How long for you to detmerine that this position is dm35?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
Dann Corbit
Posts: 12034
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

How long for you to detmerine that this position is dm35?

Post by Dann Corbit » Fri Aug 14, 2020 3:27 am

Took my computer about 8 hours. I guess the right computer + program would be faster. I wonder how long it took Ellen, in 1879.
acd 71; bm g5; ce 32698; c0 "Mrs. J. W. (Ellen E {nee Strong}) Gilbert declares mate in 35 in correspondence game with George Gossip"; c3 "g5"; dm 35; pm g5; pv g5 hxg5 hxg5 Kc4 Kf4 e5 g4 e6 g3 Bd8 g2 Bxc7+ Kf5 Bh2 Kxe6 Kc3 Kf5 Kd3 Kg4 Bg1 Kf3 Bd4 Be4+ Kc4 Ke2 Kb4 Bc6 Kc3 Kf1 Kd2 g1=Q Bxg1 Kxg1 Kd3 Kf1 Kd2 Kf2 Kc3 Ke1 Kc4 Kd1 Kb4 Kc2 Ka3 Kb1 Kb4 Kxb2 Ka5 a3 Kb6 a2 Kxc6 Ka3 Kd5 b2 c6 b1=Q c7 Qf5+ Kc6 a1=Q c8=Q Qxc8+ Kd5 Qd7+ Ke4 Qad4+ Kf3 Q7g4+;
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

peter
Posts: 2199
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: How long for you to detmerine that this position is dm35?

Post by peter » Fri Aug 14, 2020 8:27 am

Dann Corbit wrote:
Fri Aug 14, 2020 3:27 am
Took my computer about 8 hours. I guess the right computer + program would be faster. I wonder how long it took Ellen, in 1879.
acd 71; bm g5; ce 32698; c0 "Mrs. J. W. (Ellen E {nee Strong}) Gilbert declares mate in 35 in correspondence game with George Gossip"; c3 "g5"; dm 35; pm g5; pv g5 hxg5 hxg5 Kc4 Kf4 e5 g4 e6 g3 Bd8 g2 Bxc7+ Kf5 Bh2 Kxe6 Kc3 Kf5 Kd3 Kg4 Bg1 Kf3 Bd4 Be4+ Kc4 Ke2 Kb4 Bc6 Kc3 Kf1 Kd2 g1=Q Bxg1 Kxg1 Kd3 Kf1 Kd2 Kf2 Kc3 Ke1 Kc4 Kd1 Kb4 Kc2 Ka3 Kb1 Kb4 Kxb2 Ka5 a3 Kb6 a2 Kxc6 Ka3 Kd5 b2 c6 b1=Q c7 Qf5+ Kc6 a1=Q c8=Q Qxc8+ Kd5 Qd7+ Ke4 Qad4+ Kf3 Q7g4+;
#33 or less:


Analysis by Stockfish 130820:
1...g5 2.hxg5 hxg5 3.Ld8 Kf4 4.e5 g4 5.Lxc7 g3 6.e6+ Kf3 7.e7 g2 8.Lh2 Kf2 9.Le5 Le8 10.Kc4 Kf1 11.Lh2 g1D 12.Lxg1 Kxg1 13.Kc3 Kf1 14.Kd2 Kf2 15.Kd3 Ke1 16.Kc3 Kd1 17.Kc4 Kd2 18.Kb4 Kc1 19.Ka3 Kc2 20.Kb4 Kxb2 21.c6 Lxc6 22.Kc5 Le8 23.Kb6 a3 24.Kc7 a2 25.Kd8 La4 26.e8D Lxe8 27.Kc7 Kc2 28.Kd6 b2 29.Ke7 b1D 30.Kf6 a1D+ 31.Kg5 De5+ 32.Kg4 Dg1+ 33.Kh3 Dh5#
-+ (-#33) Tiefe: 65/66 00:00:07 230MN
...switched to English abbr., which I had forgotten to before:
1...g5 2.hxg5 hxg5 3.Bd8 Kf4 4.e5 g4 5.Bxc7 g3 6.e6+ Kf3 7.e7 g2 8.Bh2 Kf2 9.Be5 Be8 10.Kc4 Kf1 11.Bh2 g1Q 12.Bxg1 Kxg1 13.Kc3 Kf1 14.Kd2 Kf2 15.Kd3 Ke1 16.Kc3 Kd1 17.Kc4 Kd2 18.Kb4 Kc1 19.Ka3 Kb1 20.Kb4 Kxb2 21.c6 Bxc6 22.Kc5 Be8 23.Kb6 a3 24.Kc7 a2 25.Kd8 Bb5 26.e8R Bxe8 27.Kc7 Kc2 28.Kd6 b2 29.Ke7 b1Q 30.Kf6 a1Q+ 31.Kg5 Qe5+ 32.Kg4 Qg1+ 33.Kh4 Qh5#
-+ (-#33) Depth: 70/66 00:06:00 14748MN

