gnugo-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gnugo-devel] Test cases


From: Gunnar Farneback
Subject: [gnugo-devel] Test cases
Date: Tue, 30 Dec 2003 20:49:03 +0100
User-agent: EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/20.7 (sparc-sun-solaris2.7) (with unibyte mode)

A couple of new test cases and test case revisions which have
accumulated while Savannah was down.

/Gunnar

Index: regression/13x13b.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/13x13b.tst,v
retrieving revision 1.12
diff -u -r1.12 13x13b.tst
--- regression/13x13b.tst       29 Nov 2003 19:33:36 -0000      1.12
+++ regression/13x13b.tst       30 Dec 2003 19:37:19 -0000
@@ -121,7 +121,7 @@
 10 reg_genmove white
 #? [!M6]
 
-# Most other moves are better than the game move at M6.
+# L6 urgent.
 loadsgf games/mertin13x13/GHGN2.sgf 52
 11 reg_genmove white
 #? [L6]
Index: regression/filllib.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/filllib.tst,v
retrieving revision 1.27
diff -u -r1.27 filllib.tst
--- regression/filllib.tst      10 Jul 2003 23:16:39 -0000      1.27
+++ regression/filllib.tst      30 Dec 2003 19:37:19 -0000
@@ -144,3 +144,7 @@
 loadsgf games/filllib12.sgf
 43 reg_genmove black
 #? [C6]
+
+loadsgf games/filllib13.sgf
+44 reg_genmove black
+#? [F4]
Index: regression/gunnar.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/gunnar.tst,v
retrieving revision 1.31
diff -u -r1.31 gunnar.tst
--- regression/gunnar.tst       29 Nov 2003 19:33:36 -0000      1.31
+++ regression/gunnar.tst       30 Dec 2003 19:37:20 -0000
@@ -236,6 +236,45 @@
 42 reg_genmove white
 #? [B19]
 
+# Q19 lets white make seki. T17 dies.
+loadsgf games/nngs/Minori-gnugo-3.5.2gf1-200312161433.sgf 130
+43 reg_genmove black
+#? [R18|S18]
+
+# T18 still gives seki. T17 dies.
+loadsgf games/nngs/Minori-gnugo-3.5.2gf1-200312161433.sgf 132
+44 reg_genmove black
+#? [T18]
+
+# The played move here was H13, caused by incorrect delta territory
+# but ultimately by a bug in break_through(). Naturally there is no
+# way for white to break through G13, H14, G15.
+loadsgf games/nngs/Minori-gnugo-3.5.2gf1-200312161433.sgf 254
+45 reg_genmove black
+#? [Q7|Q6|J16|J17|E7|N4]
+
+# Don't leave aji around, capture the white stones solidly.
+loadsgf games/nngs/joshj-gnugo-3.5.2gf1-200312171536.sgf 140
+46 restricted_genmove black O11 O9 N10 N9 M9
+#? [O11]
+
+# F15 is about 6 points plus some reverse followup. A4 is three points
+# reverse sente.
+loadsgf games/nngs/gnugo-3.5.2gf1-wingjk-200312301242.sgf 114
+47 restricted_genmove white F15 A4
+#? [F15]
+
+# Saving M13 is small compared to stopping a black incursion.
+loadsgf games/nngs/gnugo-3.5.2gf1-wingjk-200312301242.sgf 114
+48 restricted_genmove white M12 P11
+#? [P11]
+
+# In 3.5.2 the breakin code hallucinates that white A10 would suddenly
+# let black break into the territory around F6.
+loadsgf games/nngs/gnugo-3.5.2gf1-wingjk-200312301242.sgf 164
+49 reg_genmove white
+#? [A10]
+
 ############ End of Tests #################
 
 # Report number of nodes visited by the tactical reading
Index: regression/nando.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/nando.tst,v
retrieving revision 1.21
diff -u -r1.21 nando.tst
--- regression/nando.tst        29 Nov 2003 19:33:36 -0000      1.21
+++ regression/nando.tst        30 Dec 2003 19:37:20 -0000
@@ -57,7 +57,7 @@
 #CATEGORY=OWL/OPTICS
 loadsgf games/nando/auto009.sgf
 11 owl_attack S8
-#? [1 (S2|T2|S1)]
+#? [1 (S2|T2|S1|T4|T3)]
 
