[gnugo-devel] Re: [computer-go] Competition of the GO-algorithms

[Gunnar Farnebäck]

I wrote, on the computer-go list:
> Btw, of the seven games I checked four were unnecessary reinforcement
> moves by GNU Go and three were missing teire.

The context is a scoring competition at
http://rusgolib.iponweb.net/AleksandrMel'nikov.html

The page has a link to the file Comp_CPP.zip which in addition to some
code contains 50 example games. In seven of those GNU Go doesn't agree
on the score. In four of the games GNU Go adds an unnecessary
reinforcement move. The four mistakes in detail are:

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

White adds a move at T12 because it misreads life and death of an
eyeshape with a cutting point (well-known problem).

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

White plays a gote move inside black's territory at S10. The problem
is a rather bogus ATTACK_EITHER move reason and that it is taken for
granted that attack_either moves are safe. There's already a FIXME on
line 1097 of value_moves.c regarding the latter. This would be nice to
solve.

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

Pattern ED91 causes black to add a stone at C2. There are many ways in
which the pattern could be revised but I propose removing it entirely.
These days the move valuation should be able to understand the
importance of such moves without an explicit j pattern.

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

After the dame at E6 has been filled, black makes an unnecessary
reinforcement at F9. The move is generated as a defend_both move but
since G8 can be sacrificed without white getting alive, it is bogus.

/Gunnar