STS re-re-re-re-re-visited

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

Moderator: Ras

Norm Pollock
Posts: 1070
Joined: Thu Mar 09, 2006 4:15 pm
Location: Long Island, NY, USA

Re: STS re-re-re-re-re-visited

Post by Norm Pollock »

I think STS is a great project for aiding in the development of chess engines. And checking the results with updated engines is a smart idea.

I do have issues though with some of the positions. I think replacements should be considered.

First, three of the positions are duplicates. It is reasonable that a position could represent two different categories, but why multi-task the same position? (The following refer to lines in SF15-sts.epd)

1q2r1k1/1b2bpp1/p2ppn1p/2p5/P3PP1B/2PB1RP1/2P1Q2P/2KR4 b - - c0 2; c1 line(s): 605 1405;
1r1r1bk1/1bq2p1p/pn2p1p1/2p1P3/5P2/P1NBB3/1P3QPP/R2R2K1 b - - c0 2; c1 line(s): 208 804;
r1r3k1/pb3p1p/1pqBp1p1/4P3/3b4/2P2P2/PR1N2PP/2RQ3K w - - c0 2; c1 line(s): 277 873;

Secondly, seven of the positions have tripled pawns in the pawn structure of one of the colors. Tripled pawns are rarely seen in top-level games because top players will avoid them. Doubled pawns are bad enough, tripled pawns are a disaster. With such an inherent weakness in a position, testing for other weaknesses is compromised.

4rbk1/p2R1p2/2p2p1p/5p2/8/PR2p3/2P3PP/5K2 w - - c0 "Ke2=10";
1kb4r/1p3pr1/3b1p1p/q2B1p2/p7/P1P3P1/1P1Q2NP/K2RR3 b - - c0 "Be5=10, Bc7=4";
8/2p1r1pk/2pq2bp/2p5/2P3Q1/pP1P2NP/P3PRK1/8 b - - c0 "Re5=10, Qe5=10, Qd8=10, Re3=10";
1k1rr2b/2q2p2/p4p1P/1p3p2/PR2bQ2/3B4/1PP5/1K1RN3 w - - c0 "Qf2=10, Qf1=9, Qc1=6, Qd2=4";
1r2r1k1/pb1n1pp1/1p1p2np/2P1p1q1/P1P1P3/2P2PP1/2QNBB1P/1R1R3K b - - c0 "Nxc5=10";
3rr1k1/p4bbp/1p1p4/3p1BPp/P2p1N1P/2P1P1K1/R1P5/3R4 w - - c0 "cxd4=10, exd4=10";
br1r2k1/p2n1pp1/np1p4/2P1p1qp/2P1P3/P1P2PP1/2QNBB1P/2R2R1K b - - c0 "Naxc5=10, Ndxc5=3";
Updated links for 40H Tools and Databases
http://40Hchess.epizy.com
http://nk-qy.info/40h
Top
peter
Posts: 3384
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: STS re-re-re-re-re-visited

Post by peter »

Norm Pollock wrote: ↑Sun Oct 30, 2022 2:53 pm I think STS is a great project for aiding in the development of chess engines. And checking the results with updated engines is a smart idea.

I do have issues though with some of the positions. I think replacements should be considered.

First, three of the positions are duplicates. It is reasonable that a position could represent two different categories, but why multi-task the same position? (The following refer to lines in SF15-sts.epd)

1q2r1k1/1b2bpp1/p2ppn1p/2p5/P3PP1B/2PB1RP1/2P1Q2P/2KR4 b - - c0 2; c1 line(s): 605 1405;
1r1r1bk1/1bq2p1p/pn2p1p1/2p1P3/5P2/P1NBB3/1P3QPP/R2R2K1 b - - c0 2; c1 line(s): 208 804;
r1r3k1/pb3p1p/1pqBp1p1/4P3/3b4/2P2P2/PR1N2PP/2RQ3K w - - c0 2; c1 line(s): 277 873;

Secondly, seven of the positions have tripled pawns in the pawn structure of one of the colors. Tripled pawns are rarely seen in top-level games because top players will avoid them. Doubled pawns are bad enough, tripled pawns are a disaster. With such an inherent weakness in a position, testing for other weaknesses is compromised.

4rbk1/p2R1p2/2p2p1p/5p2/8/PR2p3/2P3PP/5K2 w - - c0 "Ke2=10";
1kb4r/1p3pr1/3b1p1p/q2B1p2/p7/P1P3P1/1P1Q2NP/K2RR3 b - - c0 "Be5=10, Bc7=4";
8/2p1r1pk/2pq2bp/2p5/2P3Q1/pP1P2NP/P3PRK1/8 b - - c0 "Re5=10, Qe5=10, Qd8=10, Re3=10";
1k1rr2b/2q2p2/p4p1P/1p3p2/PR2bQ2/3B4/1PP5/1K1RN3 w - - c0 "Qf2=10, Qf1=9, Qc1=6, Qd2=4";
1r2r1k1/pb1n1pp1/1p1p2np/2P1p1q1/P1P1P3/2P2PP1/2QNBB1P/1R1R3K b - - c0 "Nxc5=10";
3rr1k1/p4bbp/1p1p4/3p1BPp/P2p1N1P/2P1P1K1/R1P5/3R4 w - - c0 "cxd4=10, exd4=10";
br1r2k1/p2n1pp1/np1p4/2P1p1qp/2P1P3/P1P2PP1/2QNBB1P/2R2R1K b - - c0 "Naxc5=10, Ndxc5=3";
I support that. In my collection out of STS I kept only about 590 of the 1500, other's being too easy for modern engines, even with very short TC, having multiple solutions makes positions even easier to be "solved" getting points for even second and third best moves yet still.

The MEA- tool is a great one, I'm trying to use it for suites of some higher difficulty over all. Single best moves in a very strict sense are rare compared to winners or best defending moves with candidate moves not much worse than best ones, still winning or holding draw. These to reward with points related to the difficulty of the solutions as for hardware- time of modern engines and the singularity of the best moves according to engines' eval- relations would make it much easier to find usable test positions anyhow.
Two postings above I tried to explain with examples, what I mean.
Pity I don't make more than about 20 postions per day as for .epd- syntax for MEA.
Yet I will show about 100 as a first one try next week, I guess, regards
Peter.
Top
smatovic
Posts: 3222
Joined: Wed Mar 10, 2010 10:18 pm
Location: Hamburg, Germany
Full name: Srdja Matovic

Re: STS re-re-re-re-re-visited

Post by smatovic »

From the developers point of view: engine development ranges from 1000 to 2000 to 3000+ Elo, with the full STS set with weak/easy positions it is still useful for amateur chess programmers.

--
Srdja
Top
User avatar
Rebel
Posts: 7299
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: STS re-re-re-re-re-visited

Post by Rebel »

peter wrote: ↑Sun Oct 30, 2022 3:10 pm I support that. In my collection out of STS I kept only about 590 of the 1500, other's being too easy for modern engines, even with very short TC, having multiple solutions makes positions even easier to be "solved" getting points for even second and third best moves yet still.
Are these 590 available for download?
90% of coding is debugging, the other 10% is writing bugs.
Top
peter
Posts: 3384
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: STS re-re-re-re-re-visited

Post by peter »

Rebel wrote: ↑Sun Oct 30, 2022 8:03 pm
peter wrote: ↑Sun Oct 30, 2022 3:10 pm I support that. In my collection out of STS I kept only about 590 of the 1500, other's being too easy for modern engines, even with very short TC, having multiple solutions makes positions even easier to be "solved" getting points for even second and third best moves yet still.
Are these 590 available for download?
https://www.dropbox.com/s/x4qmr9ucykk0r76/594.epd?dl=0

These are the ones that are to be used as single best moves too, no MEA- points in .epd.

