Endgame study, Costeff 1996

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

Moderators: hgm, Rebel, chrisw

afzzq

Endgame study, Costeff 1996

Post by afzzq »

Can any engine solve the following study, white to play and win? It is from Gady Costeff, 1996:

[d]5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

White's first move 1. Rb1 is easy to find. But after 1...a5, White's only move is 2 Rd1! Can engines solve that?

1.Rb1 a5 2.Rd1 ( 2.Ra1 a6 3.Kf7 ( 3.Rd1 d5 4.e6 d4 5.e7 d3 6.e8=Q d2 7.Rxd2 Nf3 8.Rf2 g1=Q 9.Qe2 Qg3 10.Qd1+ Ng1 11.Qd5+ Nf3 ( 11...Qf3+ 12.Rxf3 Nxf3 13.Qg8 Ng5 14.Ke7 ) 12.Rxf3 Qxf3+ 13.Qxf3+ Kg1 ) 3...d5 4.e6 d4 5.e7 d3 6.e8=B d2 ) 2...a6 3.Ra1 d5 4.e6 d4 5.e7 d3 6.e8=B d2 7.Bh5 d1=Q 8.Bxd1 Nh3 9.Bc2+ Ng1 10.Bb1 Nf3 11.Be4+ Ng1 12.Ra2 1-0

Stockfish didn't get far.
User avatar
AdminX
Posts: 6362
Joined: Mon Mar 13, 2006 2:34 pm
Location: Acworth, GA

Re: Endgame study, Costeff 1996

Post by AdminX »

Yace spots the 2nd move Rd1 in no time at all.

[d]5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Yace Paderborn experimental - Book by Daniel Mehrmann:

1.Rb1 a5 2.Rd1 a6 3.Kf7 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rxg2 Kxg2 9.Qe2+ Kg1 10.Qd1+ Kg2 11.Qc2+ Kg1 12.Qc1+ Kg2 13.Qb2+ Kg3 14.Qc1
-+ (-5.72 ++) Depth: 6/11 00:00:00 10kN
+- (3.57) Depth: 19/43 00:01:32 219mN

So does Crafty:

[d]5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Crafty 23.00:

1.Kg7
-+ (-8.03) Depth: 1/1 00:00:00 0kN

1.Rb1 a5 2.Rd1 a6 3.Kf7 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rf2 g1Q 9.Rxf3 Qa7+ 10.Kg6 Qd4 11.Kf5 Qd1 12.Qe4 Qd7+ 13.Kf6 Qd6+ 14.Kg5 Qd2+
= (0.00) Depth: 8/11 00:00:00 4kN
+- (2.49) Depth: 24/42 00:00:15 70538kN
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
User avatar
AdminX
Posts: 6362
Joined: Mon Mar 13, 2006 2:34 pm
Location: Acworth, GA

Re: Endgame study, Costeff 1996

Post by AdminX »

Here is Zappa Mexico II

[d]5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Zappa Mexico II:

1.Rb1 a5 2.Rd1 a6 3.Kf7 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rd3 g1Q 9.Qc6 Qg2 10.Rxf3 Qxf3+ 11.Qxf3+ Kg1 12.Qe3+ Kg2 13.Qg5+ Kf2 14.Qd2+ Kg1 15.Qd1+ Kg2 16.Qd2+ Kg1
= (0.00) Depth: 15/36 00:00:01 588kN
+- (5.04) Depth: 19/57 00:00:32 92195kN, tb=2403
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
afzzq

Re: Endgame study, Costeff 1996

Post by afzzq »

AdminX wrote:Yace spots the 2nd move Rd1 in no time at all.

Analysis by Yace Paderborn experimental - Book by Daniel Mehrmann:

1.Rb1 a5 2.Rd1 a6 3.Kf7 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rxg2 Kxg2 9.Qe2+ Kg1 10.Qd1+ Kg2 11.Qc2+ Kg1 12.Qc1+ Kg2 13.Qb2+ Kg3 14.Qc1
-+ (-5.72 ++) Depth: 6/11 00:00:00 10kN
+- (3.57) Depth: 19/43 00:01:32 219mN

So does Crafty:

Analysis by Crafty 23.00:

1.Kg7
-+ (-8.03) Depth: 1/1 00:00:00 0kN

1.Rb1 a5 2.Rd1 a6 3.Kf7 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rf2 g1Q 9.Rxf3 Qa7+ 10.Kg6 Qd4 11.Kf5 Qd1 12.Qe4 Qd7+ 13.Kf6 Qd6+ 14.Kg5 Qd2+
= (0.00) Depth: 8/11 00:00:00 4kN
+- (2.49) Depth: 24/42 00:00:15 70538kN
Impressive. But I think they are finding the right move 2. Rd1 but for the wrong reason. Are they seeing that White has to play 6 e8B and not 6 e8 e8Q? I think if they follow the line they give, it will end in a draw.

More formally, Can the engines find move 3, which must be
3 Ra1 and not 3 Kf7? (Post below this one clarifies)
Last edited by afzzq on Fri May 14, 2010 12:31 pm, edited 1 time in total.
afzzq

Re: Endgame study, Costeff 1996

Post by afzzq »

Important edit to original post: Can an engine find the first THREE moves?

The first move, 1 Rb1 is not that hard.
The second move, after 1 Rb1 a5 2 Rd1 has been found by some engines.

