Is the counting in cgoban broken?

Recently I’ve been watching an anime series called “Hikaru no Go”  and promptly I’ve become interested in Go too. 影響が強くされるとよく言われますよ。それはいいかどうか、分からないんですですけど。。。

Looking around on the net, there’s a nice page to learn the basics of Go: Interactive Way to Go and as usual there are a couple of open source programs to install from.

While gnugo seems to be one of the better Go AIs, the ascii interface is just not very appealing. Fortunately, you can use cgoban as an interface.

I’ve been playing a couple of weeks, mostly on 9×9 boards because the standard 19×19 board takes too long to complete. Today, I played one of these matches. In the endgame, when both players passed, the program asks you to remove the dead stones and then calculates the score. This is such a screenshot:

My question now is: Why aren’t the two spaces (marked in red)  in the right lower corner and in the right upper corner not counted as territory for black?
As far as my understanding goes, they should be counted. Any ideas?

Tags: , , , , ,

One Response to “Is the counting in cgoban broken?”

  1. Hiroki Mori says:

    Hi, Chris,
    Those two black groups on the right side are dead as they are.
    They each have only one eye and there’s no space to make another thus white doesn’t even have to fill the liberties to capture these stones.
    In this case, at the end of the game, you may have to manually specify the dead stones by clicking on them because the software has some difficulties in judging the stones are dead or alive.
    In this game, white wins by a lot because all the right side are white’s territories and the 14 dead black stones have to be filled in black territories.