BTW, I have the 111 positions of ERET- suite from Walter Eigenmann

https://glarean-magazin.ch/2017/03/05/c ... test-eret/

with direct download- link from there:

https://glarean-magazin.ch/wp-content/u ... hess-1.zip

edited for MEA here (hope it's ok for Walter Eigenmann, didnwrite a mail to him about that not until right now)

Code: Select all

r1bqk1r1/1p1p1n2/p1n2pN1/2p1b2Q/2P1Pp2/1PN5/PB4PP/R4RK1 w q - bm Rxf4; c0 "Rf4=50"; id "ERET 1";
r1n2N1k/2n2K1p/3pp3/5Pp1/b5R1/8/1PPP4/8 w - - bm Ng6+; c0 "Ng6+=100"; id "2";
r1b1r1k1/1pqn1pbp/p2pp1p1/P7/1n1NPP1Q/2NBBR2/1PP3PP/R6K w - - bm f5; c0 "f5=50"; id "3";
5b2/p2k1p2/P3pP1p/n2pP1p1/1p1P2P1/1P1KBN2/7P/8 w - - bm Nxg5; c0 "Nxg5=25"; id "4";
r3kbnr/1b3ppp/pqn5/1pp1P3/3p4/1BN2N2/PP2QPPP/R1BR2K1 w kq - bm Bxf7+; c0 "Bxf7+=25"; id "5";
r2r2k1/1p1n1pp1/4pnp1/8/PpBRqP2/1Q2B1P1/1P5P/R5K1 b - - bm Nc5; c0 "Nc5=25"; id "6";
2rq1rk1/pb1n1ppN/4p3/1pb5/3P1Pn1/P1N5/1PQ1B1PP/R1B2RK1 b - - bm Nde5; c0 "Nde5=25"; id "7";
r2qk2r/ppp1bppp/2n5/3p1b2/3P1Bn1/1QN1P3/PP3P1P/R3KBNR w KQkq - bm Qxd5; c0 "Qxd5=25"; id "8";
rnb1kb1r/p4p2/1qp1pn2/1p2N2p/2p1P1p1/2N3B1/PPQ1BPPP/3RK2R w Kkq - bm Ng6; c0 "Ng6=30"; id "9";
5rk1/pp1b4/4pqp1/2Ppb2p/1P2p3/4Q2P/P3BPP1/1R3R1K b - - bm d4; c0 "d4=25"; id "10";
r1b2r1k/ppp2ppp/8/4p3/2BPQ3/P3P1K1/1B3PPP/n3q1NR w - - bm dxe5 Nf3; c0 "Nf3=25, dxe5=25"; id "11";
1nkr1b1r/5p2/1q2p2p/1ppbP1p1/2pP4/2N3B1/1P1QBPPP/R4RK1 w - - bm Nxd5; c0 "Nxd5=25, f4=5"; id "12";
1nrq1rk1/p4pp1/bp2pn1p/3p4/2PP1B2/P1PB2N1/4QPPP/1R2R1K1 w - - bm Qd2 Bc2; c0 "Qd2=25, Bc2=25"; id "13";
5k2/1rn2p2/3pb1p1/7p/p3PP2/PnNBK2P/3N2P1/1R6 w - - bm Nf3; c0 "Nf3=20, Nxb3=5, Kf2=1, g3=1"; id "14";
8/p2p4/r7/1k6/8/pK5Q/P7/b7 w - - bm Qd3+; c0 "Qd3=20"; id "15";
1b1rr1k1/pp1q1pp1/8/NP1p1b1p/1B1Pp1n1/PQR1P1P1/4BP1P/5RK1 w - - bm Nc6; c0 "Nc6=20"; id "16";
1r3rk1/6p1/p1pb1qPp/3p4/4nPR1/2N4Q/PPP4P/2K1BR2 b - - bm Rb7 Rxb2; c0 "Rb7=100, Rxb2=75"; id "17";
r1b1kb1r/1p1n1p2/p3pP1p/q7/3N3p/2N5/P1PQB1PP/1R3R1K b kq - bm Qg5; c0 "Qg5=25"; id "18";
3kB3/5K2/7p/3p4/3pn3/4NN2/8/1b4B1 w - - bm Nf5; c0 "Nf5=25"; id "19";
1nrrb1k1/1qn1bppp/pp2p3/3pP3/N2P3P/1P1B1NP1/PBR1QPK1/2R5 w - - bm Bxh7+; c0 "Bxh7=25"; id "20";
3rr1k1/1pq2b1p/2pp2p1/4bp2/pPPN4/4P1PP/P1QR1PB1/1R4K1 b - - bm Rc8 a3 h5; c0 "Rc8=25, a3=10, h5=10"; id "21";
r4rk1/p2nbpp1/2p2np1/q7/Np1PPB2/8/PPQ1N1PP/1K1R3R w - - bm h4; c0 "h4=25"; id "22";
r3r2k/1bq1nppp/p2b4/1pn1p2P/2p1P1QN/2P1N1P1/PPBB1P1R/2KR4 w - - bm Ng6+; c0 "Ng6+=25"; id "23";
r2q1r1k/3bppbp/pp1p4/2pPn1Bp/P1P1P2P/2N2P2/1P1Q2P1/R3KB1R w KQ - am b3; c0 "b3=-25"; id "24";
2kb4/p7/r1p3p1/p1P2pBp/R2P3P/2K3P1/5P2/8 w - - bm Bxd8; c0 "Bxd8=25"; id "25";
rqn2rk1/pp2b2p/2n2pp1/1N2p3/5P1N/1PP1B3/4Q1PP/R4RK1 w - - bm Nxg6; c0 "Nxg6=25"; id "26";
8/3Pk1p1/1p2P1K1/1P1Bb3/7p/7P/6P1/8 w - - bm g4; c0 "g4=25"; id "27";
4rrk1/Rpp3pp/6q1/2PPn3/4p3/2N5/1P2QPPP/5RK1 w - - am Rxb7; c0 "Rxb7=-25"; id "28";
2q2rk1/2p2pb1/PpP1p1pp/2n5/5B1P/3Q2P1/4PPN1/2R3K1 w - - bm Rxc5; c0 "Rxc5=25"; id "29";
rnbq1r1k/4p1bP/p3p3/1pn5/8/2Np1N2/PPQ2PP1/R1B1KB1R w KQ - bm Nh4; c0 "Nh4=25"; id "30";
4b1k1/1p3p2/4pPp1/p2pP1P1/P2P4/1P1B4/8/2K5 w - - bm b4; c0 "b4=25"; id "31";
8/7p/5P1k/1p5P/5p2/2p1p3/P1P1P1P1/1K3Nb1 w - - bm Ng3; c0 "Ng3=100"; id "32";
r3kb1r/ppnq2pp/2n5/4pp2/1P1PN3/P4N2/4QPPP/R1B1K2R w KQkq - bm Nxe5; c0 "Nxe5=25"; id "33";
b4r1k/6bp/3q1ppN/1p2p3/3nP1Q1/3BB2P/1P3PP1/2R3K1 w - - bm Rc8; c0 "Rc8=25"; id "34";
r3k2r/5ppp/3pbb2/qp1Np3/2BnP3/N7/PP1Q1PPP/R3K2R w KQkq - bm Nxb5; c0 "Nxb5=25, Bxb5=5, Qxa5=5"; id "35";
r1k1n2n/8/pP6/5R2/8/1b1B4/4N3/1K5N w - - bm b7+; c0 "b7+=25"; id "36";
