[gnugo-devel] trevor_1_14.2

[Trevor Morris]

As part of this tuning exercise, I removed a few patterns that
didn't make sense to me or were causing problems.  The removal
helped or made no difference, so feeling that less is more, it 
seems good.


http://www.public32.com/games/go/trevor_1_14.2
 - tuning.
 - updated regress.pl
 - regression revisions/corrections


Index: patterns/conn.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/conn.db,v
retrieving revision 1.8
diff -u -b -B -r1.8 conn.db
--- patterns/conn.db    2001/10/30 16:42:00     1.8
+++ patterns/conn.db    2001/11/09 20:09:41
@@ -952,18 +954,18 @@
 # CC30x - one space jump connections
 #
 ####################################
-
-Pattern CC301
-
-O.O
-...
-
-:8,C
 
-OaO
-dbc
-
-;xplay_attack(a,b,c,c) && xplay_attack(a,b,d,d)
+#Pattern CC301
+#
+#O.O
+#...
+#
+#:8,C
+#
+#OaO
+#dbc
+#
+#;xplay_attack(a,b,c,c) && xplay_attack(a,b,d,d)
 
 
 Pattern CC302
Index: patterns/connections.c
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/connections.c,v
retrieving revision 1.17
diff -u -b -B -r1.17 connections.c
--- patterns/connections.c      2001/11/09 19:03:43     1.17
+++ patterns/connections.c      2001/11/09 20:09:43
@@ -197,8 +197,14 @@
               && dragon[pos].origin != first_dragon) {
        second_dragon = dragon[pos].origin;
        /* A second dragon found, we amalgamate them at once. */
-       TRACE("Pattern %s joins dragons %1m, %1m\n",
-             pattern->name, first_dragon, second_dragon);
+       /* Want this output if verbose or DEBUG_DRAGONS is on, but not both.*/
+       if (verbose) {
+         TRACE("Pattern %s joins %C dragons %1m, %1m\n",
+             pattern->name, color, first_dragon, second_dragon);
+       } else {
+         DEBUG(DEBUG_DRAGONS, "Pattern %s joins %C dragons %1m, %1m\n",
+             pattern->name, color, first_dragon, second_dragon);
+       }
        fflush(stdout);
        join_dragons(second_dragon, first_dragon);
        /* Now look for another second dragon. */
Index: patterns/owl_attackpats.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/owl_attackpats.db,v
retrieving revision 1.21
diff -u -b -B -r1.21 owl_attackpats.db
--- patterns/owl_attackpats.db  2001/11/07 19:33:39     1.21
+++ patterns/owl_attackpats.db  2001/11/09 20:09:49
@@ -3575,6 +3575,43 @@
 :8,s,value(40)
 
 
+Pattern A1342
+# tm new pattern (3.1.14)
+
+|.XO
+|..X
+|*.X
+|.XX
+|.O?
+
+:8,s,value(50)
+
+
+Pattern A1343
+# tm new pattern (3.1.14)
+# See strategy4:208
+
+|.Xx    1-2 point often vital in this shape
+|.oX
+|.*.
++---
+:8,s,value(50)
+
+
+Pattern A1343
+# tm new pattern (3.1.14)
+# See strategy4:208
+
+|XX?     Try destroying X's eye.
+|.X?
+|*Xx    
+|OoX
+|...
++---
+
+:8,s,value(50)
+
+
 #########################################################
 #                                                       #
 #               Invasion patterns                       #
Index: patterns/owl_defendpats.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/owl_defendpats.db,v
retrieving revision 1.22
diff -u -b -B -r1.22 owl_defendpats.db
--- patterns/owl_defendpats.db  2001/11/07 19:33:39     1.22
+++ patterns/owl_defendpats.db  2001/11/09 20:09:58
@@ -707,6 +707,7 @@
 O.X?     expand eye-space
 O.*.
 .x..
+----
 
 :8,s,value(10)
 
@@ -3645,6 +3645,20 @@
 
 ; owl_escape_value(a)>0 &&
 ; oplay_attack(*,a,c)
+
+
+
+Pattern D1136
+#tm new pattern (3.1.14)
+#See strategy:206
+
+OOO    very low value - can't hurt to try!
+..X
+X*.
+???
+
+:8,b,value(5)
+
 
 
 ########################################################
Index: patterns/patterns.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/patterns.db,v
retrieving revision 1.22
diff -u -b -B -r1.22 patterns.db
--- patterns/patterns.db        2001/11/01 23:38:30     1.22
+++ patterns/patterns.db        2001/11/09 20:10:22
@@ -6456,6 +6456,25 @@
 ;!omoyo(a)
 
 
+Pattern EB1808a
+
+??.??              defend!
+O...O
+..*..
+.....
+-----
+
+:8,OXceb
+
+??.??
+O...O
+..a..
+.....
+-----
+
+;!omoyo(a)
+
+
 Pattern EB1809
 
