Re: [gnugo-devel] Scoring pb + GTP questions

[Gunnar Farneback]

Max wrote:
> 1/
> It is probably a known problem that GG does not evaluate correctly this
> position and fails in scoring the game (Go++ and Wh have same problem too )

No, this is not a case which would be expected to fail.

   A B C D E F G H J K L M N O P Q R S T
19 . . . . X O O . . . X . O . O O X O . 19
18 . . . . X X O . . O X O O O O X X . . 18
17 . . . . X O O O . O O X X O O X X X X 17
16 . . . X X O X X O O X X X X X X O X . 16
15 . . . X O O X X X X X O X O X O O O X 15
14 . . . X X O X O . . X O X O O . . O X 14
13 . . . X O O O X X X X O X X O . . O X 13
12 . . . X X O . O O X X O O O O O O O X 12
11 . . . X O . O X O O O O O X X X X X . 11
10 . . . X O O O X X O O X O O O O X . X 10
 9 X X . . X X X O X X O X O O . O X X O 9
 8 O X . . X O O O X X X X X O O X X O O 8
 7 O X . . X O O O O X X X X X O O O O X 7
 6 O O X X X X X O O X . X X X X X X X X 6
 5 . . O O O X O O O X X X X O O X O X . 5
 4 . O . O . O . O X X O X O . . O O O X 4
 3 . . . . . O . O X X O O O O O O . O X 3
 2 . . . . . O . O X X X O O X X O O X . 2
 1 . . . . . . O X X X . X X . X X X . X 1
   A B C D E F G H J K L M N O P Q R S T

Although the black dragon in the lower right is topologically
one-eyed, it is unconditionally alive by invincibility analysis.
Apparently we have some bug here. Both --score estimate and --score
final give W+115.5. Even more interesting is --score aftermath which
returns B+9.5, somehow reaching the result that while black is alive,
the enclosed white dragon only lives in seki, with O4, P4, and R3 as
dame points. I'll investigate this.

> 2/
> It should be usefull that GG returns something like "pass"  to the top_moves
> GTP command when the move list is empty (end of game) instead of saying
> nothing.

I don't agree. It would require more special case code in typical
applications.

> 3/ Is there a GTP command which tell GG to do a  _basic_ evaluation
> for a given move (not included in the top_moves list ).

No, there's infrastructure missing for that. The best you can do is to
issue a "restricted_genmove color vertex" followed by top_moves. In
some situations the recently introduced all_move_values command might
be an alternative, but the semantics are different (specifically the
endgame patterns cutoff).

/Gunnar