Interesting study

Technical questions regarding Openings, Middlegames, Endings etc.
Wadih Khoury
Posts: 604
Joined: Sun Jul 12, 2020 8:14 pm

Interesting study

Post by Wadih Khoury » Sat Nov 07, 2020 7:04 pm

My son was so excited that he solved this study that I thought I'd share it here.
As I understand, even engines can't crack it (I tried on Stockfish and Leela and they failed), and it would appear that some 80's super GMs couldn't solve until Tal found the answer.


White to play.

Matthew Turner
Posts: 3600
Joined: Fri May 16, 2008 11:54 am

Re: Interesting study

Post by Matthew Turner » Sat Nov 07, 2020 7:59 pm

I see the idea and I know where the Kings are at the end. I think there is a general clue to the answer if you consider why each White piece has made it onto the board :D

Roger Lancaster
Posts: 1910
Joined: Tue Mar 17, 2015 2:44 pm

Re: Interesting study

Post by Roger Lancaster » Sun Nov 08, 2020 11:47 am

I could half-remember this so I was sure I'd seen it elsewhere. The solution is at https://www.chess.com/article/view/the- ... ess-puzzle [and doubtless other places too] with the main line running to 11 moves so your son did well to solve it.

Wadih Khoury
Posts: 604
Joined: Sun Jul 12, 2020 8:14 pm

Re: Interesting study

Post by Wadih Khoury » Sun Nov 08, 2020 12:25 pm

What I found fascinating was that the computer engines still have blind spots.

MartinCarpenter
Posts: 3044
Joined: Tue May 24, 2011 10:58 am

Re: Interesting study

Post by MartinCarpenter » Mon Nov 09, 2020 9:49 am

In SF's case its a purposeful choice to limit their search space to make them stronger in real games. There's probably some way to force it to do a more exhaustive search than usual which would let it solve the puzzle.
(Really old engines definitely had this, no idea if it has survived.).

Leela is far too human to have a chance :)

User avatar
Paul Robert Jackson
Posts: 266
Joined: Sun Jun 14, 2020 8:10 pm

Re: Interesting study

Post by Paul Robert Jackson » Sat Nov 21, 2020 8:56 am

Dutch endgame composer Gijs van Breukelen.
There is a ChessBase article.
https://en.chessbase.com/post/solution- ... able-study
Paul Robert Jackson