By what about the third move, the really hard one:

1 Rb1 a5 2 Rd1 a6 3 Ra1!!
User avatar
AdminX
Posts: 6362
Joined: Mon Mar 13, 2006 2:34 pm
Location: Acworth, GA

Re: Endgame study, Costeff 1996

Post by AdminX »

[d]5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Yace Paderborn experimental - Book by Daniel Mehrmann:

1.Rb1 a5 2.Rd1 a6 3.Ra1 d5 4.e6 d4 5.Rxg1+ hxg1Q 6.e7
+- (8.47) Depth: 6/2 00:00:00 0kN
+- (8.47) Depth: 20/51 00:12:39 1817mN
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
User avatar
AdminX
Posts: 6362
Joined: Mon Mar 13, 2006 2:34 pm
Location: Acworth, GA

Re: Endgame study, Costeff 1996

Post by AdminX »

I restarted this one by closing and restarting Fritz to make sure there was no cached data.

5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Yace Paderborn experimental - Book by Daniel Mehrmann:

1.Rb1 a5 2.Ra1 a6 3.Rd1 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rd3 g1Q 9.Qc6 Qg2 10.Rxf3 Qxf3+ 11.Qxf3+ Kg1 12.Qe3+ Kg2 13.Qe4+ Kg1 14.Qe1+ Kg2
-+ (-5.72 ++) Depth: 6/11 00:00:00 10kN
+- (7.09) Depth: 22/52 00:15:27 2140mN
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
afzzq

Re: Endgame study, Costeff 1996

Post by afzzq »

AdminX wrote:I restarted this one by closing and restarting Fritz to make sure there was no cached data.

5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Yace Paderborn experimental - Book by Daniel Mehrmann:

1.Rb1 a5 2.Ra1 a6 3.Rd1 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rd3 g1Q 9.Qc6 Qg2 10.Rxf3 Qxf3+ 11.Qxf3+ Kg1 12.Qe3+ Kg2 13.Qe4+ Kg1 14.Qe1+ Kg2
-+ (-5.72 ++) Depth: 6/11 00:00:00 10kN
+- (7.09) Depth: 22/52 00:15:27 2140mN
Well, this specific line is a draw after 2. Ra1. I downloaded a version of Yace and it found the right line after I played 3 Ra1, although not before. Where did you get Paderborn experimental, what is the hash size, are you using tablebases?
User avatar
AdminX
Posts: 6362
Joined: Mon Mar 13, 2006 2:34 pm
Location: Acworth, GA

Re: Endgame study, Costeff 1996

Post by AdminX »

afzzq wrote:
AdminX wrote:I restarted this one by closing and restarting Fritz to make sure there was no cached data.

5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Yace Paderborn experimental - Book by Daniel Mehrmann:

1.Rb1 a5 2.Ra1 a6 3.Rd1 d5 4.e6 d4 5.e7 d3 6.e8Q d2 7.Rxd2 Nf3 8.Rd3 g1Q 9.Qc6 Qg2 10.Rxf3 Qxf3+ 11.Qxf3+ Kg1 12.Qe3+ Kg2 13.Qe4+ Kg1 14.Qe1+ Kg2
-+ (-5.72 ++) Depth: 6/11 00:00:00 10kN
+- (7.09) Depth: 22/52 00:15:27 2140mN
Well, this specific line is a draw after 2. Ra1. I downloaded a version of Yace and it found the right line after I played 3 Ra1, although not before. Where did you get Paderborn experimental, what is the hash size, are you using tablebases?

I downloaded it from Jim Ablett's website. Note that I let it continue to run and the correct move order was produced. The hash was set to 1024 and there were no tablebase hits. I do have 3,4,5, and 6 men installed.

http://www.mediafire.com/file/ygyxd1djy ... 003_ja.zip

[d]5K2/p2p4/p7/4P3/P7/8/1R4pp/6nk w - - 0 1

Analysis by Yace Paderborn experimental - Book by Daniel Mehrmann:

1.Rb1 a5 2.Rd1 a6 3.Ra1 d5 4.e6 d4 5.e7 d3 6.e8B d2
-+ (-5.72 ++) Depth: 6/11 00:00:00 10kN
+- (7.10) Depth: 23/54 00:40:37 4293mN
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
BBauer
Posts: 658
Joined: Wed Mar 08, 2006 8:58 pm

Re: Endgame study, Costeff 1996

Post by BBauer »

This study looks hard.
Even the following position is not easy for stockfish
[D]5K2/4P3/p7/p7/P7/3p4/6pp/R5nk w - - 0 6

Code: Select all

  35	01:09	 173.090.595	2.497.015	 0,00	e7e8Q d3d2 Ra1b1 d2d1R Rb1xd1
  36	01:47	 275.643.679	2.553.840	 0,00	e7e8Q d3d2 Ra1b1 d2d1R Rb1xd1
  37	02:28	 388.631.259	2.617.556	 0,00	e7e8Q d3d2 Ra1b1 d2d1R Rb1xd1
  38	03:19	 525.283.403	2.626.916	 0,00	e7e8Q d3d2 Ra1b1 d2d1R Rb1xd1
  39	04:39	 755.546.691	2.701.632	 0,00	e7e8Q d3d2 Ra1b1 d2d1R Rb1xd1
  39+	05:37	 919.549.681	2.728.059	+0,08	e7e8B d3d2 Be8h5 d2d1N Bh5xd1 Ng1h3 Bd1c2+ 
  39+	06:02	 976.027.933	2.693.367	+10,34	e7e8B d3d2 Be8h5 d2d1N Bh5xd1 Ng1h3 Bd1c2+ 
  39+	06:02	 976.357.799	2.692.242	+20,68	e7e8B d3d2 Be8h5 d2d1N Bh5xd1 Ng1h3 Bd1c2+ 
  39+	06:02	 976.538.042	2.691.551	+82,74	e7e8B d3d2 Be8h5 d2d1N Bh5xd1 Ng1h3 Bd1c2+ 
kind regards
Bernhard