Author: Dieter Buerssner
Date: 15:55:04 02/16/01
Go up one level in this thread
On February 16, 2001 at 15:35:40, Lin Harper wrote:
> 8/8/8/2p5/R1P5/P5p1/P3K1p1/6k1 w
>
> [...]But can your comp find the mate
> sequence?
As Uri allready mentioned, this is clearly a position, where programs, that do
extensions, when there is only one legal move possible, will have a big
advantage. Yace does not do this normally, and needed almost 10 minutes to find
a mate in 17 for Ra8 and soon a mate in 16 for Ra7. Then I included the 1 legal
move extension (which I have left out normally, but I am not certain which is
better), and it showed in under 5 seconds mate in 16 on my AMD K6-2 475 with 20
M hash:
754424 4.566 13.41 13. 1.Ra7 Kh2 2.Rh7+ Kg1 3.Kd2 Kf1 4.Rf7+ Kg1 5.a4
Kh2 6.Rh7+ Kg1 7.a5 Kf1 8.Rf7+ Kg1 9.a6 Kh2
10.Rh7+ Kg1 11.a7 Kf1 12.Rf7+ Kg1 13.a8=Q Kh2
{1421}
754955 4.594 Mat16 14++ 1.Ra7 Kh2 2.Rh7+ Kg1 3.Kd2 Kf1 4.Rf7+ Kg1 5.a4
Kh2 6.Rh7+ Kg1 7.a5 Kf1 8.Rf7+ Kg1 9.a6 Kh2
10.Rh7+ Kg1 11.a7 Kf1 12.Rf7+ Kg1 13.a8=Q Kh2
14.Qh8+ Kg1 15.Qa1+ Kh2 16.Rh7# {1421}
1127651 6.260 Mat16 14t 1.Ra7 Kh2 2.Rh7+ Kg1 3.Kd2 Kf1 4.Rf7+ Kg1 5.a4
Kh2 6.Rh7+ Kg1 7.a5 Kf1 8.Rf7+ Kg1 9.a6 Kh2
10.Rh7+ Kg1 11.a7 Kf1 12.Rf7+ Kg1 13.a8=Q Kh2
14.Qh8+ Kg1 15.Qa1+ Kh2 16.Rh7# {1421}
1345118 7.187 Mat16 14. 1.Ra7 Kh2 2.Rh7+ Kg1 3.Kd2 Kf1 4.Rf7+ Kg1 5.a4
Kh2 6.Rh7+ Kg1 7.a5 Kf1 8.Rf7+ Kg1 9.a6 Kh2
10.Rh7+ Kg1 11.a7 Kf1 12.Rf7+ Kg1 13.a8=Q Kh2
14.Qh8+ Kg1 15.Qa1+ Kh2 16.Rh7# {1421}
BTW. Especially in this positions, some programs with carefully tuned search may
not find the mate too fast. More than one move are clearly winning, and every
program will see this soon. An author might decide, to not do aggressive
extensions, when a position found in the search allready looks won. This will
waste some time, when this is a "refutation-line", and will show a better score
when this is a PV. But in general, I think, it will not the playing strength of
the program, to aggressively extend won lines.
This also may be a good position for testing the hash implementation. I get
extremely good hash efficiency here, and other than the famous Fine 70, this
position may also help to test the correct handling of mate scores in with the
hash tables.
Regards,
Dieter
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.