 #CATEGORY=OWL/OPTICS
 # Note that K12 connects out (sacrificing L11 if necessary).
Index: regression/ninestones.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/ninestones.tst,v
retrieving revision 1.38
diff -u -r1.38 ninestones.tst
--- regression/ninestones.tst   29 Nov 2003 19:33:36 -0000      1.38
+++ regression/ninestones.tst   30 Dec 2003 19:37:20 -0000
@@ -59,9 +59,10 @@
 150 reg_genmove black
 #? [C5|D5]
 
+# Black is actually trapped.
 loadsgf games/nngs/thor-gnugo-3.3.12-200212170003.sgf 64
 190 owl_defend J9
-#? [1 (H8|L9)]
+#? [0]
 
 loadsgf games/nngs/thor-gnugo-3.3.12-200212170003.sgf 108
 220 reg_genmove black
Index: regression/nngs2.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/nngs2.tst,v
retrieving revision 1.42
diff -u -r1.42 nngs2.tst
--- regression/nngs2.tst        29 Nov 2003 19:33:36 -0000      1.42
+++ regression/nngs2.tst        30 Dec 2003 19:37:20 -0000
@@ -78,8 +78,8 @@
 
 # same bug as century-2002:30
 loadsgf games/nngs/gnugo-3.3.2-Programmer-200206021522.sgf 52
-160 reg_genmove black
-#? [J14]*
+160 reg_genmove white
+#? [J14]
 
 # R9 is clearly worse than Q9.
 loadsgf games/nngs/kumu-gnugo-3.2-200205070903.sgf 116
@@ -115,10 +115,14 @@
 240 reg_genmove black
 #? [F17]
 
-loadsgf games/nngs/huck-gnugo-3.3.6-200209060837.sgf 196
-250 reg_genmove black
-#? [T18]*
+# # T18 is not very big, maybe a few points. There are clearly bigger
+# # points on the board. Removed.
+# loadsgf games/nngs/huck-gnugo-3.3.6-200209060837.sgf 196
+# 250 reg_genmove black
+# #? [T18]*
 
+# White can't kill or make seki but black may be forced to play
+# several moves inside own territory to remove white's stones later.
 loadsgf games/nngs/huck-gnugo-3.3.6-200209060837.sgf 204
 260 reg_genmove black
 #? [S18]*
@@ -186,10 +190,10 @@
 #? [J4|K2|K3|K4|K5|L2]*
 
 #CATEGORY=SEMEAI
-# If we move first, black can only get a ko in the semeai.
+# K1 wins the semeai. Other moves let black get ko.
 loadsgf games/nngs/gnugo-3.3.8-nailer-200209190045.sgf 212
 470 reg_genmove white
-#? [H6|H5|H3]*
+#? [K1]*
 
 # H4 only forces black to connect. H3 is locally more interesting, or
 # a move to secure life.
@@ -245,7 +249,7 @@
 
 loadsgf games/nngs/gnugo-3.3.9-nailer-200210192319.sgf 152
 580 reg_genmove white
-#? [K19]*
+#? [N19]*
 
 # We still can live.
 loadsgf games/nngs/gnugo-3.3.9-nailer-200210192319.sgf 224
Index: regression/nngs4.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/nngs4.tst,v
retrieving revision 1.52
diff -u -r1.52 nngs4.tst
--- regression/nngs4.tst        29 Nov 2003 19:33:37 -0000      1.52
+++ regression/nngs4.tst        30 Dec 2003 19:37:20 -0000
@@ -170,7 +170,7 @@
 
 # G12 allows white to live.
 loadsgf games/nngs/joshj-gnugo-3.3.16-200302232226.sgf 185
-430 reg_genmove black
+430 restricted_genmove black H10 G9 G10 G11 M6 L7 K7 G12 H12 F9 L12 M12
 #? [H10|G9|G10|G11|M6|L7|K7]*
 
 loadsgf games/nngs/juka-gnugo-3.3.16-200303031959.sgf 20
Index: regression/owl1.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/owl1.tst,v
retrieving revision 1.80
diff -u -r1.80 owl1.tst
--- regression/owl1.tst 29 Nov 2003 19:33:37 -0000      1.80
+++ regression/owl1.tst 30 Dec 2003 19:37:20 -0000
@@ -298,6 +298,18 @@
 333 owl_defend C17
 #? [0]*
 
+loadsgf games/nngs/joshj-gnugo-3.5.2gf1-200312171536.sgf 119
+334 owl_attack F5
+#? [1 (E7|F6)]
+
+loadsgf games/nngs/joshj-gnugo-3.5.2gf1-200312171536.sgf 180
+335 owl_defend K12
+#? [1 O6]
+
+loadsgf games/nngs/gnugo-3.5.2gf1-G28-200312282240.sgf 203
+336 owl_attack M5
+#? [1 (N2|O1)]
+
 ########### end of tests #####################
 
