RE: [gnugo-devel] min_value revision

[Portela Fernand]

I completely agree that the problem should be addressed (because I recently
observed a good number of similar cases in twogtp matches, see below). But
it seems to me that such mechanism is already coded :

/* If several moves have an identical minimum value, then GNU Go uses the
 * following secondary criterion (unless min_value and max_value agree, and
 * unless min_value is bigger than 25, in which case it probably comes from
 * a J or U pattern): 
 */
if (move[pos].min_value < 25)
  move[pos].min_value += tot_value / 200;

What I would like to question here is the reason why this is not applied
to J patterns. Changing to something like

if (move[pos].min_value < 39)

would allow for these to be differentiated too. U patterns are less of a
problem, since having 2 urgent moves in the same position is rather
unlikely. 

There must have been a good reason for keeping J patterns out, but I have
to confess that I wonder why. For the record, an example from a twogtp game
finished 10 mins ago :

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

O has just invaded at R17, X replied Q17. Unfortunately, both R16 and
C10 are "J" moves. And O ... tenukies :(
R16 has a "normal" value of 17.48, C10 gets valued 15.77

/nando