EGTB Testsuite

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

Moderators: hgm, Rebel, chrisw

MoKy

Re: EGTB Testsuite

Post by MoKy »

Can you test this one?
[D]8/8/1P3k1r/8/2K1b1P1/3R1r2/7b/4N3 w - -
IMO the best move is Rxf3 and probably draw in BB:N ending due to 50 moves rule.
1.Rxf3+ Bxf3 2.g5+ Ke6! 3.gxh6 Bd5+ 4.Kd3 Ke5 5.Nc2 Bg1
Top
User avatar
Kirill Kryukov
Posts: 492
Joined: Sun Mar 19, 2006 4:12 am

Re: EGTB Testsuite

Post by Kirill Kryukov »

MoKy wrote:I think for this test are preferable drawn positions. Here is a study by Akobia+Becker 2005 [D]8/8/6p1/7R/b6K/5k2/3n4/2n5 w - - 0 1 bm Rh6=

And I add the next one by Kekely 2003
[D]3N4/7P/5pK1/1p5B/2k1b3/8/4rn2/8 w - - 0 1 bm Kxf6=
I analyzed these two positions and added to the collection with my comments. First one is "imperfect test", second is "possibly useful". Thanks!
Top
User avatar
Kirill Kryukov
Posts: 492
Joined: Sun Mar 19, 2006 4:12 am

Re: EGTB Testsuite

Post by Kirill Kryukov »

Jouni wrote:Here's traditional position to test, if 4 piece tb/bb access is working in the search:

[D]8/k7/8/2R5/8/4q3/8/4B2K w - - bm Bg3

Other moves lose I think. Actually I have seen solution without bases, but
it took a lot time. With bases You got solution instantly.

Jouni
Thanks and congratulations! Because as far as I understand we got our first perfect test position! Added to the list.
Top
User avatar
Kirill Kryukov
Posts: 492
Joined: Sun Mar 19, 2006 4:12 am

Re: EGTB Testsuite

Post by Kirill Kryukov »

MoKy wrote:Can you test this one?
[D]8/8/1P3k1r/8/2K1b1P1/3R1r2/7b/4N3 w - -
IMO the best move is Rxf3 and probably draw in BB:N ending due to 50 moves rule.
1.Rxf3+ Bxf3 2.g5+ Ke6! 3.gxh6 Bd5+ 4.Kd3 Ke5 5.Nc2 Bg1
Solution of a perfect endgame test should not depend on a 50 moves rule, otherwise not all kinds of tablebases can be tested. Glaurung 2.0.1 finds 1.Rxf3+ in 0 seconds (without any tables), I did not check other engines. Thanks anyway!
Top
Dann Corbit
Posts: 12537
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: EGTB Testsuite

Post by Dann Corbit »

Many of these will qualify (from an epd test set called pawntest). I guess that about 1/3 will meet your criteria. I will leave it to you to filter out the ones that you want.

I suggest also a basic test where all the answers are known and can be found by direct lookup, just to validate that the tablebase files are working.

