Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Endgame Suite

Author: Frank Schneider

Date: 13:44:49 01/20/98

Go up one level in this thread


Testresults for Gromit 1.2 (K6/200, most 4-man-tablebases):
#  seconds
1: 0
2: 2
3: ?
4: 3
5: 57
6: 2
7: ?
8: 0
9: ?
10: 1
11,12,13,14: ?
15: 70
16: 0
17: 0
18: ?
19: ?
20: 0
21: 1612 ( ;-)
22: ?
The testing was done in a lazy way: I didn't wait too long. Therefore
some
correct moves may be unstable and some ? could have been found.
From this and the DarkThought-results it seems that there are many
positions
that are solved quickly and are easy for programs.
I suspect, Gromit plays some good moves for the wrong reasons - but that
is
a problem with most tests.

(The extreme-case is a test that is so difficult that no program finds a
single
move. A random movegenerator (or a very weak program) would be much
better that
the good programs)

Frank



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.