1k6/bPN2pp1/Pp2p3/p1p5/2pn4/3P4/PPR5/1K6 w - - bm Na8; c0 "Na8=100, Rf2=10"; id "37";
8/6N1/3kNKp1/3p4/4P3/p7/P6b/8 w - - bm exd5; c0 "exd5=25"; id "38";
r1b1k2r/pp3ppp/1qn1p3/2bn4/8/6P1/PPN1PPBP/RNBQ1RK1 w kq - bm a3; c0 "a3=25, e4=5"; id "39";
r3kb1r/3n1ppp/p3p3/1p1pP2P/P3PBP1/4P3/1q2B3/R2Q1K1R b kq - bm Bc5; c0 "Bc5=30"; id "40";
3q1rk1/2nbppb1/pr1p1n1p/2pP1Pp1/2P1P2Q/2N2N2/1P2B1PP/R1B2RK1 w - - bm Nxg5; c0 "Nxg5=25"; id "41";
8/2k5/N3p1p1/2KpP1P1/b2P4/8/8/8 b - - bm Kb7; c0 "Kb7=25"; id "42";
2r1rbk1/1pqb1p1p/p2p1np1/P4p2/3NP1P1/2NP1R1Q/1P5P/R5BK w - - bm Nxf5; c0 "Nxf5=25, gxf5=5"; id "43";
rnb2rk1/pp2q2p/3p4/2pP2p1/2P1Pp2/2N5/PP1QBRPP/R5K1 w - - bm h4; c0 "h4=25"; id "44";
5rk1/p1p1rpb1/q1Pp2p1/3Pp2p/4Pn2/1R4N1/P1BQ1PPP/R5K1 w - - bm Kh1; c0 "Kh1=40, Rb4=20, Re1=20"; id "45";
8/4nk2/1p3p2/1r1p2pp/1P1R1N1P/6P1/3KPP2/8 w - - bm Nd3; c0 "Nd3=100"; id "46";
4kbr1/1b1nqp2/2p1p3/2N4p/1p1PP1pP/1PpQ2B1/4BPP1/r4RK1 w - - bm Nxb7; c0 "Nxb7=25"; id "47";
r1b2rk1/p2nqppp/1ppbpn2/3p4/2P5/1PN1PN2/PBQPBPPP/R4RK1 w - - bm cxd5; c0 "cxd5=25"; id "48";
r1b1kq1r/1p1n2bp/p2p2p1/3PppB1/Q1P1N3/8/PP2BPPP/R4RK1 w kq - bm Nc3; c0 "Nc3=30, f4=20, f3=20" "; id "49";
r4r1k/p1p3bp/2pp2p1/4nb2/N1P4q/1P5P/PBNQ1PP1/R4RK1 b - - bm Nf3+; c0 "Nf3+=30, g5=5, Bxc2=5; id "50";
6k1/pb1r1qbp/3p1p2/2p2p2/2P1rN2/1P1R3P/PB3QP1/3R2K1 b - - bm Bh6; c0 "Bh6=25"; id "51";
2r2r2/1p1qbkpp/p2ppn2/P1n1p3/4P3/2N1BB2/QPP2PPP/R4RK1 w - - bm b4; c0 "b4=25, g3=5, Qb1=5"; id "52";
r1bq1rk1/p4ppp/3p2n1/1PpPp2n/4P2P/P1PB1PP1/2Q1N3/R1B1K2R b KQ - bm c4; c0 "c4=25"; id "53";
2b1r3/5pkp/6p1/4P3/QppqPP2/5RPP/6BK/8 b - - bm c3; c0 "c3=25"; id "54";
r2q1rk1/1p2bpp1/p1b2n1p/8/5B2/2NB4/PP1Q1PPP/3R1RK1 w - - bm Bxh6; c0 "Bxh6=25"; id "55";
r2qr1k1/pp2bpp1/2pp3p/4nbN1/2P4P/4BP2/PPPQ2P1/1K1R1B1R w - - bm Be2; c0 "Be2=25, g4=10"; id "56";
r2qr1k1/pp1bbp2/n5p1/2pPp2p/8/P2PP1PP/1P2N1BK/R1BQ1R2 w - - bm d6; c0 "d6=20, e4=10, Nc3=10, Ng1=10"; id "57";
8/8/R7/1b4k1/5p2/1B3r2/7P/7K w - - bm h4+; c0 "h4+=25"; id "58";
rq6/5k2/p3pP1p/3p2p1/6PP/1PB1Q3/2P5/1K6 w - - bm Qd3; c0 "Qd3=25"; id "59";
q2B2k1/pb4bp/4p1p1/2p1N3/2PnpP2/PP3B2/6PP/2RQ2K1 b - - bm Qxd8; c0 "Qxd8=20, Bxe5=10, exf3=5"; id "60";
4rrk1/pp4pp/3p4/3P3b/2PpPp1q/1Q5P/PB4B1/R4RK1 b - - bm Rf6; c0 "Rf6=25"; id "61";
rr1nb1k1/2q1b1pp/pn1p1p2/1p1PpNPP/4P3/1PP1BN2/2B2P2/R2QR1K1 w - - bm g6; c0 "g6=25"; id "62";
r3k2r/4qn2/p1p1b2p/6pB/P1p5/2P5/5PPP/RQ2R1K1 b kq - bm Kf8; c0 "Kf8=25"; id "63";
8/1pp5/p3k1pp/8/P1p2PPP/2P2K2/1P3R2/5r2 b - - am Rxf2+; c0 "Rxf2+=-25"; id "64";
1r3rk1/2qbppbp/3p1np1/nP1P2B1/2p2P2/2N1P2P/1P1NB1P1/R2Q1RK1 b - - bm Qb6; c0 "Qb6=20, Rfc8=15, Bxb5=5"; id "65";
8/2pN1k2/p4p1p/Pn1R4/3b4/6Pp/1P3K1P/8 w - - bm Ke1; c0 "Ke1=25"; id "66";
5r1k/1p4bp/3p1q2/1NpP1b2/1pP2p2/1Q5P/1P1KBP2/r2RN2R b - - bm f3; c0 "f3=25"; id "67";
r3kb1r/pbq2ppp/1pn1p3/2p1P3/1nP5/1P3NP1/PB1N1PBP/R2Q1RK1 w kq - bm a3; c0 "a3=25"; id "68";
5rk1/n2qbpp1/pp2p1p1/3pP1P1/PP1P3P/2rNPN2/R7/1Q3RK1 w - - bm h5; c0 "h5=25"; id "69";
r5k1/1bqp1rpp/p1n1p3/1p4p1/1b2PP2/2NBB1P1/PPPQ4/2KR3R w - - bm a3; c0 "a3=20, Rh3=10"; id "70";
1r4k1/1nq3pp/pp1pp1r1/8/PPP2P2/6P1/5N1P/2RQR1K1 w - - bm f5; c0 "f5=25, Ne4=10"; id "71";
q5k1/p2p2bp/1p1p2r1/2p1np2/6p1/1PP2PP1/P2PQ1KP/4R1NR b - - bm Qd5; c0 "Qd5=25"; id "72"
r4rk1/ppp2ppp/1nnb4/8/1P1P3q/PBN1B2P/4bPP1/R2QR1K1 w - - bm Rxe2; c0 "Rxe2=25, Nxe2=15, Qxe2=10"; id "73";
1r3k2/2N2pp1/1pR2n1p/4p3/8/1P1K1P2/P5PP/8 w - - bm Kc4; c0 "Kc4=25, a4=10"; id "74";
6r1/6r1/2p1k1pp/p1pbP2q/Pp1p1PpP/1P1P2NR/1KPQ3R/8 b - - bm Qf5; c0 "Qf5=75"; id "75";
r1b1kb1r/1p1npppp/p2p1n2/6B1/3NPP2/q1N5/P1PQ2PP/1R2KB1R w Kkq - bm Bxf6; c0 "Bxf6=25"; id "76";
