A bit problematic position for engines

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
EroSennin
Posts: 130
Joined: Fri Apr 09, 2010 1:26 am

A bit problematic position for engines

Post by EroSennin » Sun Jul 17, 2016 11:55 am

Image

7r/ppp2kpp/8/2b5/8/3KnP2/P5PP/RN4NR b - -

Haha had to edit my position as this is a much cooler starting position.

styx
Posts: 338
Joined: Tue Mar 13, 2012 8:59 pm
Location: Germany

Re: A bit problematic position for engines

Post by styx » Sun Jul 17, 2016 12:25 pm

what is the solution? SF does not find anything better than Rd8+ (0.00) up to depth 53

EroSennin
Posts: 130
Joined: Fri Apr 09, 2010 1:26 am

Re: A bit problematic position for engines

Post by EroSennin » Sun Jul 17, 2016 12:37 pm

I found this in Smerdon's Scandinavian. Black might be winning the endgame. I guess I don't wanna ruin the fun yet with a few moves.

MikeGL
Posts: 943
Joined: Thu Sep 01, 2011 12:49 pm

Re: A bit problematic position for engines

Post by MikeGL » Sun Jul 17, 2016 1:38 pm

Maybe the first rank control + Q side pawns is too much for a minor piece.

[pgn]
[Event "Computer chess game"]
[Site "Talkchess"]
[Date "2016.07.17"]
[Round "?"]
[White "ENgine1"]
[Black "Engine2"]
[Result "1-0"]
[Time "21:22:32"]
[WhiteElo "2400"]
[TimeControl "0+1"]
[SetUp "1"]
[FEN "7r/ppp2kpp/8/2b5/8/3KnP2/P5PP/RN4NR b - - 0 1"]
[Termination "unterminated"]
[PlyCount "22"]
[WhiteType "human"]
[BlackType "human"]

1. ... Rd8+ 2. Ke2 Rd1 3. a4 Rc1 4. Kd2 Rd1+ 5. Ke2 a6 6. Ra3 Rxb1 7. Rxe3
Bxe3 8. Kxe3 b5 9. Kd3 c5 10. axb5 axb5 11. g4 c4+ 12. Kd4 +-


[/pgn]

syzygy
Posts: 4602
Joined: Tue Feb 28, 2012 10:56 pm

Re: A bit problematic position for engines

Post by syzygy » Sun Jul 17, 2016 1:47 pm

EroSennin wrote:I found this in Smerdon's Scandinavian. Black might be winning the endgame. I guess I don't wanna ruin the fun yet with a few moves.
Chances are that the "solution" does not really work now that SF clearly cannot find it. If you post it, people can verify that it actually works before spending a lot of time letting their engines search for it.

Vinvin
Posts: 4691
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: A bit problematic position for engines

Post by Vinvin » Sun Jul 17, 2016 3:23 pm

MikeGL wrote:Maybe the first rank control + Q side pawns is too much for a minor piece.

[Event "Computer chess game"]
[Site "Talkchess"]
[Date "2016.07.17"]
[Round "?"]
[White "ENgine1"]
[Black "Engine2"]
[Result "1-0"]
[Time "21:22:32"]
[WhiteElo "2400"]
[TimeControl "0+1"]
[SetUp "1"]
[FEN "7r/ppp2kpp/8/2b5/8/3KnP2/P5PP/RN4NR b - - 0 1"]
[Termination "unterminated"]
[PlyCount "22"]
[WhiteType "human"]
[BlackType "human"]

1. ... Rd8+ 2. Ke2 Rd1 3. a4 Rc1 4. Kd2 Rd1+ 5. Ke2 a6 6. Ra3 Rxb1 7. Rxe3
Bxe3 8. Kxe3 b5 9. Kd3 c5 10. axb5 axb5 11. g4 c4+ 12. Kd4 +-
It's a case of long term pinning.
Like this position :
[d]5rk1/pp3p1p/7b/2pR4/8/2P4P/P1PNr1P1/2K4R b - - bm Rxd2;

Engines should really be modified in case of long term pinning : reduce the value of the piece and extend the search more (to find an escape of the pin) because the value of a pinned piece is way less than it's usual value. In the original position the knight worth about 0. In the second position the evaluation is even more difficult because black can exchange we he wants.


