Re: [gnugo-devel] Re: [computer-go] Competition of the GO-algorithms

[Gunnar Farnebäck]

I wrote:
> Thanks. The major problem here is that C11 isn't tested as an
> additional owl defense for B16. It seems most appropriate to force
> that by adding a connection move reason, so I propose the following
> revision of ED91:

Appended as patch below. No effects on the regressions.

- tuning

/Gunnar

Index: patterns/patterns.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/patterns.db,v
retrieving revision 1.134
diff -u -r1.134 patterns.db
--- patterns/patterns.db        24 Nov 2004 04:35:51 -0000      1.134
+++ patterns/patterns.db        4 Dec 2004 20:03:50 -0000
@@ -10928,25 +10928,23 @@
 
 Pattern ED91
 # db New pattern (3.1.28)
+# gf Changed to connection pattern. (3.7.2)
 
-???..o?          Extend after the push
-???XXoo
-?O..O*.
-?O....o
-o......
--------
+??XXoo          Extend after the push
+O..O*.
+O....o
+......
+------
 
-:8,j
+:8,C
 
-???..o?
-???XXfo
-?O.aO*.
-?Ocbd.o
-o.e....
--------
+??XXoo
+a..O*.
+a....o
+......
+------
 
-; oplay_attack(*,a,b,c,d,e,e) || !oplay_defend(*,a,b,d,c,d)
-> replace(f,*)
+;!oplay_disconnect(*,*,a)
 
 
 Pattern ED92