r3r1k1/1bq2ppp/p1p2n2/3ppPP1/4P3/1PbB4/PBP1Q2P/R4R1K w - - bm gxf6; c0 "gxf6=25"; id "77";
r4rk1/ppq3pp/2p1Pn2/4p1Q1/8/2N5/PP4PP/2KR1R2 w - - bm Rxf6; c0 "Rxf6=25"; id "78";
r1bqr1k1/3n1ppp/p2p1b2/3N1PP1/1p1B1P2/1P6/1PP1Q2P/2KR2R1 w - - bm Qxe8+; c0 "Qxe8+=25"; id "79";
5rk1/1ppbq1pp/3p3r/pP1PppbB/2P5/P1BP4/5PPP/3QRRK1 b - - bm Bc1; c0 "Bc1=25, b6=10"; id "80";
r3r1kb/p2bp2p/1q1p1npB/5NQ1/2p1P1P1/2N2P2/PPP5/2KR3R w - - bm Bg7; c0 "Bg7=25"; id "81";
8/3P4/1p3b1p/p7/P7/1P3NPP/4p1K1/3k4 w - - bm g4; c0 "g4=25"; id "82";
3q1rk1/7p/rp1n4/p1pPbp2/P1P2pb1/1QN4P/1B2B1P1/1R3RK1 w - - bm Nb5; c0 "Nb5=25"; id "83";
4r1k1/1r1np3/1pqp1ppB/p7/2b1P1PQ/2P2P2/P3B2R/3R2K1 w - - bm Bg5; c0 "Bg5=25, Bg7=15"; id "84";
r4rk1/q4bb1/p1R4p/3pN1p1/8/2N3P1/P4PP1/3QR1K1 w - - bm Ng4; c0 "Ng4=25, Rd6=15"; id "85";
r3k2r/pp2pp1p/8/q2Pb3/2P5/4p3/B1Q2PPP/2R2RK1 w kq - bm c5; c0 "c5=25"; id "86";
r3r1k1/1bnq1pbn/p2p2p1/1p1P3p/2p1PP1B/P1N2B1P/1PQN2P1/3RR1K1 w - - bm e5; c0 "e5=25, Nf1=10, b3=10, Kh1=5"; id "87";
8/4k3/p2p2p1/P1pPn2p/1pP1P2P/1P1NK1P1/8/8 w - - bm g4; c0 "g4=25"; id "88";
8/2P1P3/b1B2p2/1pPRp3/2k3P1/P4pK1/nP3p1p/N7 w - - bm e8=N; c0 "e8=N=75, c8=N=25"; id "89";
4K1k1/8/1p5p/1Pp3b1/8/1P3P2/P1B2P2/8 w - - bm f4; c0 "f4=100, Bd1=100"; id "90";
8/6p1/3k4/3p1p1p/p2K1P1P/4P1P1/P7/8 b - - bm g6 Kc6; c0 "g6=15, Kc6=15"; id "91";
r1b2rk1/ppp3p1/4p2p/4Qpq1/3P4/2PB4/PPK2PPP/R6R b - - am Qxg2; c0 "Qxg2=-25"; id "92";
2b1r3/r2ppN2/8/1p1p1k2/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nd6+; c0 "Nd6+=100"; id "93";
2k2Br1/p6b/Pq1r4/1p2p1b1/1Ppp2p1/Q1P3N1/5RPP/R3N1K1 b - - bm Rf6; c0 "Rf6=50"; id "94";
r2qk2r/ppp1b1pp/2n1p3/3pP1n1/3P2b1/2PB1NN1/PP4PP/R1BQK2R w KQkq - bm Nxg5; c0 "Nxg5=25"; id "95";
8/8/4p1Pk/1rp1K1p1/4P1P1/1nP2Q2/p2b1P2/8 w - - bm Kf6; c0 "Kf6=35"; id "96";
2k5/p7/Pp1p1b2/1P1P1p2/2P2P1p/3K3P/5B2/8 w - - bm c5; c0 "c5=10, Bg1=10, Be1=10, Be3=10"; id "97";
8/6pp/5k2/1p1r4/4R3/7P/5PP1/5K2 w - - am Ke2; c0 "Ke2=-25"; id "98";
3q1r1k/4RPp1/p6p/2pn4/2P5/1P6/P3Q2P/6K1 w - - bm Re8; c0 "Re8=35"; id "99";
rn2k2r/3pbppp/p3p3/8/Nq1Nn3/4B1P1/PP3P1P/R2Q1RK1 w k - bm Nf5; c0 "Nf5=35"; id "100";
r1b1kb1N/pppnq1pB/8/3p4/3P4/8/PPPK1nPP/RNB1R3 b q - bm Ne5; c0 "Ne5=25"; id "101";
N4rk1/pp1b1ppp/n3p1n1/3pP1Q1/1P1N4/8/1PP2PPP/q1B1KB1R b K - bm Nxb4; c0 "Nxb4=25"; id "102";
4k1br/1K1p1n1r/2p2pN1/P2p1N2/2P3pP/5B2/P2P4/8 w - - bm Kc8; c0 "Kc8=100"; id "103";
r1bqkb1r/ppp3pp/2np4/3N1p2/3pnB2/5N2/PPP1QPPP/2KR1B1R b kq - bm Ne7; c0 "Ne7=25"; id "104";
r3kb1r/pbqp1pp1/1pn1pn1p/8/3PP3/2PB1N2/3N1PPP/R1BQR1K1 w kq - bm e5 d5; c0 "e5=10, d5=10, Bb2=5, c4=5"; id "105";
r2r2k1/pq2bppp/1np1bN2/1p2B1P1/5Q2/P4P2/1PP4P/2KR1B1R b - - bm Bxf6; c0 "Bxf6=25"; id "106";
1r1r2k1/2pq3p/4p3/2Q1Pp2/1PNn1R2/P5P1/5P1P/4R2K b - - bm Rb5; c0 "Rb5=25"; id "107";
8/5p1p/3P1k2/p1P2n2/3rp3/1B6/P4R2/6K1 w - - bm Ba4; c0 "Ba4=25"; id "108";
2rbrnk1/1b3p2/p2pp3/1p4PQ/1PqBPP2/P1NR4/2P4P/5RK1 b - - bm Qxd4+; c0 "Qxd4+=25"; id "109";
4r1k1/1bq2r1p/p2p1np1/3Pppb1/P1P5/1N3P2/1R2B1PP/1Q1R2BK w - - bm c5; c0 "c5=25"; id "110";
8/8/8/8/4kp2/1R6/P2q1PPK/8 w - - bm a3; c0 "a3=25"; id "ERET 111";
Pity they don't work bug- free that way, strange enough they do up to 999msec/position and single threaded, but trying longer TC MEA starts working but stops after about half of the positions without storing the run.
Guess I should not combine bm- syntax with points for moves? On the other hand that worked as long as there were less positons collected and as said it does with not too long TCs, single threaded better than SMP.