Another case where a rook worth about 0 point for a long time because it's a completely ineffective piece after Rxe5! fxe5 Rf1! :
[d]4q1kr/p6p/1prQPppB/4n3/4P3/2P5/PP2B2P/R5K1 w - - bm Qxe5;

EroSennin
Posts: 130
Joined: Fri Apr 09, 2010 1:26 am

Re: A bit problematic position for engines

Post by EroSennin » Sun Jul 17, 2016 7:37 pm

Well I haven't analysed it completely just checked a few things.
1... Rd8 2. Ke2 Rd1 3. a4 Bd4 4.Ra3 Rxb1 5.Rxe3 Bxe3 6.Kxe3 a5 7.Kd3 c6 followed b5 and maybe Ke6. Hard to say if it is really completely winning.
3.f4 b5 4.g4 b4 5.h4 a5 again seems quite good for black

Vinvin
Posts: 4691
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: A bit problematic position for engines

Post by Vinvin » Mon Jul 18, 2016 11:46 pm

EroSennin wrote:Well I haven't analysed it completely just checked a few things.
1... Rd8 2. Ke2 Rd1 3. a4 Bd4 4.Ra3 Rxb1 5.Rxe3 Bxe3 6.Kxe3 a5 7.Kd3 c6 followed b5 and maybe Ke6.
Here I got a nice + for black :

Code: Select all

FEN: 8/1p3kpp/2p5/p7/P7/3K1P2/6PP/1r4NR w - - 0 8

Stockfish_160716_x64_modern_fast:

 35/54	00:29	 405 164k	13 689k	-0,91	8.h3 b5 9.g4 Ke6 10.h4 bxa4 11.Rh2 Rxg1 12.Ra2 Rd1+ 13.Ke3 g6 14.Rxa4 Rd5 15.Rc4 Kd6 16.f4 Rb5 17.Rd4+ Ke6 18.Re4+ Kf7 19.Rd4 Rb3+ 20.Kf2 Rb7 21.Ra4 Rb5 22.Ke3 Kf6 23.Rd4 Ke7 24.Re4+ Kd7 25.Rd4+ Rd5 26.Ra4 Ke6 27.Ra1 Rb5 28.Rc1 c5 29.Ra1 Kd5
 36/54	00:29	 405 164k	13 689k	-0,75	8.h4 Ke6 9.g4 b5 10.h5 b4 11.h6 g6 12.f4 Ra1 13.Kd2 c5 14.f5+ gxf5 15.gxf5+ Kf6 16.Kc2 c4 17.Rh4 Rxg1 18.Rxc4 Rf1 19.Rc6+ Kg5 20.Rc7 Rxf5 21.Rxh7 Kg6 22.Rh8 Rh5 23.Rb8 Kxh6 24.Rb6+ Kg5 25.Rb5+ Kg6 26.Rb6+ Kf7 27.Rb7+ Ke6 28.Rb6+ Kd7 29.Ra6 Rf5 30.Kb2 Ke7 31.Kb3 Rh5 32.Kc2 Rd5 33.Kb3 Kd7 34.Kc4 Rg5
 36/54	00:29	 405 164k	13 689k	-0,75	8.g4 b5 9.h4 Ke6 10.h5 b4 11.h6 g6 12.f4 Ra1 13.Kd2 c5 14.f5+ gxf5 15.gxf5+ Kf6 16.Kc2 c4 17.Rh4 Rxg1 18.Rxc4 Rf1 19.Rc6+ Kg5 20.Rc7 Rxf5 21.Rxh7 Kg6 22.Rh8 Rh5 23.Rb8 Kxh6 24.Rb6+ Kg5 25.Rb5+ Kg6 26.Rb6+ Kf7 27.Rb7+ Ke6 28.Rb6+ Kd7 29.Ra6 Rf5 30.Kb2 Ke7 31.Kb3 Rh5 32.Kc2 Rd5 33.Kb3 Kd7 34.Kc4

peter
Posts: 2012
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: A bit problematic position for engines

Post by peter » Tue Jul 19, 2016 5:31 am

Hi Jean!
EroSennin wrote:1... Rd8 2. Ke2 Rd1 3. a4 Bd4 4.Ra3 Rxb1 5.Rxe3 Bxe3 6.Kxe3 a5 7.Kd3 c6 followed b5 and maybe Ke6.
Good position indeed, thanks for it.
After your line above Black's winning, but what about 7.h4 instead of Kd3?

E.g:

7r/ppp2kpp/8/2b5/8/3KnP2/P5PP/RN4NR b - - 0 1


1... Rd8+ 2. Ke2 Rd1 3. a4 Bd4 4. Ra3 Rxb1 5. Rxe3 Bxe3 6. Kxe3 a5

7. h4 b5 8. h5 b4 9. Rh4 b3 10. Ne2 b2 11. Nc3 Rc1 12. Kd2 Rxc3 13. Rh1 Rc5 14. Rb1 Rxh5 15. Rxb2 Ke6
and I don't think the 1 plus-pawn is enough in this Rook- ending
Peter.

Post Reply