Code: Select all

6B1/1pp3p1/3p4/n1k2P2/2P3P1/4K3/8/8 b - - bm Nc6; id "Koundinya Veluri, 9/25/2001@22:56:30 on CCC";
7k/p5p1/2n1p2p/3p4/4P2P/Prp2N2/5PPK/3R4 b - - bm Rb5 dxe4; id "Ron Murawski -- Unstoppable passed pawn";
8/8/2p5/8/8/5k2/P7/5K2 w - - bm a4; id "J&S42.01";
8/8/8/6K1/8/5k2/P7/8 w - - bm Kf5; id "J&S42.02";
8/4k3/8/8/4PK2/8/8/8 w - - bm Ke5; id "J&S42.03";
8/2k5/4P3/8/1K6/8/8/8 w - - bm Kc5; id "J&S42.04";
8/8/8/8/4P1k1/8/5K2/8 w - - bm Ke3; id "J&S42.05";
2k5/8/8/8/1PK5/8/8/8 b - - bm Kb8 Kc7; id "J&S42.06";
8/2k5/8/8/6K1/8/7P/8 w - - bm Kf5 Kg5 Kh5; id "J&S42.07";
k7/P7/1P6/1K6/8/8/8/8 w - - bm Kc5; id "J&S42.08";
6k1/8/5KP1/6P1/8/8/8/8 w - - bm g7; id "J&S42.09";
8/8/K1p1k3/2P5/8/8/8/8 w - - bm Kb7; id "J&S42.10";
8/p4K2/P7/8/8/8/1k6/8 w - - bm Ke6; id "J&S42.11";
8/8/8/7p/1PK2k2/8/8/8 w - - bm b5; id "J&S42.12";
8/8/1p6/8/8/6P1/k1K5/8 w - - bm Kc3; id "J&S42.13";
8/5p2/2k5/K7/8/1P6/8/8 b - - bm Kd5; id "J&S42.14";
8/8/4K3/4P2p/8/5k2/8/8 w - - bm Kf5; id "J&S42.15";
7K/8/k1P5/7p/8/8/8/8 w - - bm Kg7; id "J&S42.16";
8/p7/6k1/1P6/6KP/8/8/8 w - - bm Kf4 h5+; id "J&S43.01";
8/3p1k2/3P4/2P2K2/8/8/8/8 w - - bm Ke5; id "J&S43.02";
3k4/8/2pP1K2/2P5/8/8/8/8 w - - bm d7; id "J&S43.03";
3k4/1p6/1P1K4/2P5/8/8/8/8 w - - bm Ke6; id "J&S43.04";
3k4/2p5/2K5/1P1P4/8/8/8/8 w - - bm Kb7; id "J&S43.05";
6k1/7p/7K/7P/8/8/6P1/8 w - - bm g3; id "J&S43.06";
8/8/1kp5/8/K1PP4/8/8/8 w - - bm Kb3; id "J&S43.07";
8/2k5/p1P5/P1K5/8/8/8/8 w - - bm Kd5; id "J&S43.08";
8/6p1/8/5P2/5P2/5K2/8/6k1 w - - bm f6; id "J&S43.09";
8/6p1/8/4K3/7P/4k2P/8/8 w - - bm Ke6; id "J&S43.10";
4K3/8/2p5/8/P2k4/8/P7/8 w - - bm a5; id "J&S43.11";
8/8/1k1K4/pP6/P7/8/8/8 w - - bm Kd5 Ke6; c0 "Kd5 is a mate in 19 as shown by Yace"; id "J&S43.12";
8/8/8/4kPp1/6P1/4K3/8/8 w - - bm Kd2 Ke2; c0 "Ke2 is a mate in 21 as shown by Yace"; id "J&S43.13";
8/6k1/8/1p1K4/5P1P/8/8/8 w - - bm Kc5; id "J&S43.14";
8/1p6/8/8/K7/2P5/1P1k4/8 w - - bm Kb3 c4; id "J&S43.15?";
8/8/6p1/7p/5k2/8/4K2P/8 b - - bm Kg4; id "J&S43.16";
8/8/4k3/3p2p1/1P1K2P1/8/8/8 w - - bm Kc5 b5; id "J&S44.01";
8/ppp5/8/PPP5/8/7k/8/7K w - - bm b6; id "J&S44.02";
8/2k5/2Pp3p/1P6/8/5K2/8/8 w - - bm Kf4; id "J&S44.03";
7k/6p1/6P1/8/8/p5K1/P7/8 w - - bm Kf4; id "J&S44.04";
8/7k/4K2p/6p1/6P1/7P/8/8 w - - bm Kf7; id "J&S44.05";
8/8/5p2/pK6/2Pk2P1/8/8/8 w - - bm g5; id "J&S44.06";
8/8/8/KP6/1p6/k4p2/5P2/8 w - - bm b6; id "J&S44.07";
1k6/8/P1P5/8/5ppp/8/6K1/8 w - - bm Kg1; id "J&S44.08";
