RE: [gnugo-devel] more owl tuning

[Portela Fernand]

Just a short note, as I just happened to see in a twogtp game
a similar position as the one described by :

> +Pattern D1386a
> +# gf New pattern. (3.3.17)
> +
> +OX.x       connect along edge to strong dragon
> +o*.O
> +----

The position in the game doesn't match, but it's close

OX.O
...O
----

The left O stone belongs to a living dragon, those on the
right (a 15 stones worm) were declared dead after the owl code
wasn't able to find a defense. The connection would have
worked.

Are there good reasons to disallow an O stone in the upper-right
corner of the D1386x patterns ? In case not, a corrective
patch is appended (sorry I can't test it on regressions, my box
is still busy running twogtp matches)

/nando

Index: patterns/owl_defendpats.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/owl_defendpats.db,v
retrieving revision 1.86
diff -u -r1.86 owl_defendpats.db
--- patterns/owl_defendpats.db  12 Feb 2003 11:16:31 -0000      1.86
+++ patterns/owl_defendpats.db  15 Feb 2003 19:06:20 -0000
@@ -6400,13 +6400,13 @@
 Pattern D1386a
 # gf New pattern. (3.3.17)

-OX.x       connect along edge to strong dragon
+OX.?       connect along edge to strong dragon
 o*.O
 ----

 :8,Eb,value(90)

-aX.x
+aX.?
 o*.b
 ----

@@ -6416,13 +6416,13 @@
 Pattern D1386b
 # gf New pattern. (3.3.17)

-OX.x       connect along edge
+OX.?       connect along edge
 o*.O
 ----

 :8,Eb,value(35)

-aX.x
+aX.?
 o*.b
 ----