Mate in 96.

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

Moderator: Ras

Hobacle
Posts: 244
Joined: Thu Mar 04, 2021 3:57 pm
Full name: Nott Q Suree

Mate in 96.

Post by Hobacle »

[fen]7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1
[/fen]



Sincerely,
Hobacle
Spliffjiffer
Posts: 436
Joined: Thu Aug 02, 2012 7:48 pm
Location: Germany

Re: Mate in 96.

Post by Spliffjiffer »

im gonna take this seriously !....great content....but so far i do not want to make my brain work...i know that THIS is art (inspite of my content)...this is surely a better "thought"...thx VERY,VERY much....im going to try it but not now....were going to see us....hobacle...im gonna take some time...you are really a composer and i am definately not...im very curious what the community can discovers...thx!
Wahrheiten sind Illusionen von denen wir aber vergessen haben dass sie welche sind.
peter
Posts: 3487
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: Mate in 96.

Post by peter »

Hobacle wrote: Fri Sep 09, 2022 2:02 pm [fen]7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1
[/fen]
Didn't we have a similar one here called Spring Break then about one year ago?
:)
Peter.
Hobacle
Posts: 244
Joined: Thu Mar 04, 2021 3:57 pm
Full name: Nott Q Suree

Re: Mate in 96.

Post by Hobacle »

peter wrote: Fri Sep 09, 2022 3:07 pm
Hobacle wrote: Fri Sep 09, 2022 2:02 pm [fen]7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1
[/fen]
Didn't we have a similar one here called Spring Break then about one year ago?
:)
I pray for the real Spring Break.
Spliffjiffer wrote: Fri Sep 09, 2022 3:01 pm im gonna take this seriously !....great content....but so far i do not want to make my brain work...i know that THIS is art (inspite of my content)...this is surely a better "thought"...thx VERY,VERY much....im going to try it but not now....were going to see us....hobacle...im gonna take some time...you are really a composer and i am definately not...im very curious what the community can discovers...thx!
Not for art this time. Actually every single person is a composer, but some of composers compose viciousness and disputes, while some of the composers compose kindness, justice and love.
acase
Posts: 1002
Joined: Wed Mar 15, 2006 1:14 am
Location: Columbus, Ohio USA
Full name: Andrew R Case

Re: Mate in 96.

Post by acase »

Hobacle wrote: Fri Sep 09, 2022 2:02 pm [fen]7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1
[/fen]



Sincerely,
Hobacle
This reminds me of this puzzle, by Gustavus Charles Reichhelm

[fen]8/6pp/5p2/k7/3p4/1Q2p3/3prpp1/3Kbqrb w - - 0 0[/fen]

Here is a video of it:


I would imagine it works the same way. After Queening the White pawn the White Queen forces Black's King to h3 where the White Queen forces the checkmate from the h5 square
Last edited by acase on Fri Sep 09, 2022 5:59 pm, edited 1 time in total.
peter
Posts: 3487
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: Mate in 96.

Post by peter »

acase wrote: Fri Sep 09, 2022 5:41 pm
Hobacle wrote: Fri Sep 09, 2022 2:02 pm [fen]7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1
[/fen]



Sincerely,
Hobacle
This reminds me of this puzzle:

[fen]8/6pp/5p2/k7/3p4/1Q2p3/3prpp1/3Kbqrb w - - 0 0[/fen]

Here is a video of it:


I would imagine it works the same way. After Queening the White pawn the White Queen forces Black's King to h3 where the White Queen forces the checkmate from the h5 square
The thread started with earlier versions, followed by some others then still:

forum3/viewtopic.php?p=907413#p907413

Yet I'm again admiring it in the new form too, even if N on e8 is already captured and the pawns' positions and number hasn't changed so much.

Wonder if Peer's Chest- parameters for the earlier version would work still

forum3/viewtopic.php?p=907440#p907440

By myself I'm experimenting with Minic at the moment, which has a MateFinder- option in UCI- parameters that works surprisingly well, even in long distances like here.

Of course no good chance to keep the full DTM in hash, but at least the right track stays in output at backward of some suboptimal line forward following the already known scheme:

7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1

Analysis by Minic 3.30:

1.a4 Kg7 2.a5 Kf8 3.a6 Kg8 4.a7 Kg7 5.a8Q Kh6 6.Qf8+ Kg6 7.Qg8+ Kf5 8.Qd8 Kg6 9.Qe8 h6 10.Qh8 f5 11.Qf8 Kg5 12.Qxe7+ Kg6 13.Qf8 Kf6 14.Qxh6+ Ke7 15.Qe3 Kd6 16.Qa3+ Kc6 17.Qb4 Kd7 18.Qc3 f6 19.Qa1 Kd6 20.Qa7 Kc6 21.Qa4+ Kd6 22.Qb4+ Kd7 23.Qb7+ Ke8 24.Qb8+ Ke7 25.Qa7+ Kd8 26.Qf7 Kc8 27.Qe8+ Kc7 28.Qb5 Kd6 29.Qb7 e4 30.Qg7 Ke6 31.Qb7
Weiß steht klar auf Gewinn: +- (98.72) Tiefe: 52 00:02:49 4059MN

Not quite sure about the line as for capturing so many of the pawns at all and if it really works by force or with Black blunders only, just wanted to show the result or Minic's pondering after Backward.
BTW Peer's Chest- parameters seem not to work here as well as with old version,
regards
Last edited by peter on Fri Sep 09, 2022 6:19 pm, edited 2 times in total.
Peter.
Chessqueen
Posts: 5685
Joined: Wed Sep 05, 2018 2:16 am
Location: Moving
Full name: Jorge Picado

Re: Mate in 96.