Image

Could you have a look at the .epd above as for the syntax, please?

Here's the batch- file:
set MT=1000
set HASH=8
set THREADS=30
set PROTOCOL=uci
set EPD=epd\MEA111.epd
set MRL=mea_results

set EXE=engines\SF15.exe
set NAME="SF15 SMP"
mea.exe --engine %EXE% --name %NAME% --hash %HASH% --threads %THREADS% --protocol %PROTOCOL% --epd %EPD% --movetime %MT%
mrl %MRL% %MRL% %EPD% %MT%
Thanks again, regards
Peter.
Top
peter
Posts: 3384
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: STS re-re-re-re-re-visited

Post by peter »

peter wrote: ↑Sun Oct 30, 2022 8:42 pm Could you have a look at the .epd above as for the syntax, please?
Edit- time over.
The numeric height of points given to the single or multiple rewarded move(s) aren't best for all of the positions so far, that was just a quick first try to get any version working soon. Started a little too high, getinng more reluctant in further progress, that has all to be be revised again for a concerted level, if syntax is ok then and I've seen some results with more hardware- TCs of more engines and more positions like these. The 594 out of STS would fit for certain TCs as well as e.g. Arasan- suite would too as for bigger collections like STS. In the 1024 here

https://www.dropbox.com/s/yu7dnjpofa6bb8n/1024.epd?dl=0

I selected positions from STS, Eret, Arasan- suite, HTC and more rather difficult ones to compensate the easier others (added some more single best move opening- positions too then again for the counter-part, what I'd like to see statistically was a balance between selectivity and sensitivity so to say, for a certain hardware- time and engine- sample there must not be too many solved and not too many unsolved positions) these are together to be used in a single best move suite, no MEA- points so far, regards
Peter.
Top
User avatar
Rebel
Posts: 7299
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: STS re-re-re-re-re-visited

Post by Rebel »

peter wrote: ↑Sun Oct 30, 2022 8:42 pm
Rebel wrote: ↑Sun Oct 30, 2022 8:03 pm
peter wrote: ↑Sun Oct 30, 2022 3:10 pm I support that. In my collection out of STS I kept only about 590 of the 1500, other's being too easy for modern engines, even with very short TC, having multiple solutions makes positions even easier to be "solved" getting points for even second and third best moves yet still.
Are these 590 available for download?
https://www.dropbox.com/s/x4qmr9ucykk0r76/594.epd?dl=0

These are the ones that are to be used as single best moves too, no MEA- points in .epd.

BTW, I have the 111 positions of ERET- suite from Walter Eigenmann

https://glarean-magazin.ch/2017/03/05/c ... test-eret/

with direct download- link from there:

https://glarean-magazin.ch/wp-content/u ... hess-1.zip

edited for MEA here (hope it's ok for Walter Eigenmann, didnwrite a mail to him about that not until right now)

