[gnugo-devel] Strategic values too high?

[SP LEE]

Here is a recent game I played with gnugo. Obviously black should play at 
B14 to connect its group, worth 28 points. However, as I experienced before, 
gnugo put a rather high strategic value on G12 and played on it.

loadsgf 0624.sgf 112
= black

showboard
=
  A B C D E F G H J K L M N O P Q R S T
19 . . . . . . . . . . . . . . . O X . . 19
18 . . X O . . O X . X X O . . X O O X . 18
17 . X X O . O . . . . . O . . X O X X . 17
16 . X O X . . . . . X X X O O O + O . . 16
15 . O O X O . O . O . O X O . X O . . . 15
14 . . X X O . . . O . O O X . X X O . . 14
13 . X O O . . . . . . X X . . . . O . . 13
12 . . X O . . . . . . . . . . . X X . . 12
11 . . X . . . . . O O . . . . X . . . . 11     WHITE (O) has captured 3 
stones
10 . . . X . . . X . X . O . . . X . . . 10     BLACK (X) has captured 0 
stones
9 . . . . . . . X . . X O . . . . . . . 9
8 . . . . . . . . . . . . . . . . . . . 8
7 . . . X . O . . . . . . . . . X . . . 7
6 . . . . . X O . . . . . . . . . . . . 6
5 . . . . . X O . . O . O . . X O . . . 5
4 . . X X X O O . O + O . O . . X O . . 4
3 . . X O O O X X X O X O O X . X X . . 3
2 . . . . O X . . . X X X O . . . . . . 2
1 . . . . . . X . . . . . . . . . . . . 1
  A B C D E F G H J K L M N O P Q R S T

 B14: owl attack/defend for D16
 B14: owl attack/defend for C16
 B14: 20.97 - change in territory
 B14:   8.43 - C16 strategic value already counted - A.
 B14:   10.08 - D16 strategic value already counted - A.
 B14:   2.81 - total followup value, added 2.81 as territorial followup
 B14: 1.40 - added due to followup (2.81) and reverse followup values 
(0.00)
 B14: 2.59 - shape (shape values +3.00(1) -0.00(0), shape factor 1.158)
 B14: 1.09 - connects strings (connect value 10, shape factor 0.219)
Move generation values B14 to 26.06

 G12: 6.66 - change in territory
 G12: 6.38 - strategic effect on J11
 G12: 9.14 - strategic effect on H10
 G12:   1.60 - total followup value, added 1.60 as territorial followup
 G12: 0.80 - added due to followup (1.60) and reverse followup values 
(0.00)
 G12: 3.62 - shape (shape values +3.00(1) -0.00(0), shape factor 1.158)
Move generation values G12 to 26.60

If we look at the individual move reasons of G12, the values seem to be 
reasonable, but I can't see in any case that the total value is 26.6. 
Therefore I think it's still questionable how we should add the strategic 
value to the final value of a move.

For example, if a move has threats on two opponent groups at the same time, 
there is a good chance that the opponent will save one of the group and only 
the smaller group can be attacked. So their strategic values can't be simply 
added.

SP Lee

_________________________________________________________________
Add photos to your e-mail with MSN 8. Get 2 months FREE*.  
http://join.msn.com/?page=features/featuredemail

0624.sgf
Description: Binary data