Update: I just realised that you can click the hint button on the bottom left to reveal the solution. So anyway enjoy the solution if you need it!
The most beautiful puzzle that my engine can never solve
Sort:
Weird. The chess.com engine saw mate coming in less than 1 second. And it didn't put a step wrong while on the road. I solved it only after a few retries on the first move.
Nevertheless, it is a beautiful puzzle. Never expected the multiple underpromotions of this version. I made one myself with a similar finish and 1 underpromotion:
4...Kc1
Why must Black play this?
I know the composition, so I just checked how fast an engine could solve it.
At my 6 year old computer Stockfish-devel gave +50 instantly, and #19 after less than half a minute. It could find a faster mate if I had the patience to leave it for a while.
We all know that chess dot com's online Stockfish is very weak, so this is not a composition which can illustrate the shortcomings of chess engines.
Stockfish 11, on my smart phone with 5 men tablebase access.
In less than 1 second, Stockfish think draw, 0.00
Just 1.22 second Stockfish knows it is tablebase winning.
About 20 seconds= SF announce Mate 20
At 60 seconds= SF announce Mate 17.
rapidflashlight wrote:
I don't see the position...
Are you on mobile or desktop? It definitely loads on a computer but there may be compatibility issues with a mobile device
drmrboss wrote:
Stockfish 11, on my smart phone with 5 men tablebase access.
In less than 1 second, Stockfish think draw, 0.00
Just 1.22 second Stockfish knows it is tablebase winning.
About 20 seconds= SF announce Mate 20
At 60 seconds= SF announce Mate 17.
I guess you all have very strong engines ^^. I just tested on chess.com's max analysis and I guess it runs without a tablebase, making the win beyond recognition from the starting position
Ellipsoul έγραψε:
drmrboss wrote:
Stockfish 11, on my smart phone with 5 men tablebase access.
In less than 1 second, Stockfish think draw, 0.00
Just 1.22 second Stockfish knows it is tablebase winning.
About 20 seconds= SF announce Mate 20
At 60 seconds= SF announce Mate 17.
I guess you all have very strong engines ^^. I just tested on chess.com's max analysis and I guess it runs without a tablebase, making the win beyond recognition from the starting position
You don't need a connection to tablebase here, the ply is not high. You just need an engine which is operating decently, and chess dot coms's online stockfish is not.
I thought that Platinum and Diamond members had the ability to analyze at high depth with multi core support? That's literally the only reason why I might consider a membership.
Ellipsoul wrote:
drmrboss wrote:
Stockfish 11, on my smart phone with 5 men tablebase access.
In less than 1 second, Stockfish think draw, 0.00
Just 1.22 second Stockfish knows it is tablebase winning.
About 20 seconds= SF announce Mate 20
At 60 seconds= SF announce Mate 17.
I guess you all have very strong engines ^^. I just tested on chess.com's max analysis and I guess it runs without a tablebase, making the win beyond recognition from the starting position
The authour of Droidfish put default cpu=1,
Today smartphone has 4 or 8 cores already. I put 8 cores for droidfish. 5 men Tablebases is around 900 MB and worth downloading. May be have a look on my topic about setting up droidfish properly ( including how to set up opening books etc)
https://www.chess.com/forum/view/general/how-to-set-up-your-droidfish-android-stockfish-properly
can you please name the composer of this beautiful study? took me 30 minutes to solve this.
capabalanca02 έγραψε:
can you please name the composer of this beautiful study? took me 30 minutes to solve this.
This was made by the famous Czech composer Mario Matouš
Aug 22, 2020
0
#17
Droidfish running on 8 threads on my phone finds the mate in 20 in just under 30 sec.
I made a puzzle and computer can solve it but you will see the end position.
Now white can nothing doing with his bishop and any engine says its better for white but white can't win and because of the 50-move rule its a draw (https://en.wikipedia.org/wiki/Fifty-move_rule) you can check it with your engine
Aron_08 έγραψε:
I made a puzzle and computer can solve it but you will see the end position.
Now white can nothing doing with his bishop and any engine says its better for white but white can't win and because of the 50-move rule its a draw (https://en.wikipedia.org/wiki/Fifty-move_rule) you can check it with your engine
"Any engine" does not mean much.
Here is an engine:
Flat equal everywhere.
Crystal is a Stockfish derivative, which (at least for correspondence chess) is more useful than the NN enhanced stuff. It is approx. 35% slower than Stockfish, but it has anti-fortress code inserted, which, as you see, works pretty well.
I guess if enough people see this post I will update this with the full solution in a later comment. Have fun! (p.s. I've never tried embedding a puzzle, I hope it worked)