[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] New testcase
From: |
Gunnar Farneback |
Subject: |
Re: [gnugo-devel] New testcase |
Date: |
Thu, 07 Nov 2002 20:12:09 +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) |
Nando wrote:
> Index: regression/blunder.tst
> [...]
> +
> +# Apparently, T9 gets overvalued because the effect on S7 and T6 is
> +# counted twice (territorially and strategically)
> +loadsgf games/blunder23.sgf
> +30 gg_genmove white
> +#? [!T9]
Good test case, but I'd like to reserve blunder.tst for true blunders,
not merely bad moves. The patch below clarifies this.
/Gunnar
Index: regression/blunder.tst
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/blunder.tst,v
retrieving revision 1.24
diff -u -r1.24 blunder.tst
--- regression/blunder.tst 27 Oct 2002 16:46:14 -0000 1.24
+++ regression/blunder.tst 7 Nov 2002 18:57:18 -0000
@@ -4,6 +4,11 @@
reset_reading_node_counter
reset_trymove_counter
+# This test suite is a collection of blunders. By "blunder" we here
+# mean a move which is worse than passing. Also included are moves
+# which are okay but which GNU Go incorrectly rejects thinking that
+# they would be blunders.
+
loadsgf games/blunder1.sgf
1 gg_genmove black
#? [!E5]