Author: Dieter Buerssner
Date: 13:01:17 06/29/03
Go up one level in this thread
On June 29, 2003 at 15:30:03, Jeremiah Penery wrote:
>8/5p1p/8/6k1/8/6P1/5PP1/7K w - -
>
>This position was part of a test posted quite some time ago. All the other
>positions were solved without huge difficulty by programs, but this one was not.
> I've seen a couple of times where programs come up with the first move, but
>they don't really know what's going on.
>
>I want to see a program get the correct line with a decent score (more than +1
>or so).
>
>The line is 1. Kh2 f6 2. Kh1! f5 3. Kg1 Kf6 4. Kf1 Ke5 5. Ke1 Kd5 6. Kd1 Ke5 7.
>Kc2! Kd4 8. Kd2 Ke4 9. Ke2 h6 10. Kf1! Ke5 11. Kg1 Kf6 12. Kh2 Kg6 13. Kh3, and
>now the win is not hard to find.
Yace, my current developement version, shows a score > 1 after 40 seconds, and
the PV is identical to the one you showed for 16 ply (later moves are from HT,
and possibly unreliable). P4, 2.53 GHz:
[nodes time score depth]
16013535 17.472 0.85 28. 1.Kh2 f6 2.Kh1 f5 3.Kg1 Kf6 4.Kf1 Ke5 5.Ke1 Kd5
6.Kd1 Ke5 7.Kc2 Ke4 8.Kc3 Ke5 9.Kd3 h6 10.Ke2
Ke4 11.Kf1 Ke5 12.Kg1 Kd5 13.f3 Ke5 14.Kh2 Kd6
15.Kh3 Kd5 16.g4 f4 17.g5 Kc4 18.g3 Kd5 {80}
20473356 40.684 1.22 29t 1.Kh2 f6 2.Kh1 f5 3.Kg1 Kf6 4.Kf1 Ke5 5.Ke1 Kd5
6.Kd1 Ke5 7.Kc2 Kd4 8.Kd2 h6 9.Ke2H h5H 10.Kf3H
f4H 11.gxf4H Kc3H 12.g3H Kb2H 13.g4H h4H 14.f5H
h3H 15.f6H Kb3H 16.Kf4H h2H 17.f7H h1=QH
18.f8=QH Qh4H 19.f3H Kc4H {HT}
21894695 53.124 1.22 29. 1.Kh2 f6 2.Kh1 f5 3.Kg1 Kf6 4.Kf1 Ke5 5.Ke1 Kd5
6.Kd1 Ke5 7.Kc2 Kd4 8.Kd2 h6 9.Ke2 h5 10.Kf3 f4
11.gxf4 Kc3 12.g3 Kb2 13.g4 h4 14.f5 h3 15.f6
[...]
96216441 6:29.2 2.90 33t 1.Kh2 f6 2.Kh1 f5 3.Kg1 Kf6 4.Kf1 Ke5 5.Ke1 Kd5
6.Kd1 Ke5 7.Kc2 Kd4 8.Kd2 h6 9.Ke1 Kd5 10.Kf1
Kd4H 11.Kg1H Kd5H 12.Kh2H h5H 13.f3H Kc5H
14.g4H Kc6H 15.gxf5H h4H 16.f6H Kc5H 17.f7H
Kd4H 18.f8=QH Ke3H {HT}
104308919 7:56.5 2.90 33. 1.Kh2 f6 2.Kh1 f5 3.Kg1 Kf6 4.Kf1 Ke5 5.Ke1 Kd5
6.Kd1 Ke5 7.Kc2 Kd4 8.Kd2 h6 9.Ke1 Kd5 10.Kf1
Kd4 11.Kg1 Kd5 12.Kh2 h5 13.f3 Kc5 14.g4 Kc6
15.gxf5 h4 16.f6 Kc5 17.f7 Kd4 18.f8=Q Ke3
136699637 10:36.1 3.30 34++ 1.Kh2 f6 2.Kh1 f5 3.Kg1 Kf6 4.Kf1 Ke5 5.Ke1 Kd5
6.Kd1 Ke5 7.Kc2 Kd4 8.Kd2 h6 9.Ke2 Kd5 10.Kf3
Ke6 11.Kf4 Kd7 12.Kxf5 Kc8 13.f3 Kb8 14.f4 Ka8
15.g4 Ka7 16.g3 Ka6 17.Kf6 Kb6H 18.f5H Kc5H
19.g5H h5H 20.Kf7H Kd6H 21.g6H Kd5H 22.f6H Ke5H
23.g4H h4H 24.g5H h3H 25.g7H h2H 26.g8=QH h1=QH
{HT}
I stopped here.
Yace PB does not solve it fast. The main difference (I think) is that the new
experimental version can use KPPKP info from RAM (so access is very fast, and
can be used deep in the search).
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.