Code: Select all

r1bqk1r1/1p1p1n2/p1n2pN1/2p1b2Q/2P1Pp2/1PN5/PB4PP/R4RK1 w q - bm Rxf4; c0 "Rf4=50"; id "ERET 1";
r1n2N1k/2n2K1p/3pp3/5Pp1/b5R1/8/1PPP4/8 w - - bm Ng6+; c0 "Ng6+=100"; id "2";
r1b1r1k1/1pqn1pbp/p2pp1p1/P7/1n1NPP1Q/2NBBR2/1PP3PP/R6K w - - bm f5; c0 "f5=50"; id "3";
5b2/p2k1p2/P3pP1p/n2pP1p1/1p1P2P1/1P1KBN2/7P/8 w - - bm Nxg5; c0 "Nxg5=25"; id "4";
r3kbnr/1b3ppp/pqn5/1pp1P3/3p4/1BN2N2/PP2QPPP/R1BR2K1 w kq - bm Bxf7+; c0 "Bxf7+=25"; id "5";
r2r2k1/1p1n1pp1/4pnp1/8/PpBRqP2/1Q2B1P1/1P5P/R5K1 b - - bm Nc5; c0 "Nc5=25"; id "6";
2rq1rk1/pb1n1ppN/4p3/1pb5/3P1Pn1/P1N5/1PQ1B1PP/R1B2RK1 b - - bm Nde5; c0 "Nde5=25"; id "7";
r2qk2r/ppp1bppp/2n5/3p1b2/3P1Bn1/1QN1P3/PP3P1P/R3KBNR w KQkq - bm Qxd5; c0 "Qxd5=25"; id "8";
rnb1kb1r/p4p2/1qp1pn2/1p2N2p/2p1P1p1/2N3B1/PPQ1BPPP/3RK2R w Kkq - bm Ng6; c0 "Ng6=30"; id "9";
5rk1/pp1b4/4pqp1/2Ppb2p/1P2p3/4Q2P/P3BPP1/1R3R1K b - - bm d4; c0 "d4=25"; id "10";
r1b2r1k/ppp2ppp/8/4p3/2BPQ3/P3P1K1/1B3PPP/n3q1NR w - - bm dxe5 Nf3; c0 "Nf3=25, dxe5=25"; id "11";
1nkr1b1r/5p2/1q2p2p/1ppbP1p1/2pP4/2N3B1/1P1QBPPP/R4RK1 w - - bm Nxd5; c0 "Nxd5=25, f4=5"; id "12";
1nrq1rk1/p4pp1/bp2pn1p/3p4/2PP1B2/P1PB2N1/4QPPP/1R2R1K1 w - - bm Qd2 Bc2; c0 "Qd2=25, Bc2=25"; id "13";
5k2/1rn2p2/3pb1p1/7p/p3PP2/PnNBK2P/3N2P1/1R6 w - - bm Nf3; c0 "Nf3=20, Nxb3=5, Kf2=1, g3=1"; id "14";
8/p2p4/r7/1k6/8/pK5Q/P7/b7 w - - bm Qd3+; c0 "Qd3=20"; id "15";
1b1rr1k1/pp1q1pp1/8/NP1p1b1p/1B1Pp1n1/PQR1P1P1/4BP1P/5RK1 w - - bm Nc6; c0 "Nc6=20"; id "16";
1r3rk1/6p1/p1pb1qPp/3p4/4nPR1/2N4Q/PPP4P/2K1BR2 b - - bm Rb7 Rxb2; c0 "Rb7=100, Rxb2=75"; id "17";
r1b1kb1r/1p1n1p2/p3pP1p/q7/3N3p/2N5/P1PQB1PP/1R3R1K b kq - bm Qg5; c0 "Qg5=25"; id "18";
3kB3/5K2/7p/3p4/3pn3/4NN2/8/1b4B1 w - - bm Nf5; c0 "Nf5=25"; id "19";
1nrrb1k1/1qn1bppp/pp2p3/3pP3/N2P3P/1P1B1NP1/PBR1QPK1/2R5 w - - bm Bxh7+; c0 "Bxh7=25"; id "20";
3rr1k1/1pq2b1p/2pp2p1/4bp2/pPPN4/4P1PP/P1QR1PB1/1R4K1 b - - bm Rc8 a3 h5; c0 "Rc8=25, a3=10, h5=10"; id "21";
r4rk1/p2nbpp1/2p2np1/q7/Np1PPB2/8/PPQ1N1PP/1K1R3R w - - bm h4; c0 "h4=25"; id "22";
r3r2k/1bq1nppp/p2b4/1pn1p2P/2p1P1QN/2P1N1P1/PPBB1P1R/2KR4 w - - bm Ng6+; c0 "Ng6+=25"; id "23";
r2q1r1k/3bppbp/pp1p4/2pPn1Bp/P1P1P2P/2N2P2/1P1Q2P1/R3KB1R w KQ - am b3; c0 "b3=-25"; id "24";
2kb4/p7/r1p3p1/p1P2pBp/R2P3P/2K3P1/5P2/8 w - - bm Bxd8; c0 "Bxd8=25"; id "25";
rqn2rk1/pp2b2p/2n2pp1/1N2p3/5P1N/1PP1B3/4Q1PP/R4RK1 w - - bm Nxg6; c0 "Nxg6=25"; id "26";
8/3Pk1p1/1p2P1K1/1P1Bb3/7p/7P/6P1/8 w - - bm g4; c0 "g4=25"; id "27";
4rrk1/Rpp3pp/6q1/2PPn3/4p3/2N5/1P2QPPP/5RK1 w - - am Rxb7; c0 "Rxb7=-25"; id "28";
2q2rk1/2p2pb1/PpP1p1pp/2n5/5B1P/3Q2P1/4PPN1/2R3K1 w - - bm Rxc5; c0 "Rxc5=25"; id "29";
rnbq1r1k/4p1bP/p3p3/1pn5/8/2Np1N2/PPQ2PP1/R1B1KB1R w KQ - bm Nh4; c0 "Nh4=25"; id "30";
4b1k1/1p3p2/4pPp1/p2pP1P1/P2P4/1P1B4/8/2K5 w - - bm b4; c0 "b4=25"; id "31";
8/7p/5P1k/1p5P/5p2/2p1p3/P1P1P1P1/1K3Nb1 w - - bm Ng3; c0 "Ng3=100"; id "32";
r3kb1r/ppnq2pp/2n5/4pp2/1P1PN3/P4N2/4QPPP/R1B1K2R w KQkq - bm Nxe5; c0 "Nxe5=25"; id "33";
b4r1k/6bp/3q1ppN/1p2p3/3nP1Q1/3BB2P/1P3PP1/2R3K1 w - - bm Rc8; c0 "Rc8=25"; id "34";
r3k2r/5ppp/3pbb2/qp1Np3/2BnP3/N7/PP1Q1PPP/R3K2R w KQkq - bm Nxb5; c0 "Nxb5=25, Bxb5=5, Qxa5=5"; id "35";
r1k1n2n/8/pP6/5R2/8/1b1B4/4N3/1K5N w - - bm b7+; c0 "b7+=25"; id "36";
1k6/bPN2pp1/Pp2p3/p1p5/2pn4/3P4/PPR5/1K6 w - - bm Na8; c0 "Na8=100, Rf2=10"; id "37";
8/6N1/3kNKp1/3p4/4P3/p7/P6b/8 w - - bm exd5; c0 "exd5=25"; id "38";
r1b1k2r/pp3ppp/1qn1p3/2bn4/8/6P1/PPN1PPBP/RNBQ1RK1 w kq - bm a3; c0 "a3=25, e4=5"; id "39";
r3kb1r/3n1ppp/p3p3/1p1pP2P/P3PBP1/4P3/1q2B3/R2Q1K1R b kq - bm Bc5; c0 "Bc5=30"; id "40";
3q1rk1/2nbppb1/pr1p1n1p/2pP1Pp1/2P1P2Q/2N2N2/1P2B1PP/R1B2RK1 w - - bm Nxg5; c0 "Nxg5=25"; id "41";
8/2k5/N3p1p1/2KpP1P1/b2P4/8/8/8 b - - bm Kb7; c0 "Kb7=25"; id "42";
2r1rbk1/1pqb1p1p/p2p1np1/P4p2/3NP1P1/2NP1R1Q/1P5P/R5BK w - - bm Nxf5; c0 "Nxf5=25, gxf5=5"; id "43";
rnb2rk1/pp2q2p/3p4/2pP2p1/2P1Pp2/2N5/PP1QBRPP/R5K1 w - - bm h4; c0 "h4=25"; id "44";
5rk1/p1p1rpb1/q1Pp2p1/3Pp2p/4Pn2/1R4N1/P1BQ1PPP/R5K1 w - - bm Kh1; c0 "Kh1=40, Rb4=20, Re1=20"; id "45";
8/4nk2/1p3p2/1r1p2pp/1P1R1N1P/6P1/3KPP2/8 w - - bm Nd3; c0 "Nd3=100"; id "46";
4kbr1/1b1nqp2/2p1p3/2N4p/1p1PP1pP/1PpQ2B1/4BPP1/r4RK1 w - - bm Nxb7; c0 "Nxb7=25"; id "47";
r1b2rk1/p2nqppp/1ppbpn2/3p4/2P5/1PN1PN2/PBQPBPPP/R4RK1 w - - bm cxd5; c0 "cxd5=25"; id "48";
r1b1kq1r/1p1n2bp/p2p2p1/3PppB1/Q1P1N3/8/PP2BPPP/R4RK1 w kq - bm Nc3; c0 "Nc3=30, f4=20, f3=20" "; id "49";
r4r1k/p1p3bp/2pp2p1/4nb2/N1P4q/1P5P/PBNQ1PP1/R4RK1 b - - bm Nf3+; c0 "Nf3+=30, g5=5, Bxc2=5; id "50";
6k1/pb1r1qbp/3p1p2/2p2p2/2P1rN2/1P1R3P/PB3QP1/3R2K1 b - - bm Bh6; c0 "Bh6=25"; id "51";
2r2r2/1p1qbkpp/p2ppn2/P1n1p3/4P3/2N1BB2/QPP2PPP/R4RK1 w - - bm b4; c0 "b4=25, g3=5, Qb1=5"; id "52";
r1bq1rk1/p4ppp/3p2n1/1PpPp2n/4P2P/P1PB1PP1/2Q1N3/R1B1K2R b KQ - bm c4; c0 "c4=25"; id "53";
2b1r3/5pkp/6p1/4P3/QppqPP2/5RPP/6BK/8 b - - bm c3; c0 "c3=25"; id "54";
r2q1rk1/1p2bpp1/p1b2n1p/8/5B2/2NB4/PP1Q1PPP/3R1RK1 w - - bm Bxh6; c0 "Bxh6=25"; id "55";
r2qr1k1/pp2bpp1/2pp3p/4nbN1/2P4P/4BP2/PPPQ2P1/1K1R1B1R w - - bm Be2; c0 "Be2=25, g4=10"; id "56";
r2qr1k1/pp1bbp2/n5p1/2pPp2p/8/P2PP1PP/1P2N1BK/R1BQ1R2 w - - bm d6; c0 "d6=20, e4=10, Nc3=10, Ng1=10"; id "57";
