[gnugo-devel] further test revision

[Arend Bayer]

Here are some more tests that I disagree with. Nicklas5:1203 and 
Nicklas4:1205 are somewhat debatable, and I would appreciate a 2nd
opinion on them.

-Arend

arb.tst:
200: Agree with Gunnar. I suggest to remove this problem, as there are too
many plausible moves (M11, H4).

Nicklas2:
1001: H8 seems to be a little bigger than B7. W H8 would kill B.
1803: A6 is a nice tesuji.

Nicklas4:
1103: C2 just as good as C3.
1201: As Gunnar commented already, P9 is probably a typo, Q9 correct. (Btw
there is a logical reason in GNU Go's move valuation here: It sees that
S7 defends R9, and that it defends Q8. From this it seems to conclude
that it can defend both.)
1205: E16 requires courage, as W can attack with W N11, B N12 and next
W L9 or K10. I don't know what is correct, nor do I know what would be
the best advice to GNU Go :-). N10 seems a plausible move to defend.

Nicklas5:
1203: I think K18 is playable, in fact I tend to favor it over the
given alternatives. After B K18, W N18, B L19 (not N19 as in the game),
W O18, B J16 B has a good attack and can make life at J18 if necessary.

1211: I think the W grouop is alive in Semeai. I assume this should
result in the status alive once the semai module is perfected?
1212: I think the P13 status is dead. Note that W can afford to answer
a move like K9 with L9, or L9 with K10.




- regression tests reviewed

Index: arb.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/arb.tst,v
retrieving revision 1.14
diff -u -r1.14 arb.tst
--- arb.tst     2001/11/10 09:26:25     1.14
+++ arb.tst     2001/11/18 01:39:38
@@ -30,11 +30,6 @@
 150 gg_genmove white
 #? [J5]
 
-# Why is this move so urgent? /gf
-loadsgf games/arb/dumonteil-2.sgf 229
-200 gg_genmove black
-#? [F14]*
-
 loadsgf games/arb/dumonteil-2.sgf 247
 201 gg_genmove black
 #? [H13]
Index: nicklas2.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/nicklas2.tst,v
retrieving revision 1.17
diff -u -r1.17 nicklas2.tst
--- nicklas2.tst        2001/11/09 17:46:17     1.17
+++ nicklas2.tst        2001/11/18 01:39:39
@@ -73,7 +73,7 @@
 
 loadsgf games/nicklas/nicklas10.sgf 18
 1001 gg_genmove black
-#? [B7]*
+#? [H8]
 
 loadsgf games/nicklas/nicklas13.sgf 49
 1301 gg_genmove black
@@ -130,7 +130,7 @@
 #SEE_ALSO=trevor.tst?310
 loadsgf games/nicklas/nicklas18.sgf 45
 1803 gg_genmove black
-#? [A8]*
+#? [A6]
 
 
 #CATEGORY=ENDGAME_TUNING
Index: nicklas4.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/nicklas4.tst,v
retrieving revision 1.6
diff -u -r1.6 nicklas4.tst
--- nicklas4.tst        2001/11/09 17:46:17     1.6
+++ nicklas4.tst        2001/11/18 01:39:39
@@ -33,13 +33,19 @@
 
 loadsgf games/nicklas/nicklas11.sgf 172
 1103 gg_genmove black
-#? [C3]*
+#? [C2|C3]
 
-# P9 must be an error. /gf
+# It seems that GNU Go makes a logical mistake here: After seeing that
+# R9 can be defended with S7, and that Q8 can be defended with S7, I
+# doubt it checks whether both will be defended by S7. /ab
 loadsgf games/nicklas/nicklas12.sgf 29
 1201 gg_genmove black
-#? [P9]*
+#? [Q9]*
 
+# I have some doubt about the correct move. I might be advisable for B
+# to defend his center group, although W E16 hurts, of course. However,
+# W's attack after W N11, B N12 and next W L9 or similar might be
+# very useful as well.
 loadsgf games/nicklas/nicklas12.sgf 95
 1205 gg_genmove black
 #? [E16]*
Index: nicklas5.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/nicklas5.tst,v
retrieving revision 1.9
diff -u -r1.9 nicklas5.tst
--- nicklas5.tst        2001/11/09 17:46:17     1.9
+++ nicklas5.tst        2001/11/18 01:39:39
@@ -29,23 +29,28 @@
 1202 gg_genmove black
 #? [!L17]
 
+# I guess K18, giving up the O17 stones and attacking H17 in exchange,
+# is playable. /ab
 loadsgf games/nicklas/nicklas12.sgf 71
 1203 gg_genmove black
-#? [N18|N19|M19]*
+#? [K18|N18|N19|M19]
 
 loadsgf games/nicklas/nicklas12.sgf 89
 1204 gg_genmove black
 #? [H15]*
 
 # This is presumably a semeai problem.
+# W L8 is alive in semeai. /ab
 loadsgf games/nicklas/nicklas12.sgf 182
 1211 dragon_status L8
-#? [critical]*
+#? [alive]
 
 # This is presumably a semeai problem.
+# As W L8 is alive in semeai with sufficiently many liberties ahead of
+# B G13, I do not think the black P13 dragon can make eyes. /ab
 loadsgf games/nicklas/nicklas12.sgf 199
 1212 dragon_status P13
-#? [critical]*
+#? [dead]
 
 loadsgf games/nicklas/nicklas12.sgf 263
 1217 gg_genmove black