Mate in 19

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

Re: Mate in 19.

Post by peter » Tue Aug 18, 2020 3:18 pm

Ajedrecista wrote:
Tue Aug 18, 2020 10:42 am

[d]r5kr/pp2q1pp/2pp3B/3n2Q1/3P4/P4R2/6PP/5RK1 w - -

It quickly returns a mate in 13 at shallow depths < 30, which means a mate in 20 from the root position. If I make the move Qg4, then it does not like ... b6 (the proposed black move for the mate in 17) because it shortens white mate. SF prefers ... a6, ... a5 or ... c5, for example. I would look at 12. ... b6, which looks fishy.

Analysis by Cfish 030820 64 POPCNT NUMA:

5.Bh6 d6 6.Qh5 Be6 7.Rae1 N8d7 8.Rxe6 Nf6 9.Qg5 Qc7 10.Rexf6 Nd5 11.R6f3 Qe7 12.Qg4 b6 13.h4 Rb8 14.h5 c5 15.Bg5 Qe2 16.h6 Re8 17.Be3 Qxf1+ 18.Rxf1 g6 19.Qd7 Re7 20.Qc8+ Re8 21.Qxe8#
+- (#17) Depth: 105/34 01:00:19 98077MN

After 12...a6 (instead of ...b6):


Analysis by Cfish 030820 64 POPCNT NUMA:

13.h4 a5 14.h5 a4 15.Bg5 Qe2 16.h6 Re8 17.Be3 Qxg2+ 18.Kxg2 Nxe3+ 19.Rxe3 Re7 20.Qc8+ Re8 21.Qxe8#
+- (#9) Depth: 88/18 00:00:19 568MN

After 12...a5:


Analysis by Cfish 030820 64 POPCNT NUMA:

13.h4 c5 14.h5 b5 15.Bxg7 Ne3 16.Rxe3 Qxe3+ 17.Kh1 Qe8 18.Bh6+ Qg6 19.hxg6 Re8 20.gxh7+ Kxh7 21.Qg7#
+- (#9) Depth: 67/18 00:00:04 99797kN

After 12...c5:


Analysis by Cfish 030820 64 POPCNT NUMA:

13.h4 c4 14.h5 c3 15.Bg5 Qe2 16.h6 Nf4 17.Bxf4 Kf7 18.Bg5+ Ke8 19.hxg7 Qf2+ 20.R1xf2 Rg8 21.Qe6#
+- (#9) Depth: 76/18 00:00:04 115MN
Peter.

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

Re: Mate in 19

Post by peter » Tue Aug 18, 2020 3:50 pm

acase wrote:
Tue Aug 18, 2020 9:59 am
Of course Peter, I wasn't absolutely sure that 19 was the quickest, but that was the best I could do.
Where is the position from?
It was discussed here before:

viewtopic.php?p=856600#p856600

But no hint to the origin there neither, even if I didn't read the whole thread.
Peter.

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

Re: Mate in 19.

Post by peter » Tue Aug 18, 2020 4:15 pm

peter wrote:
Tue Aug 18, 2020 3:18 pm

Analysis by Cfish 030820 64 POPCNT NUMA:

5.Bh6 d6 6.Qh5 Be6 7.Rae1 N8d7 8.Rxe6 Nf6 9.Qg5 Qc7 10.Rexf6 Nd5 11.R6f3 Qe7 12.Qg4 b6 13.h4 Rb8 14.h5 c5 15.Bg5 Qe2 16.h6 Re8 17.Be3 Qxf1+ 18.Rxf1 g6 19.Qd7 Re7 20.Qc8+ Re8 21.Qxe8#
+- (#17) Depth: 105/34 01:00:19 98077MN
Btw it's not Cfish- dependent, if you start the Backward already from the right lines:

rnbq2kr/pp1p2pp/1np5/8/3P1B2/P7/6PP/R2Q1RK1 w - - 0 1

Analysis by Stockfish 160820:

