Re: [gnugo-devel] fuseki db

[Douglas Ridgway]

On Mon, 6 Dec 2004, Arend Bayer wrote:

> Gunnar wrote:
> 
> >      488 (actually dominates most alternatives based on amateur
> >      results)
> >      616 (does fine, is by far most popular)
> >      649 (ditto)
> >      714 (followup to 616. This move is better than ave, but see
> >      below)
> >      1000 (seems ordinary)
> 
> > > So should I make a patch putting those five patterns back in, or just edit
> > > the db files?
> > 
> > Either is fine with me but it would be good if someone more offered
> > their opinion on those patterns.
> 
> I don't like 616 and 649 for GNU Go. I think the others are fine.

Here's a patch putting back 488, 714, and 1000, the patterns no one has
spoken against, and including regression test patches for 1000 and 488.

doug.


--- patterns/fuseki19.db.orig   Tue Dec 14 20:43:57 2004
+++ patterns/fuseki19.db        Tue Dec 14 20:54:26 2004
@@ -48540,40 +48540,38 @@
 # # gf Removed. (3.5.3)
 # # Frowned upon, see gifu03:601.
 # 
+# dr Reinserted 2004-12-13
 #
-# Pattern F-H0-1377
-# # pre : 0x220d654c
-# # post: 0x235d1a72
-# # frequency: 20/196
-# # unique players: 9
-# # wins: 15/90
-#
-# # success: 75.0% vs 45.9% for this position overall, dchisq = 6.81132
-#
-# +-------------------+
-# |...................|
-# |...................|
-# |...................|
-# |...X...........X...|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...............*X..|
-# |...................|
-# |..O............O...|
-# |........O..........|
-# |...................|
-# |...................|
-# +-------------------+
-# 
-# :8,-,value(9)
+Pattern F-H0-1377
+# pre : 0x220d654c
+# post: 0x235d1a72
+# frequency: 20/196
+# unique players: 9
+# wins: 15/90
+# success: 75.0% vs 45.9% for this position overall, dchisq = 6.81132
++-------------------+
+|...................|
+|...................|
+|...................|
+|...X...........X...|
+|...................|
+|...................|
+|...................|
+|...................|
+|...................|
+|...................|
+|...................|
+|...................|
+|...................|
+|...............*X..|
+|...................|
+|..O............O...|
+|........O..........|
+|...................|
+|...................|
++-------------------+
 
+:8,-,value(9)
 
 Pattern F-H0-1378
 # pre : 0x220d654c
@@ -73207,38 +73205,39 @@
 #Pattern Fuseki1000
 ## 5/15
 #
-# Pattern F-H0-2066
-# # pre : 0x0ed9073c
-# # post: 0x319ca3dd
-# # frequency: 26/105
-# # unique players: 24
-# # wins: 18/60
-#
-# # success: 69.2% vs 57.1% for this position overall, dchisq = 1.55128
+# dr Reinserted 2004-12-13
+
+Pattern F-H0-2066
+# pre : 0x0ed9073c
+# post: 0x319ca3dd
+# frequency: 26/105
+# unique players: 24
+# wins: 18/60
 #
-# +-------------------+
-# |...................|
-# |...................|
-# |.............O.....|
-# |...O...........X...|
-# |.............*.....|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...............X...|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...O.....X.....X...|
-# |...................|
-# |...................|
-# |...................|
-# +-------------------+
-# 
-# :8,-,value(24)
+# success: 69.2% vs 57.1% for this position overall, dchisq = 1.55128
++-------------------+
+|...................|
+|...................|
+|.............O.....|
+|...O...........X...|
+|.............*.....|
+|...................|
+|...................|
+|...................|
+|...................|
+|...............X...|
+|...................|
+|...................|
+|...................|
+|...................|
+|...................|
+|...O.....X.....X...|
+|...................|
+|...................|
+|...................|
++-------------------+
+
+:8,-,value(24)
 
