gnugo-devel
[Top][All Lists]
Advanced

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

[gnugo-devel] Owl mistake and thrashing dragons.


From: Gunnar Farneback
Subject: [gnugo-devel] Owl mistake and thrashing dragons.
Date: Thu, 16 Jan 2003 22:58:04 +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)

GNU Go made a couple of rather serious mistakes in the NNGS game
Temnik-gnugo-3.3.15-200301161937.sgf (available in CVS and on the
recent games page).

The first one I want to discuss is at move 186.

   A B C D E F G H J K L M N O P Q R S T
19 . . . X X O . O . . . . . . . . . . . 19
18 . . X . X O . . O O X X . O X X . . . 18
17 . . X . X X O O O X O O O O O X . . . 17
16 . . X X O X X O . X X X O X X X . . . 16
15 . X X O O O O O X X X O X X . . . . . 15
14 . X O O . X O X X O O O . . . X X . . 14
13 X O . . . . O X . X . X X . . . . X X 13
12 . O . . O O O O X X X . . X . O O O X 12
11 . O O O X X X X O O O X X X O . O X O 11     WHITE has captured 3 stones
10 . X O O O X O O . + . . . . . O . X . 10     BLACK has captured 6 stones
 9 X . X O O X X O O . . . . . . . O . X 9
 8 . . X X O X . X O . . . . . . X . O O 8
 7 . . X O O X X X O . . . . . . . O O . 7
 6 . . . X X O . . O . . . . . X . O X O 6
 5 . . . X . O . . . . X . . . . . X X X 5
 4 . . . X O . . O . + . . . . X X X O . 4
 3 . X X O . O . X O . X . . X . O O O . 3
 2 . X O O . . O O . . . . X . O . . . . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

White is struggling to live on the right edge. Now GNU Go finds the S8
dragon critical with both attack and defense points at T7! This is
just horrible. T7 has no effect whatsoever on the life of the group.
In fact it shouldn't even spend any time at all trying it.

In the following moves white is allowed to connect out, but let's skip
possible mistakes there. Fast forward to move 220 where white has
allowed the group to become cut off once more while attempting to pull
out the M15 stones.

   A B C D E F G H J K L M N O P Q R S T
19 . . . X X O . O . . . . . O X . . . . 19
18 . . X . X O . . O O X X X O X X . . . 18
17 . . X . X X O O O X O O O O O X . . . 17
16 . . X X O X X O . X X X O X X X . . . 16
15 . X X O O O O O X X X O X X . . . . . 15
14 . X O O . X O X X O O O O O . X X X . 14
13 X O . . . . O X . X O X X . O . O X X 13
12 . O . . O O O O X X X . . X . O O O X 12
11 . O O O X X X X O O O X X X O . O X X 11     WHITE has captured 3 stones
10 . X O O O X O O . + O O X O O O O X X 10     BLACK has captured 10 stones
 9 X . X O O X X O O . O X . X X . O O X 9
 8 . . X X O X . X O . X X . X . X . O O 8
 7 . . X O O X X X O . O X X . O . O O X 7
 6 . . . X X O . . O . X . . . X . O X . 6
 5 . . . X . O . . . . X . . . . . X X X 5
 4 . . . X O . . O . + . . . . X X X O . 4
 3 . X X O . O . X O . X . . X . O O O . 3
 2 . X O O . . O O . X . . X . O . . . . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

White is dead, even way dead, but starts thrashing with P7. GNU Go has
a very clear lead and can easily afford a local reply. As it happens
we have a policy to attack thrashing dragons, but the implementation
is not good enough, as this example will show.

220 tenuki (bad)
222 tenuki (bad)
224 owl attack (correct)
226 owl attack (correct)
228 attacks thrashing dragon (ineffective)
230 attacks thrashing dragon (ineffective)
232 tenuki (extremely bad)

At move 232 GNU Go misreads a connection and is subsequently cut with
a lost semeai as a result (yes, white's group got away and took P4
with it). But even disregarding the connection mistake, which is far
from trivial, black should have reinforced the position as a thrashing
dragon counter-measure.

Suggestions how to improve these five moves?

Below is a new test case for the connection mistake at move 232. Feel
free too add test cases for the other problems, if you feel confident
about the correct answers.

- new test case

/Gunnar

Index: regression/connection.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/connection.tst,v
retrieving revision 1.50
diff -u -r1.50 connection.tst
--- regression/connection.tst   11 Jan 2003 16:06:03 -0000      1.50
+++ regression/connection.tst   16 Jan 2003 21:54:27 -0000
@@ -333,6 +333,10 @@
 100 disconnect L12 M13
 #? [(2|3) J15]*
 
+loadsgf games/nngs/Temnik-gnugo-3.3.15-200301161937.sgf 232
+101 disconnect N2 L5
+#? [1 M3]
+
 # Report number of nodes visited by the tactical reading
 10000 get_reading_node_counter
 #? [0]&




reply via email to

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