[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] problem with score feature of 3.2
From: |
bump |
Subject: |
Re: [gnugo-devel] problem with score feature of 3.2 |
Date: |
Sun, 23 May 2004 06:04:05 -0700 |
> The first two positions involve sekis. I looked at the
> third. At the end of the game, the problem is that the
> D2 stone is not considered ``captured territory''
> because the status of its neighbor the A8 dragon is not
> ALIVE. The position looks like this:
I see that the e-mail that I sent to the list two days ago
finally made it.
In the meantime I looked at the other three positions submitted
by Adam. The patch doesn't fix any of them.
The first two involve sekis, and I haven't looked at them
closely. The fourth is not fixed by the patch and has to do with
nested inessential strings. This is the position after both
players add one stone:
A B C D E F G H J
9 X X X X . X . X X 9
8 X O O O X . O O O 8
7 . O + O O X X X X 7
6 . . . . . X X O . 6
5 O O O . O X O O . 5
4 O X X O O O X . . 4
3 X X + X X O X O . 3
2 . . X X O . O X . 2
1 . . X O O O . . . 1
A B C D E F G H J
Inside the dead B dragon are nakade stones at J8, which
are inessential. The problem is that the B stones at J9
are also counted inessential. The function
captured_territory() should recognize that J9 is captured
territory but J8 is not. I do not yet have a patch for
this.
As Gunnar noted, --score aftermath does get this problem
right and is generally pretty reliable for scoring at
the end of the game. The algorithms in score.c can
also be used to get a score in the middle game so
they address a different purpose. This file is a
rewrite of some old code of Jerome Dumonteil based on
Bouzy's dissertation, really obsoleted by influence.c
but kept around anyway.
Dan