Re: [gnugo-devel] more break-in evaluation

[Arend Bayer]

Dan wrote:
> Arend wrote:
>
> > - Very few, maybe once in 8-10 games, we get a crucial improvement, i.e.
> >   s.th. between 10 pts more territory up to saving/keeping dead a
> >   20stone group. (Sometimes by gnugo guessing right really nice tesujis.)
>
> Could you post some examples of this?

Sure: ("Game move" is always the move played by CVS, i.e. with break-in.)

RikiTiki-gnugo-3.3.22-200307051101.sgf
Move 18 (black): GNU Go plays C8 (9.54) - Game move F8 (6.17)
./showboard.sh /home/arend/Go/gnugo-devel/interface/gnugo 
RikiTiki-gnugo-3.3.22-200307051101.sgf 18
=
   A B C D E F G H J
 9 . . . . . . . . . 9
 8 . . . . . . . . . 8
 7 . X O . O . X . . 7
 6 . X O . . O . . . 6
 5 . . X O + . O X . 5
 4 . . X O . O X . . 4
 3 . . X X . O X . . 3
 2 . . . . . . . . . 2     WHITE (O) has captured 0 stones
 1 . . . . . . . . . 1     BLACK (X) has captured 0 stones
   A B C D E F G H J

F8 seems _the_ most efficient way to protect the cut at H6.


Move 220 (black): GNU Go plays L13 (6.60) - Game move L11 (8.53)
./showboard.sh /home/arend/Go/gnugo-devel/interface/gnugo 
wildcat-gnugo-3.3.22-200307061953.sgf 220
=
   A B C D E F G H J K L M N O P Q R S T
19 O . O O O O O O X . X O . . . . . . . 19
18 X O . O X O O X X X O O O . O O X X . 18
17 X X O X X X O O X . O X X O X X O O . 17
16 . X X X . X O O X + O . X . O X X O . 16
15 X X X O O O X X O O . . . X X . . X . 15
14 O O O O X X . X . O X . . . . . X . . 14
13 O . X O X . X . O X . X . . . . . O . 13
12 X O X O O X X . . O X . . . . . O X . 12
11 X X O O X O . . O O . O . O X O X O . 11     WHITE (O) has captured 8 stones
10 O X X X X . . O . + O X . O X O . . . 10     BLACK (X) has captured 8 stones
 9 . X . . . . O X O O X . X O X O . . . 9
 8 . X . O . . O . X X X . . O X X O O . 8
 7 . O . O X . O X . . O O O X . . X . . 7
 6 . . O X X . O . X . . . . X . X . . . 6
 5 . . O . . X O . X O . . . . . . . X . 5
 4 . O X X X O . O X O O . . O . X X O . 4
 3 . . O O X O . O O O X O . O X . O . . 3
 2 . . O X X X X X X X X . O X X . . . . 2
 1 . O X . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

L11 saves the big group at J8. Even in sente (well, in the game, gnugo
even go the cut at N8).

Move 66 (black): GNU Go plays M5 (11.91) - Game move N17 (7.04)
./showboard.sh /home/arend/Go/gnugo-devel/interface/gnugo 
Zahlman-gnugo-3.3.22-200307050045.sgf 66
=
   A B C D E F G H J K L M N O P Q R S T
19 . . . . . . . . . . . . . . . . . . . 19
18 . . . . . . . . O . . . . O X . . . . 18
17 . . . . . . . X . . . . . O X . . . . 17
16 . . . X . X . . . O . . O X . X . . . 16
15 . . X . . . . . . . . . . . . . . . . 15
14 . . O X . . . . . . . . . . . X . . . 14
13 . . O . . . . . . . . . . . . . . . . 13
12 . . . . . . . . . . . . . . . X O X . 12
11 . . O . . . . . . . . . . . X O X . . 11     WHITE (O) has captured 1 stones
10 . . . + . . . . . + . . . . X O X . . 10     BLACK (X) has captured 1 stones
 9 . . . . . . . . . . . . . . . O . X . 9
 8 . . O . . . . . . . . . . . O O O O . 8
 7 . . . . . . . . . . . . . O X X X . . 7
 6 . . X . . . . . . . . . . . O O X . . 6
 5 . . . . . X . . X . X . . . . . X X . 5
 4 . . . X . . . X . X O . . . . O O O . 4
 3 . . X . . O . O X O . O . . . . . . . 3
 2 . . X O . . . O O . O . . . . . . . . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