 ..X..              secure fourth line territory
@@ -7255,6 +7274,16 @@
 :8,OXb
 
 
+Pattern CB220
+
+*.
+XO     hane at head of two stones.
+XO
+..
+
+:8,shape(2)
+
+
 Pattern CB221
 
 ????O?            hane
@@ -12025,21 +12054,23 @@
 ; !dead(A)
 
 
-Pattern EJ83
-
-..OXOO        make shape and defend cut
-.*.O..
-......
-------
-
-:8,J
 
-..OXCC
-.*bOa.
-......
-------
-
-; xplay_attack(a,b,C)
+#Pattern EJ83
+#This pattern breaks arb:101 - let's remove it and see what happens!
+#
+#..OXOO        make shape and defend cut
+#.*.O..
+#......
+#------
+#
+#:8,J
+#
+#..OXCC
+#.*bOa.
+#......
+#------
+#
+#; xplay_attack(a,b,C)
 
 
 Pattern EJ84
@@ -12754,12 +12785,12 @@
 :8,OX,value(12)
 
 ?Xxx
-O...
+A...
 ..*.
 ....
 ----
 
-; xmoyo(*)
+; xmoyo(*) && lib(A) > 2
 
 
 Pattern LE6
Index: regression/arb.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/arb.tst,v
retrieving revision 1.13
diff -u -b -B -r1.13 arb.tst
--- regression/arb.tst  2001/11/09 17:46:17     1.13
+++ regression/arb.tst  2001/11/09 20:10:22
@@ -12,7 +12,7 @@
 
 loadsgf games/arb/game01.sgf 53
 101 gg_genmove black
-#? [!PASS]
+#? [(C2|D2|F2|G2|F1|D1|E1)]
 
 loadsgf games/arb/game01.sgf 54
 102 gg_genmove white
Index: regression/buzco.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/buzco.tst,v
retrieving revision 1.6
diff -u -b -B -r1.6 buzco.tst
--- regression/buzco.tst        2001/11/07 05:41:30     1.6
+++ regression/buzco.tst        2001/11/09 20:10:23
@@ -18,7 +18,7 @@
 
 loadsgf games/buzco1.sgf 70
 5 gg_genmove black
-#? [C10|C11|E9|A4]*
+#? [C10|C11|E9|A4|F10]*
 
 loadsgf games/buzco1.sgf 82
 6 gg_genmove black
Index: regression/capture.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/capture.tst,v
retrieving revision 1.4
diff -u -b -B -r1.4 capture.tst
--- regression/capture.tst      2001/10/21 19:37:48     1.4
+++ regression/capture.tst      2001/11/09 20:10:23
@@ -76,8 +76,9 @@
 
 # This test case originally said "defend G5", which must be an error.
 # Changed it to defend the weak stone at F5 instead. /gf
+# F6 works to defend F3 -trevor
 loadsgf golois/web000103.sgf
 18 defend F5
-#? [0]*
+#? [1 F6]*
 
 
Index: regression/regress.pl
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/regress.pl,v
retrieving revision 1.9
diff -u -b -B -r1.9 regress.pl
--- regression/regress.pl       2001/11/07 19:33:39     1.9
+++ regression/regress.pl       2001/11/09 20:10:27
@@ -155,7 +155,8 @@
 }
 
 if (!$goprog) {
-  $goprog = "../interface/gnugo.exe --mode gtp --quiet -t";  
+  my $s = (lc ($^O) eq 'mswin32') ? '\\' : '/';
+  $goprog = "..${s}interface${s}gnugo.exe --mode gtp --quiet -t -w
-d0x1000";  
 }
 
 die $helpstring unless defined $goprog;
@@ -181,7 +182,8 @@
 $goprog_out = new FileHandle;          # stdout of computer player
 $goprog_err = new FileHandle;          # stdout of computer player
 print "Go program: $goprog\n" if $verbose > 1;
-$pidg = open3($goprog_in, $goprog_out, $goprog_err, $goprog);
+$pidg = open3($goprog_in, $goprog_out, $goprog_err, $goprog)
+  or die "Couldn't launch GNU Go: $!";
 print "goprog pid: $pidg\n" if $verbose > 1;
 my ($goprog_exe) = split (" ", $goprog);
 $goprog_timestamp = (stat $goprog_exe)->mtime;
@@ -240,11 +242,6 @@
   $testfile = shift;
   print "$testfile\n" if $verbose;
   
-  foreach (@counters) {
-    go_command("reset_${_}_counter");
-    eat();
-  }
-
   -e "html" or mkdir "html" or die "Couldn't create html";
   -e "html/$testfile" or mkdir "html/$testfile" or die "Couldn't create
html/$testfile";
   