Code: Select all

5.Bh6 d6 6.Qh5 Be6 7.Rae1 N8d7 8.Rxe6 Nf6 9.Qg5 Qc7 10.Rexf6 Nd5 11.R6f3 Qe7 12.Qg4 b5 13.h4 b4 14.h5 b3 15.Bg5 Qe2 16.h6 Re8 17.Be3 Qxg2+ 18.Qxg2 g6 19.Rf7 Nf4 20.Rg7+ Kf8 21.Rxf4# 
+-  (#17)    Depth: 67/34   00:00:08  164MN
...
5.Bh6 d6 6.Qh5 Be6 7.Rae1 N8d7 8.Rxe6 Nf6 9.Qg5 Qc7 10.Rexf6 Nd5 11.R6f3 Qe7 12.Qg4 b5 13.h4 b4 14.h5 b3 15.Bg5 Qe2 16.h6 Re8 17.Be3 Qxg2+ 18.Qxg2 g6 19.Rf7 Nf4 20.Rg7+ Kf8 21.Rxf4# 
+-  (#17)    Depth: 75/34   00:00:55  1192MN
...
5.Bh6 d6 6.Qh5 Be6 7.Rae1 N8d7 8.Rxe6 Nf6 9.Qg5 Qc7 10.Rexf6 Nd5 11.R6f3 Qe7 12.Qg4 b5 13.h4 b4 14.h5 b3 15.Bg5 Qe2 16.h6 Re8 17.Be3 Qxg2+ 18.Qxg2 g6 19.Rf7 Nf4 20.Rg7+ Kf8 21.Rxf4# 
+-  (#17)    Depth: 80/34   00:02:46  3622MN
...
5.Bh6 d6 6.Qh5 Be6 7.Rae1 N8d7 8.Rxe6 Nf6 9.Qg5 Qc7 10.Rexf6 Nd5 11.R6f3 Qe7 12.Qg4 b5 13.h4 b4 14.h5 b3 15.Bg5 Qe2 16.h6 Re8 17.Be3 Qxg2+ 18.Qxg2 g6 19.Rf7 Nf4 20.Rg7+ Kf8 21.Rxf4# 
+-  (#17)    Depth: 92/34   00:15:12  20502MN
Peter.

User avatar
Ajedrecista
Posts: 1537
Joined: Wed Jul 13, 2011 7:04 pm
Location: Madrid, Spain.
Contact:

Re: Mate in 19.

Post by Ajedrecista » Tue Aug 18, 2020 4:57 pm

Hello Ignacio:
Ignacio wrote:
Tue Aug 18, 2020 12:23 pm
Ajedrecista wrote:
Tue Aug 18, 2020 10:42 am
[...]

Code: Select all

FEN: r5kr/pp2q1pp/2pp3B/3n2Q1/3P4/P4R2/6PP/5RK1 w - - 3 12 

ChestUCI_52:
FEN: r5kr/pp2q1pp/2pp3B/3n2Q1/3P4/P4R2/6PP/5RK1 w - -   (9+11)
Stellungs-Analyse:  C0/R0/K1/P7/X32   W:8/41
Suche nach Spezial-Matt [C0/R0/K1/P7/X32] in 10 ...  (Hash=512MB)
Suche abgeschlossen ...  (Zeit=525.30s)
Kein Matt in 10 gefunden !  (08:45)
[...]
[...]

Wrong parameters for Chest.

[...]
How to choose the right parameters? I start a Normal Mode analysis and Chest returns 'Stellungs-Analyse: Cc/Rr/Kk/Pp/Xx W:w/y', I stop the analysis, I set those parameters in the configuration window in Arena [Cc/Rr/Kk/Pp/Xx] while change the search mode to Special and I run the analysis again.

By the way, what are the meanings of these parameters? In the example of the position that I analyzed:

Code: Select all

FEN: r5kr/pp2q1pp/2pp3B/3n2Q1/3P4/P4R2/6PP/5RK1 w - -   (9+11)
Stellungs-Analyse:  C0/R0/K1/P7/X32   W:8/41

Arena configuration window:
C: ChecksOnly.
R: ThreatDepth.
K: KingSquares.
P: PieceLimit.
X: MaxMoves.