And here two typical examples of the 1-5pt category:

Move 202 (white): GNU Go plays D3 (15.67) - Game move E4 (13.33)
./showboard.sh /home/arend/Go/gnugo-devel/interface/gnugo 
gnugo-3.3.22-nailer-200307051649.sgf 202
=
   A B C D E F G H J K L M N O P Q R S T
19 . . . . . . . . . . . . . . . . . O . 19
18 . . . . . . O . X O . . . . . O O X X 18
17 . . . O O O X . X O . O . O . O X . . 17
16 O O . + . X X . . X O . . . . X X . . 16
15 O O O O X X X . . X . . . O . . . . . 15
14 X O X X O X O X X . . X X X X . X X . 14
13 . X X O O O O X . . . O O O . X O O . 13
12 X X O O O . O X X . . . X O O O X O . 12
11 . X X O . X O X . . X X X X X O X . . 11     WHITE (O) has captured 6 stones
10 X O X X X X X O O + . O . O O X O O . 10     BLACK (X) has captured 6 stones
 9 O O O O X O O O . . . . O . X X O . O 9
 8 . X . O O X X O . . . . O X X . X O . 8
 7 . . X . X O X X X O . . . O O X X X . 7
 6 . X . X X O O X O . . . . . . O O O . 6
 5 . X X O O X O O . . O . . . . O X . . 5
 4 . . . + . X . . . X O . . . . O X . O 4
 3 . . X . O . . O . X X . O O O X . X O 3
 2 . . X O . . . . O O X X X X X . X . X 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

This one is very subtle. If white plays D3, black can break in upt to F1
with the hane at D1 and the atari at D4. Therefor, E4 is a little better
territorially.

Move 90 (black): GNU Go plays N2 (32.98) - Game move M4 (29.27)
./showboard.sh /home/arend/Go/gnugo-devel/interface/gnugo 
wildcat-gnugo-3.3.22-200307061937.sgf 90
=
   A B C D E F G H J K L M N O P Q R S T
19 . . . . . . . . . . . . . . . . . . . 19
18 . . . . . . . . . . . . . . X . O . . 18
17 . . X X . . . X O . X . . . O X O . . 17
16 . . O X X . . X O X . . . X X X . . . 16
15 . . O . . . O O X X . . . . . . O O . 15
14 . . O X X O . . O X O . . . . . O X O 14
13 . . . O O . . . O O X X . . . . O X . 13
12 . . . . . . . . . . O . . . . O X . X 12
11 . . . . . . . . . . . . . . . O X X . 11     WHITE (O) has captured 0 stones
10 . . X + O . . . . + . . . . . O X . . 10     BLACK (X) has captured 1 stones
 9 . . . . . . . . . . . . . . . O X . . 9
 8 . . . . . . . . . . . . X . O X X X . 8
 7 . . X . . . . . . . . . . . . . . O . 7
 6 . . . . . . . . . . . . X X O . O . . 6
 5 . . . . . . X . . . . O . . O . . . . 5
 4 . . X X X . . . X + . . O X . X O . . 4
 3 . . . O X O . O . . X O X O . X O . . 3
 2 . . . O O . . . . . . X . . . . . . . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T
It requires some careful reading, but M4 looks clearly better to me.
Point is that black gets both the cut at N5 (or captures N4) _and_ can
save Q4. Of course this is a little lucky for the break-in code, but not
completely atypical.


(And if you wonder about the output, yes, I did write a script to
automatically generate ascii boards from the --replay <color> output.)

Arend