8/8/R7/1b4k1/5p2/1B3r2/7P/7K w - - bm h4+; c0 "h4+=25"; id "58";
rq6/5k2/p3pP1p/3p2p1/6PP/1PB1Q3/2P5/1K6 w - - bm Qd3; c0 "Qd3=25"; id "59";
q2B2k1/pb4bp/4p1p1/2p1N3/2PnpP2/PP3B2/6PP/2RQ2K1 b - - bm Qxd8; c0 "Qxd8=20, Bxe5=10, exf3=5"; id "60";
4rrk1/pp4pp/3p4/3P3b/2PpPp1q/1Q5P/PB4B1/R4RK1 b - - bm Rf6; c0 "Rf6=25"; id "61";
rr1nb1k1/2q1b1pp/pn1p1p2/1p1PpNPP/4P3/1PP1BN2/2B2P2/R2QR1K1 w - - bm g6; c0 "g6=25"; id "62";
r3k2r/4qn2/p1p1b2p/6pB/P1p5/2P5/5PPP/RQ2R1K1 b kq - bm Kf8; c0 "Kf8=25"; id "63";
8/1pp5/p3k1pp/8/P1p2PPP/2P2K2/1P3R2/5r2 b - - am Rxf2+; c0 "Rxf2+=-25"; id "64";
1r3rk1/2qbppbp/3p1np1/nP1P2B1/2p2P2/2N1P2P/1P1NB1P1/R2Q1RK1 b - - bm Qb6; c0 "Qb6=20, Rfc8=15, Bxb5=5"; id "65";
8/2pN1k2/p4p1p/Pn1R4/3b4/6Pp/1P3K1P/8 w - - bm Ke1; c0 "Ke1=25"; id "66";
5r1k/1p4bp/3p1q2/1NpP1b2/1pP2p2/1Q5P/1P1KBP2/r2RN2R b - - bm f3; c0 "f3=25"; id "67";
r3kb1r/pbq2ppp/1pn1p3/2p1P3/1nP5/1P3NP1/PB1N1PBP/R2Q1RK1 w kq - bm a3; c0 "a3=25"; id "68";
5rk1/n2qbpp1/pp2p1p1/3pP1P1/PP1P3P/2rNPN2/R7/1Q3RK1 w - - bm h5; c0 "h5=25"; id "69";
r5k1/1bqp1rpp/p1n1p3/1p4p1/1b2PP2/2NBB1P1/PPPQ4/2KR3R w - - bm a3; c0 "a3=20, Rh3=10"; id "70";
1r4k1/1nq3pp/pp1pp1r1/8/PPP2P2/6P1/5N1P/2RQR1K1 w - - bm f5; c0 "f5=25, Ne4=10"; id "71";
q5k1/p2p2bp/1p1p2r1/2p1np2/6p1/1PP2PP1/P2PQ1KP/4R1NR b - - bm Qd5; c0 "Qd5=25"; id "72"
r4rk1/ppp2ppp/1nnb4/8/1P1P3q/PBN1B2P/4bPP1/R2QR1K1 w - - bm Rxe2; c0 "Rxe2=25, Nxe2=15, Qxe2=10"; id "73";
1r3k2/2N2pp1/1pR2n1p/4p3/8/1P1K1P2/P5PP/8 w - - bm Kc4; c0 "Kc4=25, a4=10"; id "74";
6r1/6r1/2p1k1pp/p1pbP2q/Pp1p1PpP/1P1P2NR/1KPQ3R/8 b - - bm Qf5; c0 "Qf5=75"; id "75";
r1b1kb1r/1p1npppp/p2p1n2/6B1/3NPP2/q1N5/P1PQ2PP/1R2KB1R w Kkq - bm Bxf6; c0 "Bxf6=25"; id "76";
r3r1k1/1bq2ppp/p1p2n2/3ppPP1/4P3/1PbB4/PBP1Q2P/R4R1K w - - bm gxf6; c0 "gxf6=25"; id "77";
r4rk1/ppq3pp/2p1Pn2/4p1Q1/8/2N5/PP4PP/2KR1R2 w - - bm Rxf6; c0 "Rxf6=25"; id "78";
r1bqr1k1/3n1ppp/p2p1b2/3N1PP1/1p1B1P2/1P6/1PP1Q2P/2KR2R1 w - - bm Qxe8+; c0 "Qxe8+=25"; id "79";
5rk1/1ppbq1pp/3p3r/pP1PppbB/2P5/P1BP4/5PPP/3QRRK1 b - - bm Bc1; c0 "Bc1=25, b6=10"; id "80";
r3r1kb/p2bp2p/1q1p1npB/5NQ1/2p1P1P1/2N2P2/PPP5/2KR3R w - - bm Bg7; c0 "Bg7=25"; id "81";
8/3P4/1p3b1p/p7/P7/1P3NPP/4p1K1/3k4 w - - bm g4; c0 "g4=25"; id "82";
3q1rk1/7p/rp1n4/p1pPbp2/P1P2pb1/1QN4P/1B2B1P1/1R3RK1 w - - bm Nb5; c0 "Nb5=25"; id "83";
4r1k1/1r1np3/1pqp1ppB/p7/2b1P1PQ/2P2P2/P3B2R/3R2K1 w - - bm Bg5; c0 "Bg5=25, Bg7=15"; id "84";
r4rk1/q4bb1/p1R4p/3pN1p1/8/2N3P1/P4PP1/3QR1K1 w - - bm Ng4; c0 "Ng4=25, Rd6=15"; id "85";
r3k2r/pp2pp1p/8/q2Pb3/2P5/4p3/B1Q2PPP/2R2RK1 w kq - bm c5; c0 "c5=25"; id "86";
r3r1k1/1bnq1pbn/p2p2p1/1p1P3p/2p1PP1B/P1N2B1P/1PQN2P1/3RR1K1 w - - bm e5; c0 "e5=25, Nf1=10, b3=10, Kh1=5"; id "87";
8/4k3/p2p2p1/P1pPn2p/1pP1P2P/1P1NK1P1/8/8 w - - bm g4; c0 "g4=25"; id "88";
8/2P1P3/b1B2p2/1pPRp3/2k3P1/P4pK1/nP3p1p/N7 w - - bm e8=N; c0 "e8=N=75, c8=N=25"; id "89";
4K1k1/8/1p5p/1Pp3b1/8/1P3P2/P1B2P2/8 w - - bm f4; c0 "f4=100, Bd1=100"; id "90";
8/6p1/3k4/3p1p1p/p2K1P1P/4P1P1/P7/8 b - - bm g6 Kc6; c0 "g6=15, Kc6=15"; id "91";
r1b2rk1/ppp3p1/4p2p/4Qpq1/3P4/2PB4/PPK2PPP/R6R b - - am Qxg2; c0 "Qxg2=-25"; id "92";
2b1r3/r2ppN2/8/1p1p1k2/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nd6+; c0 "Nd6+=100"; id "93";
2k2Br1/p6b/Pq1r4/1p2p1b1/1Ppp2p1/Q1P3N1/5RPP/R3N1K1 b - - bm Rf6; c0 "Rf6=50"; id "94";
r2qk2r/ppp1b1pp/2n1p3/3pP1n1/3P2b1/2PB1NN1/PP4PP/R1BQK2R w KQkq - bm Nxg5; c0 "Nxg5=25"; id "95";
8/8/4p1Pk/1rp1K1p1/4P1P1/1nP2Q2/p2b1P2/8 w - - bm Kf6; c0 "Kf6=35"; id "96";
2k5/p7/Pp1p1b2/1P1P1p2/2P2P1p/3K3P/5B2/8 w - - bm c5; c0 "c5=10, Bg1=10, Be1=10, Be3=10"; id "97";
8/6pp/5k2/1p1r4/4R3/7P/5PP1/5K2 w - - am Ke2; c0 "Ke2=-25"; id "98";
3q1r1k/4RPp1/p6p/2pn4/2P5/1P6/P3Q2P/6K1 w - - bm Re8; c0 "Re8=35"; id "99";
rn2k2r/3pbppp/p3p3/8/Nq1Nn3/4B1P1/PP3P1P/R2Q1RK1 w k - bm Nf5; c0 "Nf5=35"; id "100";
r1b1kb1N/pppnq1pB/8/3p4/3P4/8/PPPK1nPP/RNB1R3 b q - bm Ne5; c0 "Ne5=25"; id "101";
N4rk1/pp1b1ppp/n3p1n1/3pP1Q1/1P1N4/8/1PP2PPP/q1B1KB1R b K - bm Nxb4; c0 "Nxb4=25"; id "102";
4k1br/1K1p1n1r/2p2pN1/P2p1N2/2P3pP/5B2/P2P4/8 w - - bm Kc8; c0 "Kc8=100"; id "103";
r1bqkb1r/ppp3pp/2np4/3N1p2/3pnB2/5N2/PPP1QPPP/2KR1B1R b kq - bm Ne7; c0 "Ne7=25"; id "104";
r3kb1r/pbqp1pp1/1pn1pn1p/8/3PP3/2PB1N2/3N1PPP/R1BQR1K1 w kq - bm e5 d5; c0 "e5=10, d5=10, Bb2=5, c4=5"; id "105";
r2r2k1/pq2bppp/1np1bN2/1p2B1P1/5Q2/P4P2/1PP4P/2KR1B1R b - - bm Bxf6; c0 "Bxf6=25"; id "106";
1r1r2k1/2pq3p/4p3/2Q1Pp2/1PNn1R2/P5P1/5P1P/4R2K b - - bm Rb5; c0 "Rb5=25"; id "107";
8/5p1p/3P1k2/p1P2n2/3rp3/1B6/P4R2/6K1 w - - bm Ba4; c0 "Ba4=25"; id "108";
2rbrnk1/1b3p2/p2pp3/1p4PQ/1PqBPP2/P1NR4/2P4P/5RK1 b - - bm Qxd4+; c0 "Qxd4+=25"; id "109";
4r1k1/1bq2r1p/p2p1np1/3Pppb1/P1P5/1N3P2/1R2B1PP/1Q1R2BK w - - bm c5; c0 "c5=25"; id "110";
8/8/8/8/4kp2/1R6/P2q1PPK/8 w - - bm a3; c0 "a3=25"; id "ERET 111";
Pity they don't work bug- free that way, strange enough they do up to 999msec/position and single threaded, but trying longer TC MEA starts working but stops after about half of the positions without storing the run.
Guess I should not combine bm- syntax with points for moves? On the other hand that worked as long as there were less positons collected and as said it does with not too long TCs, single threaded better than SMP.