4K3/8/8/1p5p/1P5P/8/8/4k3 w - - bm Ke7; id "J&S44.09";
k7/P7/1P6/4p3/4Pp2/5K2/8/8 w - - bm Ke2; id "J&S44.10";
8/8/4k3/1pp5/2p1PP2/8/3K4/8 b - - bm b4; id "J&S44.11";
8/8/1p4K1/p1p5/P1P3k1/1P6/8/8 w - - bm Kf6; id "J&S44.12";
8/p2p4/8/8/8/k7/5P1P/7K w - - bm f4; id "J&S45.01";
8/8/8/1P2kp2/P2p2p1/6P1/3K4/8 b - - bm f4; id "J&S45.02";
8/5k2/6p1/5pKp/8/6P1/5P1P/8 w - - bm Kh6 f3 f4; c0 "Kh6 gives a huge score by ply 23 according to Yace"; id "J&S45.03";
k7/2p1pp2/2P3p1/4P1P1/5P2/p7/Kp3P2/8 w - - bm f5; id "J&S45.04";
8/2p5/1pPp4/1P1Pp3/4Pp1k/5P2/5KP1/8 w - - bm g3+; id "J&S45.05";
1k6/6p1/1Pp1pp1p/2P5/7P/8/5PP1/6K1 w - - bm h5; id "J&S45.06";
8/8/1p1k4/5ppp/PPK1p3/6P1/5PP1/8 b - - bm f4; id "J&S45.07";
8/5pp1/4p2p/3k3P/1p1P2P1/1P1K1P2/8/8 b - - bm g5; id "J&S45.08";
8/8/4k3/6p1/2pPpP2/4P2P/6K1/8 w - - bm f5+; id "J&S45.09";
3k4/2p5/1pKp4/p2P4/2P5/P7/1P6/8 w - - bm c5; c0 "This one is questionable."; id "J&S45.10";
8/2pp2pp/8/2PP1P2/1p5k/8/PP4p1/6K1 w - - bm f6; id "J&S45.11";
7k/8/5PpK/Pp1P2pp/3P4/8/5p2/8 w - - bm a6; id "J&S45.12";
8/p3k1p1/4r3/2ppNpp1/PP1P4/2P3KP/5P2/8 b - - bm Rxe5; id "WAC.087";
8/k1b5/P4p2/1Pp2p1p/K1P2P1P/8/3B4/8 w - - bm Be3 b6+; id "WAC.100";
8/1p3pp1/7p/5P1P/2k3P1/8/2K2P2/8 w - - bm f6; id "LCTII.FIN.01";
6k1/6p1/8/4K1P1/8/7P/8/8 w - - bm Kf4; id "bote.1";
k7/4p3/4p3/8/8/3P1P2/5P2/K7 w - - bm Kb2; id "bote.2";
8/4p3/2kp4/4p3/6K1/5P2/3P4/8 w - - bm Kg5; id "bote.3";
k7/8/1p6/p1p5/2P4K/8/PP6/8 w - - bm a4; id "bote.4";
8/8/p7/8/1P6/7p/P4k1P/3K4 w - - bm a3; id "bote.5";
8/5p1p/8/6k1/8/6P1/5PP1/7K w - - bm Kh2; id "bote.6";
8/2p4p/7P/2K5/2P5/8/k7/8 w - - bm Kb4; id "(JAFM) F. J. Prokop,1943";
8/6p1/p7/5P1p/1Pk5/8/5KPP/8 w - - bm h4; id "(JAFM) Plaskett-Flear Hastings 1984/85";
8/8/8/1P2kp2/P2p2p1/6P1/3K4/8 b - - bm f4; id "(JAFM) Stoltz-Nimzowitsch,1928"; pv f4! gxf4 Kd6! a5 g3 a6 Kc7!;
8/6K1/8/ppp2k2/8/1P6/1P5P/8 w - - bm h4; ce 0; id "(JAFM) Lasker-Tarrasch,1914"; pv h4 Kg4 Kg6!;
k7/8/pp6/2p4K/8/PPP5/8/8 w - - bm c4; id "(Simon Finn on CCC) Grigoriev 1933"; pv c4 Kb7 a4 Kc6 Kg5 Kc7 Kf6 Kd7 a5 Kd6 Kf5 Kc7 Ke6 Kc6 Ke7 Kc7 Ke8 Kf8 axb6 Kb7 Kd7 Kxb6 Kc8 Kb6;
7k/6p1/p4p1p/2p5/1pPP4/1P3R1K/r7/B7 w - - bm dxc5; id "(ML) Gabdushev-Hasangatin, Bor 2000";
8/p7/K6p/1P4p1/2k5/8/6PP/8 w - - bm h3; c0 "Draws."; id "(JAFM) CCC#191202";
8/5p2/7p/5pk1/8/5KPP/8/8 b - - bm Kf6 h5; c0 "h5 is a mate in 43 according to Yace"; id "(JAFM) CCC#191597";
Top
User avatar
Kirill Kryukov
Posts: 492
Joined: Sun Mar 19, 2006 4:12 am