What I worked out in this position (and other ones too):
C0/R0/K1/P7/X32   W:8/41
C0 -> ?
R0 -> ?
K1 -> ? (Sometimes the number of squares where the side-to-move king can move in the first move of the position).
P7 -> ? [Sometimes the number of squares where the side-to-move pawns (or number of pieces maybe?) can move in the first move of the position].
X32 -> ?
W:8/41 -> 8 side-to-move pieces can move in the first move of the position (all except d4) and perft(1) = 41 in this position.
Regards from Spain.

Ajedrecista.

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

Re: Mate in 19.

Post by peter » Tue Aug 18, 2020 6:14 pm

Ajedrecista wrote:
Tue Aug 18, 2020 4:57 pm
By the way, what are the meanings of these parameters?
Binary comes along with a file called English, that's the English readme, to be able to open it in Arena, you have to rename it to English.txt, but you can open directly from the installation- directory with an editor, you know?

Retro- analysis running with SF's #17- line, back to 10.h5 it's said to be correct, but from there it will take it's time yet probably too.
If I get a result in reasonable time, I'll show it.
Peter.

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

Re: Mate in 19.

Post by peter » Tue Aug 18, 2020 7:01 pm

peter wrote:
Tue Aug 18, 2020 6:14 pm
Retro- analysis running with SF's #17- line, back to 10.h5 it's said to be correct, but from there it will take it's time yet probably too.
If I get a result in reasonable time, I'll show it.
Last output since about half an hour now:
FEN: r5kr/4q1pp/1ppp3B/p2n4/3P2QP/P4R2/6P1/5RK1 w - a6 (9+11)
Stellungs-Analyse: C0/R0/K1/P7/X35 W:8/42
RetroAnalysis-Suche nach Spezial-Matt [C0/R0/K2/P8/X36] in 8 ... (Hash=1024MB)
8.05 0:33 +M8 10.h5 860
8/08 0:33 +M8 10.h5 c5 11.Lxg7 Se3 12.Txe3 Dxe3+ 13.Kh1 De8 14.Lh6+ Dg6 15.hxg6 Te8 16.gxh7+ Kxh7 17.Dg7+ 868
Matt in 8 gefunden ! (1 Lösung in 00:33)
8/08 0:33 +M8 10.h5 c5 11.Lxg7 Se3 12.Txe3 Dxe3+ 13.Kh1 De8 14.Lh6+ Dg6 15.hxg6 Te8 16.gxh7+ Kxh7 17.Dg7+ 868
Bester Zug: h4-h5 Zeit: 0:33.977 min K/s: 868.083 Knoten: 29.397.646
FEN: r5kr/p3q1pp/1ppp3B/3n4/3P2Q1/P4R2/6PP/5RK1 w - - (9+11)
Stellungs-Analyse: C0/R0/K1/P7/X33 W:8/43
RetroAnalysis-Suche nach Spezial-Matt [C0/R0/K2/P8/X36] in 9 ... (Hash=1024MB)
If #9 with these parameters takes so long already, I don't think, root- position will be to be solved in that way.
Maybe Franz Huber could help. I'm even rather sure, he could, times before, when if it came to this, I was always astonished about the parameters he found.
Peter.

User avatar
Ajedrecista
Posts: 1537
Joined: Wed Jul 13, 2011 7:04 pm
Location: Madrid, Spain.
Contact:

Re: Mate in 19.

Post by Ajedrecista » Tue Aug 18, 2020 7:01 pm

Hello Peter:
peter wrote:
Tue Aug 18, 2020 6:14 pm
Binary comes along with a file called English, that's the English readme, to be able to open it in Arena, you have to rename it to English.txt, but you can open directly from the installation- directory with an editor, you know?
I have open it with Notepad many times in the past. I have no doubts that the info is well explained but I have not taken advantage of it. It might be beyond my understanding!