Post by Chessqueen »

This one has fewer pieces, but it is more complicated to solve :roll:
[fen]6k1/5n2/8/8/8/5n2/1RK5/1N6 w - - 100 1 [/fen]
acase
Posts: 1002
Joined: Wed Mar 15, 2006 1:14 am
Location: Columbus, Ohio USA
Full name: Andrew R Case

Re: Mate in 96.

Post by acase »

Chessqueen wrote: Fri Sep 09, 2022 6:08 pm This one has fewer pieces, but it is more complicated to solve :roll:
[fen]6k1/5n2/8/8/8/5n2/1RK5/1N6 w - - 100 1 [/fen]
Not hard at all with Tablebases :lol:
and by the way you would have to ignore the 50 move rule for the win otherwise it is a draw.
User avatar
xr_a_y
Posts: 1872
Joined: Sat Nov 25, 2017 2:28 pm
Location: France

Re: Mate in 96.

Post by xr_a_y »

peter wrote: Fri Sep 09, 2022 5:52 pm
acase wrote: Fri Sep 09, 2022 5:41 pm
Hobacle wrote: Fri Sep 09, 2022 2:02 pm [fen]7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1
[/fen]



Sincerely,
Hobacle
This reminds me of this puzzle:

[fen]8/6pp/5p2/k7/3p4/1Q2p3/3prpp1/3Kbqrb w - - 0 0[/fen]

Here is a video of it:


I would imagine it works the same way. After Queening the White pawn the White Queen forces Black's King to h3 where the White Queen forces the checkmate from the h5 square
The thread started with earlier versions, followed by some others then still:

forum3/viewtopic.php?p=907413#p907413

Yet I'm again admiring it in the new form too, even if N on e8 is already captured and the pawns' positions and number hasn't changed so much.

Wonder if Peer's Chest- parameters for the earlier version would work still

forum3/viewtopic.php?p=907440#p907440

By myself I'm experimenting with Minic at the moment, which has a MateFinder- option in UCI- parameters that works surprisingly well, even in long distances like here.

Of course no good chance to keep the full DTM in hash, but at least the right track stays in output at backward of some suboptimal line forward following the already known scheme:

7k/4pp1p/5p2/4p3/8/5pp1/P3prrb/4Kbqn w - - 0 1

Analysis by Minic 3.30:

1.a4 Kg7 2.a5 Kf8 3.a6 Kg8 4.a7 Kg7 5.a8Q Kh6 6.Qf8+ Kg6 7.Qg8+ Kf5 8.Qd8 Kg6 9.Qe8 h6 10.Qh8 f5 11.Qf8 Kg5 12.Qxe7+ Kg6 13.Qf8 Kf6 14.Qxh6+ Ke7 15.Qe3 Kd6 16.Qa3+ Kc6 17.Qb4 Kd7 18.Qc3 f6 19.Qa1 Kd6 20.Qa7 Kc6 21.Qa4+ Kd6 22.Qb4+ Kd7 23.Qb7+ Ke8 24.Qb8+ Ke7 25.Qa7+ Kd8 26.Qf7 Kc8 27.Qe8+ Kc7 28.Qb5 Kd6 29.Qb7 e4 30.Qg7 Ke6 31.Qb7
Weiß steht klar auf Gewinn: +- (98.72) Tiefe: 52 00:02:49 4059MN

Not quite sure about the line as for capturing so many of the pawns at all and if it really works by force or with Black blunders only, just wanted to show the result or Minic's pondering after Backward.
BTW Peer's Chest- parameters seem not to work here as well as with old version,
regards
I'm happy you find Minic's mateFinder option usefull. :D

Did you get this 31 moves deep line with mateFinder on ? This looks impossible to me as this option will disable almost all forward pruning so that it is very hard to go deep in the line. Or maybe you are using a very big TT and many threads ?
peter
Posts: 3487
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: Mate in 96.

Post by peter »

xr_a_y wrote: Fri Sep 09, 2022 9:14 pm I'm happy you find Minic's mateFinder option usefull. :D

Did you get this 31 moves deep line with mateFinder on ? This looks impossible to me as this option will disable almost all forward pruning so that it is very hard to go deep in the line. Or maybe you are using a very big TT and many threads ?
Hi Vivien!
Yes, MateFinder was checked, 32G hash, 30 threads of 16x30GHz were used, and output was copied after quite some forward of an interactive line and backward to the starting position with already full hash.
MateFinder option works fine for positions like this one and Minic seems to have gained much tactical strength at all as for the test- positions I tried so far, regards
Peter.