occurred while loading the content.

Territory & dead stones?

Expand Messages
  • William M. Shubert
    While the SGF standard mailing list is sort of active, let me first say that I agree with Robert Jasiek in that a lot of things should be put in general
    Message 1 of 2 , Dec 10, 1998
    • 0 Attachment
      While the SGF standard mailing list is sort of active, let me first say
      that I agree with Robert Jasiek in that a lot of things should be put in
      general comments. IMHO, most properties are for
      machine-readable-and-understandable data; the comments are for things
      that humans are meant to read and understand. "Played during the Qing
      dynasty" sounds like something you would never expect a computer to
      understand, only a human, so it belongs in a comment.

      Now for my problem. When a game is over, you need to mark dead stones
      and territory. I believe that programs usually mark territory with TW[]
      and TB[], and mark dead stones by placing a TB/TW marker over them
      (since usually a dead stone is in the territory of the opponent). But
      there is one case I can think of where you have dead stones that are
      *not* territory, and that's in an eye of a group in seki in Japanese
      rules. E.g.:

      5 |
      6 | O O O # # # #
      7 | # # # O O O #
      8 | # . # O . O #
      9 | # O # . O O #
      +----------------
      A B C D E F G H

      Note the stone at B9. According to Japanese rules, it is dead and thus
      worth one point to Black at the end of the game. But it
      IS NOT TERRITORY. If it were marked with "TB[]", then Black would get
      two points - one for the dead stone, and one for the point of territory.
      This would be wrong. This wouldn't matter too much, but cgoban does
      recognize seki and scores it correctly. It just doesn't know how to save
      the scored poisition to an SGF file, so at the moment it (incorrectly)
      labels the dead stone as territory.

      So how do we mark this situation? Or does this fall into the
      "it's-so-rare-that-who-cares" category? Seem to me that it would have
      been nice to have a "DE[]" parameter to mark dead stones separate from
      the TB/TW parameters, but it may be too late to add something like this
      now.
      --
      Bill Shubert (wms@...)
      http://www.hevanet.com/wms/
    • Robert Jasiek
      ... What one really needs to mark are the scoring points on the board. If one liked to mark other things, one would have to use several stages of diagrams.
      Message 2 of 2 , Dec 11, 1998
      • 0 Attachment
        > Now for my problem. When a game is over, you need to mark dead stones
        > and territory.

        What one really needs to mark are the scoring points on the board.
        If one liked to mark other things, one would have to use several
        stages of diagrams. E.g. with the Nihon Kiin 1989 rules and at the
        game end one might illustrate in the following order:

        1 - living stones (distinguished: uncapturable stones, capturable stones
        [but what about stones that are alive _and_ dead in case of incomplete
        and incorrect play?]), dead stones, eye points, dame (all distinguished)
        2 - stones of groups not in seki, stones of groups in seki, [optionally
        any remaining stones], points of territory (all distinguished)

        Not to mention extra troubles during the confirmation phase.
        AYCS, rules intricacies are not useful for SGF properties. One
        can still illustrate them by using labels, e.g. Stage 2 as above
        might be shown with N, S, T labels (for stones of groups not in seki,
        stones of groups in seki, points of territory). Or you offer three
        diagrams for these.

        > I believe that programs usually mark territory with TW[]
        > and TB[], and mark dead stones by placing a TB/TW marker over them
        > (since usually a dead stone is in the territory of the opponent).

        With some rule sets this can be reasonable. With most Japanese rule
        sets it is not.

        > there is one case I can think of where you have dead stones that are
        > *not* territory, and that's in an eye of a group in seki in Japanese
        > rules.

        Correct play would use the proper [encore] method of the exact Japanese
        rule set to capture the dead stone(s).

        > According to Japanese rules, it is dead and thus
        > worth one point to Black at the end of the game.

        It is only worth one point if being removed BEFORE the [completion of
        the]
        end of the game. Otherwise it remains within a group in seki, where
        there
        is no territory from which stones would be allowed to be removed.

        > If it were marked with "TB[]"

        With Japanese rules a mistake.

        > This wouldn't matter too much, but cgoban does
        > recognize seki and scores it correctly.

        See above.

        > It just doesn't know how to save
        > the scored poisition to an SGF file, so at the moment it (incorrectly)
        > labels the dead stone as territory.

        If you want to store the scored board points as well, you must not
        include any seki points (under Japanese rules). If the game ended as
        shown, then the dead stone in seki does not belong to territory,
        is not marked, and does not score at all.

        > So how do we mark this situation?

        Do not mark it. Players should have played it out correctly.

        Also see:
        --
        robert jasiek
        http://www.snafu.de/~jasiek/rules.html
        http://www.snafu.de/~jasiek/j1989com.html
        http://www.snafu.de/~jasiek/wagc.html
      Your message has been successfully submitted and would be delivered to recipients shortly.