 
 Pattern F-H0-2067
@@ -82680,39 +82679,40 @@
 # # 4/26
 # Looks risky
 # 
-# Pattern F-H0-2329
-# # pre : 0x1a329e6b
-# # post: 0x01ae4faa
-# # frequency: 12/91
-# # unique players: 11
-# # wins: 8/44
-#
-# # success: 66.7% vs 48.4% for this position overall, dchisq = 1.61186
-#
-# +-------------------+
-# |...................|
-# |...................|
-# |...................|
-# |...X...........O...|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |...X...............|
-# |...................|
-# |...................|
-# |...................|
-# |...................|
-# |......X*...........|
-# |...X.XO............|
-# |.....O.........O...|
-# |...................|
-# |...................|
-# +-------------------+
+# dr Reinserted 2004-12-13
 #
-# :8,-,value(11)
+Pattern F-H0-2329
+# pre : 0x1a329e6b
+# post: 0x01ae4faa
+# frequency: 12/91
+# unique players: 11
+# wins: 8/44
 
+# success: 66.7% vs 48.4% for this position overall, dchisq = 1.61186
+
++-------------------+
+|...................|
+|...................|
+|...................|
+|...X...........O...|
+|...................|
+|...................|
+|...................|
+|...................|
+|...................|
+|...X...............|
+|...................|
+|...................|
+|...................|
+|...................|
+|......X*...........|
+|...X.XO............|
+|.....O.........O...|
+|...................|
+|...................|
++-------------------+
+
+:8,-,value(11)
 
 Pattern F-H0-2330
 # pre : 0x1a329e6b

Index: regression/nngs4.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/nngs4.tst,v
retrieving revision 1.63
diff -u -w -u -r1.63 nngs4.tst
--- regression/nngs4.tst        11 Sep 2004 18:55:12 -0000      1.63
+++ regression/nngs4.tst        15 Dec 2004 04:05:47 -0000
@@ -345,10 +345,12 @@
 800 reg_genmove white
 #? [E4]
 
-#O5 is bad
-loadsgf games/nngs/gnugo-3.5.2gf1-thi-200312210524.sgf 8
-810 reg_genmove white
-#? [D10|K4|K3|R6|Q6|Q2]
+# dr test removed 2004-12-13 same time as pattern Fuseki1000 
+#         aka pre 0x0ed9073c / post 0x319ca3dd was reinserted
+# #O5 is bad
+# loadsgf games/nngs/gnugo-3.5.2gf1-thi-200312210524.sgf 8
+# 810 reg_genmove white
+# #? [D10|K4|K3|R6|Q6|Q2]
 
 #some kind of base for the invasion is urgent
 loadsgf games/nngs/gnugo-3.5.2gf1-thi-200312210524.sgf 20

Index: regression/gifu03.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/gifu03.tst,v
retrieving revision 1.7
diff -u -w -u -r1.7 gifu03.tst
--- regression/gifu03.tst       7 Jan 2004 23:10:32 -0000       1.7
+++ regression/gifu03.tst       15 Dec 2004 04:11:28 -0000
@@ -183,10 +183,12 @@
 507 reg_genmove white
 #? [G13]
 
-# Q14 is not common in the Chinese fuseki. (But is it worth regressing over?)
-loadsgf games/cgf2003/GnuGo-KCC.sgf 7
-601 restricted_genmove black R12 O16 Q14
-#? [R12|O16|Q11|P14]
+## Q14 is not common in the Chinese fuseki. (But is it worth regressing over?)
+# dr deleted 2004-12-13 when reinserting pattern Fuseki488 / pre 0x220d654c
+# / post 0x235d1a72
+# loadsgf games/cgf2003/GnuGo-KCC.sgf 7
+# 601 restricted_genmove black R12 O16 Q14
+# #? [R12|O16|Q11|P14]
 
 # Making moyo here is much better than the game move F14. The
 # important thing is that O5 has some strategic effect on white while