 # Report number of nodes visited by the tactical reading
Index: regression/reading.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/reading.tst,v
retrieving revision 1.63
diff -u -r1.63 reading.tst
--- regression/reading.tst      30 Dec 2003 13:20:10 -0000      1.63
+++ regression/reading.tst      30 Dec 2003 19:37:20 -0000
@@ -805,6 +805,11 @@
 193 defend D3
 #? [1 F3]
 
+# See also strategy:13.
+loadsgf games/incident108b.sgf 292
+194 attack L16
+#? [0]
+
 # Report number of nodes visited by the tactical reading
 10000 get_reading_node_counter
 #? [0]&
Index: regression/seki.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/seki.tst,v
retrieving revision 1.3
diff -u -r1.3 seki.tst
--- regression/seki.tst 29 Nov 2003 19:33:37 -0000      1.3
+++ regression/seki.tst 30 Dec 2003 19:37:21 -0000
@@ -116,7 +116,7 @@
 #? [B2]
 
 # Both A2 and B1 lead to seki. Strictly speaking A2 is a better move
-# since it leaves black the option to take sente and let black make a
+# since it leaves black the option to take sente and let white make a
 # thousand-year ko. For now we accept both.
 play white B2
 203 reg_genmove black
Index: regression/strategy.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/strategy.tst,v
retrieving revision 1.57
diff -u -r1.57 strategy.tst
--- regression/strategy.tst     29 Nov 2003 19:33:37 -0000      1.57
+++ regression/strategy.tst     30 Dec 2003 19:37:21 -0000
@@ -71,6 +71,7 @@
 #? [N4]
 
 # incident 113
+# See also reading:194
 loadsgf games/incident108b.sgf 292
 13 reg_genmove black
 #? [N18]
Index: regression/strategy4.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/strategy4.tst,v
retrieving revision 1.74
diff -u -r1.74 strategy4.tst
--- regression/strategy4.tst    29 Nov 2003 19:33:37 -0000      1.74
+++ regression/strategy4.tst    30 Dec 2003 19:37:21 -0000
@@ -8,8 +8,9 @@
 # strictly speaking tenuki is possible here. Q1 is of course bad
 # tm - Indeed GNU Go has finally figured out that tenuki is best!
 # tm - yes, Q1 is bad.  T5 also wastes a point.
+# gf - However, policy is to restrain thrashing dragons, so T5 is correct.
 151 reg_genmove white
-#? [!(Q1|T5)]*
+#? [T5]*
 
 loadsgf games/gwe.sgf 241
 152 reg_genmove white
@@ -179,7 +180,7 @@
 
 loadsgf games/xxlin.sgf 106
 184 reg_genmove black
-#? [H13|F15]*
+#? [H13|F15|Q13]
 
 # We may want to move this to a dedicated seki or semeai test suite in
 # the future.
@@ -202,10 +203,8 @@
 188 reg_genmove black
 #? [B6|C4]
 
-# The dragon at A2 is misclassified as inessential.
-
 loadsgf games/pooo.sgf 50
-189 reg_genmove black
+189 restricted_genmove black A14 C14 D14 E14 D15 E15 C19
 #? [E14]*
 
 loadsgf games/pooo.sgf 52
Index: regression/thrash.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/thrash.tst,v
retrieving revision 1.7
diff -u -r1.7 thrash.tst
--- regression/thrash.tst       5 Nov 2003 13:55:11 -0000       1.7
+++ regression/thrash.tst       30 Dec 2003 19:37:21 -0000
@@ -35,3 +35,18 @@
 8 reg_genmove black
 #? [H10]*
 
+# The N19 dragon should possibly be critical.
+loadsgf games/nngs/ccwills-gnugo-3.5.2-200312112026.sgf 95
+9 reg_genmove black
+#? [N12|N11]*
+
+# Don't let the dragon get out. It's in fact probably critical.
+loadsgf games/nngs/ccwills-gnugo-3.5.2-200312112026.sgf 125
+10 reg_genmove black
+#? [J15]*
+
+# White can't be allowed to thrash in the corner. (Clearly critical dragon.)
+loadsgf games/nngs/ccwills-gnugo-3.5.2-200312112026.sgf 173
+11 reg_genmove black
+#? [S3]*
+




reply via email to

[Prev in Thread] Current Thread [Next in Thread]