@@ -264,7 +262,7 @@
         if (/^FINISHED PROBLEM: (.*)/) {
           close TRACER or die "Couldn't close temp trace file";
           my $thisgtp = $1;
-          my ($num) = $thisgtp =~ /^(\d+)/ or die "Missing number: $thisgtp";
+          my ($num) = $thisgtp =~ /^(\d+)/ or confess "Missing number:
$thisgtp";
           rename "tracer.ttt", "$num.trace"
               or die "Couldn't rename tracer: $testfile, $num";
           open (TRACER, ">tracer.ttt");
@@ -274,8 +272,11 @@
       exit;
   }
   }
-  
   
+  foreach (@counters) {
+    go_command("reset_${_}_counter");
+    eat();
+  }
 
   #main bit.
   $pidt = open ($testfile_out,"<$testfile");
@@ -406,9 +407,12 @@
 
   unless ($one_gg_process) {
     go_command("quit");
-    print "waiting\n" if $verbose > 2;
+    print "waiting\n" if $verbose > 1;
     waitpid $pidg, 0;
-    print "done waiting\n" if $verbose > 2;
+    print "done waiting\n" if $verbose > 1;
+    print "waiting on child\n" if $verbose > 1;
+    waitpid $childpid, 0;
+    print "done waiting on child\n" if $verbose > 1;
   }
 }
 
@@ -720,7 +724,7 @@
   # ignore empty lines
     my $line = "";
     while ($line eq "") {
-    chop($line = <$goprog_out>) or die "No response!";
+    chop($line = <$goprog_out>) or confess "No response!";
     $line =~ s/\s*$//smg;
     }
   <$goprog_out>;
Index: regression/regress.plx
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/regress.plx,v
retrieving revision 1.3
diff -u -b -B -r1.3 regress.plx
Index: regression/strategy4.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/strategy4.tst,v
retrieving revision 1.22
diff -u -b -B -r1.22 strategy4.tst
--- regression/strategy4.tst    2001/11/09 17:46:17     1.22
+++ regression/strategy4.tst    2001/11/09 20:10:31
@@ -111,6 +111,9 @@
 172 gg_genmove white
 #? [!N9]
 
+#CATEGORY=KO
+#DESCRIPTION=GNU Go lacks sufficient Ko understanding
+#SEVERITY=4
 # There's no way to win the ko by making atari from above. This move
 # must be at A9. This is a good example showing why only having the
 # difference between a ko where we make the first threat or not is
@@ -146,13 +149,17 @@
 178 gg_genmove black
 #? [P4|P3]
 
+#DESCRIPTION=Tenuki also seems OK here.
+#Added E2 (3.1.14) -trevor
 loadsgf games/xxlin.sgf 20
 179 gg_genmove black
-#? [P3|O2|P2|O4]*
+#? [P3|O2|P2|O4|E2]*
 
-loadsgf games/xxlin.sgf 30
-180 gg_genmove black
-#? [Q8]*
+#The point of this problem is that Q8 is locally better than M7.
+#However, GNU Go reasonably tenukis.
+#loadsgf games/xxlin.sgf 30
+#180 gg_genmove black
+##? [Q8]*
 
 loadsgf games/xxlin.sgf 38
 181 gg_genmove black
@@ -227,21 +234,25 @@
 196 gg_genmove black
 #? [M17]*
 
+#Tenuki at S18 is not unthinkable. (3.1.14) -trevor
 loadsgf games/bretz.sgf 130
 197 gg_genmove black
-#? [K3]*
+#? [K3|S18]*
 
+#Tenuki at S18 is not unthinkable. (3.1.14) -trevor
 loadsgf games/bretz.sgf 148
 198 gg_genmove black
-#? [C10]*
+#? [C10|S18]*
 
+#Tenuki at S18 is not unthinkable. (3.1.14) -trevor
 loadsgf games/bretz.sgf 164
 199 gg_genmove black
-#? [N5]*
+#? [N5|S18]*
 
+#Tenuki at S18 is not unthinkable. (3.1.14) -trevor
 loadsgf games/bretz.sgf 212
 200 gg_genmove black
-#? [P6|P7|Q7]*
+#? [P6|P7|Q7|S18]*
 
 loadsgf games/juka1.sgf 8
 201 gg_genmove black
@@ -263,6 +274,11 @@
 205 gg_genmove black
 #? [!O11]*
 
+#CATEGORY=SEMEAI
+#DESCRIPTION=K6 does take K8 to no eyes, but...
+#SEVERITY=8
+#GNU Go misses that K10 (large B group) has way few liberties
+#than O9 group, hense missing importance of cutting off K8.
 loadsgf games/strategy45.sgf
 206 gg_genmove black
 #? [L8]*