I tried Chest again with DefenderMoves enabled after Qg4: 72 minutes in my slow PC and again one ply short to reach the desired depth, just as my previous try of 30 minutes or so. Chest has reached the last ply at around 73 minutes (more than 600 MN), so very long time is expected until the final ply search will be completed. I have not got enough time so I stopped the analysis after 78 minutes.

I have been told that experience, trial and error are expected when looking for the right parameters, while K2/P9/X42 for the mate in 10 just after ... Qe7 could be fine although long time is expected until the solution is found. Thanks for the tips!
peter wrote:
Tue Aug 18, 2020 6:14 pm
Retro- analysis running with SF's #17- line, back to 10.h5 it's said to be correct, but from there it will take it's time yet probably too.
If I get a result in reasonable time, I'll show it.
I think I have never tested RetroAnalysis. Thank you in advance.

I agree with your other post: Ignacio and you are very good using Chest. I have been reading your 'Mate in X' posts for years. Though Franz is simply incredible, hats off. :)

Regards from Spain.

Ajedrecista.

Ignacio
Posts: 150
Joined: Wed Mar 08, 2006 7:15 pm

Re: Mate in 19

Post by Ignacio » Tue Aug 18, 2020 7:33 pm

Here is your mate in 10:

FEN: r5kr/pp2q1pp/2pp3B/3n2Q1/3P4/P4R2/6PP/5RK1 w - - 3 8

Chest 5.2:
FEN: r5kr/pp2q1pp/2pp3B/3n2Q1/3P4/P4R2/6PP/5RK1 w - - (9+11)
Position-Analysis: C0/R0/K1/P7/X32 W:8/41
Search for 1. Special-Mate [C0/R0/K2/P9/X42] in 126 ... (Hash=2044MB)
10 3:08:42 6.085.802.965 1.043.794 +M10 Dg5-g4
Search completed ... (Time=11321.96s)
Mate in 10 found ! (1. Solution in 3:08:41)
10/10 3:12:15 11.745.866.641 1.037.441 +M10 Dg5-g4 c6-c5 h2-h4 c5-c4 h4-h5 c4-c3 Ah6xg7 Cd5-e3 Tf3xe3 De7xe3+ Rg1-h1 De3-e8 Ag7-h6+ De8-g6 h5xg6 Ta8-e8 g6xh7+ Rg8xh7 Dg4-g7+

Gustav is better, 12 minutes.

Best.

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

Re: Mate in 19

Post by peter » Tue Aug 18, 2020 8:10 pm

Ignacio wrote:
Tue Aug 18, 2020 7:33 pm
Here is your mate in 10:
Well fine, but we are looking for the #17 at the root position, aren't we?
:)
Gustav has the fine feature, that you can define a certain field at which King has to be mated at, maybe it has to be h7 here, but even trying this it's pondering at root- position now for over an hour.

Here we'd need help from Olaf Jenkner, I'd say. For sure he'd know which parameter for Drohtiefe, maximale Zugzahl und Ausnahmen would be to be chosen here, without that, I don't think Gustav will make it overnight, but if I don't have any other work to be done by the PC it's installed on, I'll let it run.
Thanks for having given it a try with Chest too. At least you got one move further on than I did, breaking off with #9 after half an hour.
Peter.

acase
Posts: 734
Joined: Wed Mar 15, 2006 12:14 am
Location: Columbus, Ohio USA
Full name: Andrew R Case

Re: Mate in 19

Post by acase » Tue Aug 18, 2020 10:00 pm

peter wrote:
Tue Aug 18, 2020 3:50 pm
acase wrote:
Tue Aug 18, 2020 9:59 am
Of course Peter, I wasn't absolutely sure that 19 was the quickest, but that was the best I could do.
Where is the position from?
It was discussed here before:

viewtopic.php?p=856600#p856600

But no hint to the origin there neither, even if I didn't read the whole thread.
It comes from an old test position originally from John Dart That I modified from about 10 years ago, see here:

viewtopic.php?f=2&t=32898&p=333988&hili ... on#p333988

Post Reply