[gnugo-devel] owl defense pattern class "B"

[Arend Bayer]

I think one of my centerum censeo's on this list is that patterns
should not have too much power. (I am aware that this is a debatable
opinion and will try to be careful with concrete decisions based on
it.) Here is another example:

I find it hard to imagine many circumstances under which the
"B" classification in owl_defendpats.db (claiming that
the played stone is definitely not connected to the goal dragon)
can make sense. Short of trying to the effect of changing all "B"
to "b" ( = possibly not connected to the goal), I added the following
simple "heuristic" to owl_update_goal():

If the stone is part of a string of which any stone belongs to the
goal, then the new stone (and, if necessary, the rest of the string)
will get added to the goal.

This is enough to ensure that we always have strings either completely
in or not at all.

owl:262         FAIL 1 E8 [0]
nngs:875        PASS C5 [C5]
ninestones:190  FAIL 1 H8 [0]

In all three cases the cvs version doesn't even try a simple
defense move because an essential stone didn't get added to the goal:

nngs:875
O first:
   A B C D E F G H J
 9 . . . . . . . . . 9
 8 . . . . . . . . . 8
 7 . . X . . . X . . 7
 6 . 2 . . . . . . . 6
 5 . 4 1 5 + . . . . 5
 4 . X O X a . . . . 4
 3 . 3 X O . . X . . 3
 2 . 8 6 7 . . . . . 2     WHITE (O) has captured 0 stones
 1 . . . . . . . . . 1     BLACK (X) has captured 0 stones
   A B C D E F G H J

After this sequence, CVS does not even consider "a", as 5 did not get
added to the goal.

ninestones:190
X first:
   A B C D E F G H J K L M N O P Q R S T
19 . . . . . . . . . . . . . . . . . . . 19
18 . . . . . . . . . . . . . . . O . . . 18
17 . . . . . . . . . . . . O O . . . . . 17
16 . . . X . . X . . X . X O X X X . . . 16
15 . . . . . . . . . . . X X O . . . . . 15
14 . . . . . . . . . . . . . O . X O . . 14
13 . . . . . . . . . . . . . . . O . . . 13
12 . . . . . . . . . . . . . . . . . . . 12
11 . . . . . . . . O . . . . . . . . . . 11   
10 . . . X . . . X O X X . . . . X . . . 10  
 9 . . . . . . 2 O X O . . . . . . . . . 9
 8 . . . . . . . 1 X O . O . O . . . . . 8
 7 . . . . . . 3 . X X O . O X . . . . . 7
 6 . . . X . a . O O X X O O X . . X . . 6
 5 . . . . . . . O X X X O X . X . . . . 5
 4 . . . X . O O X X X O O X X . X . . . 4
 3 . . . . . . . O O O X X . . . . . . . 3
 2 . . . . . . . . . . . . . . . . . . . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

Again CVS doesnt consider 3. Not that 3 is a good move, but not having
1 in the goal is a very bad reason for trying this.
(Unfortunately, gnugo doesnt find a better reply than "a" after this.)

Arend


Total nodes: 1566796001 2795743 10986434