Re: EGTB Testsuite

Post by Kirill Kryukov »

Thanks Dann! This will keep me busy for a while. :-)
Dann Corbit wrote:I suggest also a basic test where all the answers are known and can be found by direct lookup, just to validate that the tablebase files are working.
Yeah this may be useful... I will add basic test cathegory.
Top
Edmund
Posts: 670
Joined: Mon Dec 03, 2007 3:01 pm
Location: Barcelona, Spain

Re: EGTB Testsuite

Post by Edmund »

Try this one please:

[D]2k4q/3R4/1P6/8/8/8/4P3/2K5 w - - 0 1

only Rd3 draws (forming a fortress)

Rc7 loses in 56
Top
Edmund
Posts: 670
Joined: Mon Dec 03, 2007 3:01 pm
Location: Barcelona, Spain

Re: EGTB Testsuite

Post by Edmund »

And another KRPPKQ

[D]3k3q/2R5/1P6/8/8/8/4P3/2K5 w - - 0 1

Rc4 wins

Engines prefer Kingmoves
Top
User avatar
David Dahlem
Posts: 900
Joined: Wed Mar 08, 2006 9:06 pm

Re: EGTB Testsuite

Post by David Dahlem »

Codeman wrote:Try this one please:

[D]2k4q/3R4/1P6/8/8/8/4P3/2K5 w - - 0 1

only Rd3 draws (forming a fortress)

Rc7 loses in 56
Easy for Sage, without tablebases ...

FEN: 2k4q/3R4/1P6/8/8/8/4P3/2K5 w - - 0 1

Sage 3.38:
4 00:00 454 3.492 0.00 Rd7d3 Kc8b7 Kc1c2
5 00:00 593 4.561 0.00 Rd7d3 Kc8b7
6 00:00 1.186 9.123 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6
7 00:00 2.364 18.184 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6 Kc2b1
8 00:00 6.142 47.246 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6 Rd3b3+ Kb6a5 Kc2b1
9 00:00 13.067 100.515 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6 Rd3b3+ Kb6a5 Rb3a3+ Ka5b4 Ra3b3+ Kb4a5
10 00:00 54.874 322.788 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6 Rd3b3+ Kb6a5 Rb3a3+ Ka5b4 Ra3b3+ Kb4a5
11 00:00 129.366 588.027 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6 Rd3b3+ Kb6a5 Rb3a3+ Ka5b4 Ra3b3+ Kb4a5
12 00:00 398.739 886.086 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6 Rd3b3+ Kb6a5 Rb3a3+ Ka5b4 Ra3b3+ Kb4a5
13 00:02 1.873.906 1.035.307 0.00 Rd7d3 Kc8b7 Kc1c2 Kb7xb6 Rd3b3+ Kb6a5 Rb3a3+ Ka5b4 Ra3b3+ Kb4a5

Regards
Dave
Top
User avatar
David Dahlem
Posts: 900
Joined: Wed Mar 08, 2006 9:06 pm

Re: EGTB Testsuite

Post by David Dahlem »

Codeman wrote:And another KRPPKQ

[D]3k3q/2R5/1P6/8/8/8/4P3/2K5 w - - 0 1

Rc4 wins

Engines prefer Kingmoves
Also easy for Sage, without tablebases ...

FEN: 3k3q/2R5/1P6/8/8/8/4P3/2K5 w - - 0 1

Sage 3.38:
4 00:00 284 2.184 0.00 Kc1b1 Qh8h1+
5 00:00 366 2.815 0.00 Kc1b1 Qh8h1+
6 00:00 704 5.415 0.00 Kc1b1 Qh8h1+ Kb1a2 Qh1g2 Ka2a1
7 00:00 1.676 12.892 0.00 Kc1b1 Qh8h1+ Kb1a2 Qh1g2 Ka2a1 Qg2xe2
8 00:00 9.987 76.823 0.00 Kc1d1 Qh8d4+ Kd1c2 Qd4xb6 Rc7c3 Qb6a5 Kc2b2
9 00:00 39.128 244.550 0.00 Kc1c2 Qh8d4 Rc7c3 Qd4xb6 e2e3 Qb6g6+ Rc3d3+ Kd8e8 Kc2d2
10 00:00 198.450 640.161 0.00 Rc7c4 Qh8h1+ Kc1c2 Qh1e1 Kc2b3 Qe1xe2 Rc4d4+ Kd8c8 b6b7+ Kc8xb7 Kb3c3
11 00:00 398.409 766.171 0.00 Rc7c4 Qh8h1+ Kc1c2 Qh1e1 Kc2b3 Qe1xe2 Rc4d4+ Kd8c8 b6b7+ Kc8xb7 Rd4b4+ Kb7c8 Kb3c3
12 00:01 911.310 893.441 0.00 Rc7c4 Qh8h1+ Kc1c2 Qh1e1 Kc2d3 Qe1g1 Kd3c2 Qg1xb6 Rc4c3 Kd8e7 Rc3d3
13 00:03 3.031.045 935.507 -0.01 Rc7c4 Qh8h6+ Kc1c2 Qh6xb6 Rc4c3 Kd8e7 Rc3d3 Qb6a7 Kc2b2 Qa7b8+ Kb2c2 Qb8c7+

Regards
Dave
Top

Return to “Computer Chess Club: General Topics”