And now I'd wait for Louis' machine again. These "tests" are nowadays about hardware with SF (or CFish or other SF- branches) only, aren't they?
Peter.

Dann Corbit
Posts: 12034
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: How long for you to detmerine that this position is dm35?

Post by Dann Corbit » Fri Aug 14, 2020 6:36 pm

peter wrote:
Fri Aug 14, 2020 8:27 am
And now I'd wait for Louis' machine again. These "tests" are nowadays about hardware with SF (or CFish or other SF- branches) only, aren't they?
yes, I imagine that they are.
I used SF NNUE only.
i think Cfish is probably fastest, but i don't use CFish because it is the one SF variant that I cannot use to write logging data directly to disk.

For SF, Mike's variants, Sting, and any other C++ SF derivtive, I can log directly to disk, and I have parsers that will eat the logs and write directly to my database.

I still use other programs once in a while, but the rising strength of SF and cronies is making the others less and less relevant since they create so much more work for me in order to use their data.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

peter
Posts: 2199
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: How long for you to detmerine that this position is dm35?

Post by peter » Sat Aug 15, 2020 7:10 am

Dann Corbit wrote:
Fri Aug 14, 2020 6:36 pm
For SF, Mike's variants, Sting, and any other C++ SF derivtive, I can log directly to disk, and I have parsers that will eat the logs and write directly to my database.

I still use other programs once in a while, but the rising strength of SF and cronies is making the others less and less relevant since they create so much more work for me in order to use their data.
Xiphos has a very fine search for mating-scores too.

And as for SF forks not to forget Crystal and CorChess, it was the latter one, that brought along first #32- lines to the position.
To verify it backward to the root position I yet took Cfish again.

Here's output after 1...g5 2.hxg5, which might be the main line, the ones with 2.h5 and 2.e5 I already had shorter for some moves before.


Analysis by Cfish 030820 64 POPCNT NUMA:

2...hxg5 3.Kc3 Kf4 4.e5 g4 5.e6 g3 6.Bd8 Ke3 7.Bxc7 Kf2 8.Kd2 g2 9.Be5 g1Q 10.Bd4+ Kf1 11.Bxg1 Kxg1 12.Kd3 Kf2 13.Kd2 Bb5 14.e7 Be8 15.Kd3 Ke1 16.Kc3 Ke2 17.Kc4 Kd2 18.Kb4 Kc2 19.Ka3 Kb1 20.c6 Bxc6 21.Kb4 Be8 22.Kc5 a3 23.bxa3 Kc2 24.Kd5 b2 25.Ke5 Kd2 26.Kf4 b1Q 27.a4 Qf1+ 28.Ke5 Ke3 29.Kd5 Qf4 30.a5 Qd4+ 31.Ke6 Ke4 32.a6 Qe5#
-+ (-#31) Depth: 88/62 00:01:39 3586MN

Yet more than this one node is difficult to be kept in hash fully as for the last ply backward, at least the two other candidates come up again, aging the hash.

So I'd take the dtm from the output after 2.hxg5 and leave it for good, still not really sure, if there couldn't be one further move spared yet maybe.
Peter.

peter
Posts: 2199
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: How long for you to detmerine that this position is dm35?

Post by peter » Sat Aug 15, 2020 9:14 am

peter wrote:
Sat Aug 15, 2020 7:10 am
So I'd take the dtm from the output after 2.hxg5 and leave it for good,
Yet could'nt resist
:-)

Analysis by Cfish 030820 64 POPCNT NUMA:

1...g5 2.hxg5 hxg5 3.e5 Kf4 4.Bd8 g4 5.Bxc7 g3 6.Bd6 g2 7.e6+ Ke3 8.Bh2 Kf2 9.Be5 g1Q 10.Bd4+ Kg2 11.Bxg1 Kxg1 12.Kc3 Kf1 13.Kd2 Kf2 14.e7 Be8 15.Kd3 Ke1 16.Kc3 Ke2 17.Kc4 Kd2 18.Kb4 Kd3 19.c6 Bxc6 20.Ka3 Be8 21.Kb4 Kc2 22.Ka3 Kc1 23.Kb4 Kxb2 24.Kc4 Kc2 25.Kd4 a3 26.Ke5 b2 27.Kf6 b1Q 28.Ke5 a2 29.Ke4 Kd2+ 30.Kd5 a1B 31.Kd6 Qb6+ 32.Kd5 Qc6#
-+ (-#32) Depth: 89/64 00:06:48 14849MN
Peter.

Zenmastur
Posts: 919
Joined: Sat May 31, 2014 6:28 am

Re: How long for you to detmerine that this position is dm35?

Post by Zenmastur » Sat Aug 15, 2020 9:20 am

peter wrote:
Sat Aug 15, 2020 9:14 am
peter wrote:
Sat Aug 15, 2020 7:10 am
So I'd take the dtm from the output after 2.hxg5 and leave it for good,
Yet could'nt resist
:-)

Analysis by Cfish 030820 64 POPCNT NUMA:

1...g5 2.hxg5 hxg5 3.e5 Kf4 4.Bd8 g4 5.Bxc7 g3 6.Bd6 g2 7.e6+ Ke3 8.Bh2 Kf2 9.Be5 g1Q 10.Bd4+ Kg2 11.Bxg1 Kxg1 12.Kc3 Kf1 13.Kd2 Kf2 14.e7 Be8 15.Kd3 Ke1 16.Kc3 Ke2 17.Kc4 Kd2 18.Kb4 Kd3 19.c6 Bxc6 20.Ka3 Be8 21.Kb4 Kc2 22.Ka3 Kc1 23.Kb4 Kxb2 24.Kc4 Kc2 25.Kd4 a3 26.Ke5 b2 27.Kf6 b1Q 28.Ke5 a2 29.Ke4 Kd2+ 30.Kd5 a1B 31.Kd6 Qb6+ 32.Kd5 Qc6#
-+ (-#32) Depth: 89/64 00:06:48 14849MN
I get this after a few hours:

Code: Select all

( 95:M32 1...g5 2.hxg5 hxg5 3.Bd8 Kf4 4.e5 g4 5.e6 g3 6.Bxc7+ Kf3 7.Be5 g2 8.Bd4 Ke2 9.Kc4 Kf1 10.e7 g1=Q 11.Bxg1 Kxg1 12.Kb4 Be8 13.Kc3 Kf1 14.Kd2 Kf2 15.Kd3 Ke1 16.Kc3 Kd1 17.Kc4 Kd2 18.Kb4 Kc2 19.Ka3 Kc1 20.c6 Bxc6 21.Kb4 Be8 22.Ka3 Kc2 23.Kb4 Kxb2 24.Kc4 a3 25.Kb4 Kc2 26.Kc4 a2 27.Kd5 a1=Q 28.Ke6 Kd2 29.Kd5 Kd3 30.Kd6 Qc3 31.Ke6 Ke4 32.Kd6 Qc6# )
Only 2 defining forces have ever offered to die for you.....Jesus Christ and the American Soldier. One died for your soul, the other for your freedom.

Post Reply