Image

Could you have a look at the .epd above as for the syntax, please?

Here's the batch- file:
set MT=1000
set HASH=8
set THREADS=30
set PROTOCOL=uci
set EPD=epd\MEA111.epd
set MRL=mea_results

set EXE=engines\SF15.exe
set NAME="SF15 SMP"
mea.exe --engine %EXE% --name %NAME% --hash %HASH% --threads %THREADS% --protocol %PROTOCOL% --epd %EPD% --movetime %MT%
mrl %MRL% %MRL% %EPD% %MT%
Thanks again, regards
MEA is crashing on position 50.

r4r1k/p1p3bp/2pp2p1/4nb2/N1P4q/1P5P/PBNQ1PP1/R4RK1 b - - bm Nf3+; c0 "Nf3+=30, g5=5, Bxc2=5; id "50";

Change to:

r4r1k/p1p3bp/2pp2p1/4nb2/N1P4q/1P5P/PBNQ1PP1/R4RK1 b - - bm Nf3+; c0 "Nf3+=30, g5=5, Bxc2=5"; id "50";

Other than that use set HASH=128 or higher with 30 threads.

I get -

Code: Select all

    EPD  : epd\eret-111.epd
    Time : 100ms
                                                Max   Total   Time   Hash          
    Engine           Score   Found  Pos   ELO  Score   Rate    ms     Mb  Cpu      
 1  Rebel 15.1        1190     42   111  1629   3290  36.2%    100    64    4
 2  Koivisto 8.0      1080     39   111  1476   3290  32.8%    100    64    4
 3  Berserk 9          795     29   111  1089   3290  24.2%    100    64    4
 4  Seer 2.5.0         785     31   111  1075   3290  23.9%    100    64    4
 5  Toga 4             360     16   111   490   3290  10.9%    100    64    4
 6  Fruit 2.1          260     13   111   355   3290   7.9%    100    64    4

                                    Created with MEA
                                          by
                                       Ferdinand
                                         Mosca
So, all okay.

With crashing regards.
90% of coding is debugging, the other 10% is writing bugs.
Top
peter
Posts: 3384
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: STS re-re-re-re-re-visited

Post by peter »

Rebel wrote: ↑Sun Oct 30, 2022 10:20 pm r4r1k/p1p3bp/2pp2p1/4nb2/N1P4q/1P5P/PBNQ1PP1/R4RK1 b - - bm Nf3+; c0 "Nf3+=30, g5=5, Bxc2=5; id "50";

Change to:

r4r1k/p1p3bp/2pp2p1/4nb2/N1P4q/1P5P/PBNQ1PP1/R4RK1 b - - bm Nf3+; c0 "Nf3+=30, g5=5, Bxc2=5"; id "50";
So I really forgot a " one more time again
:!:
Couldn't believe, it yet did work several times before it started crashing then later on without me changing anything in .epd. Whenever I had such a syntax- error before, it didn't work at once at all till I had found the bug deleting one line after another
:)
Anyhow, it works now

Edit: only once with SF, second one try with LC0 crashed again
:(
Tried to restrict again to the 50 that ran without error message in console, yet no result stored. Window closes too fast to read the end of messages fully, but there's something written about margin time at the end.
Spooky, maybe it will get better after Halloween again
:twisted:
Peter.
Top
peter
Posts: 3384
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: STS re-re-re-re-re-visited

Post by peter »

peter wrote: ↑Sun Oct 30, 2022 10:47 pm but there's something written about margin time at the end.
Another one try with LC0 worked with 300msec, with 1000 it doesn't.
Peter.
Top
Ferdy
Posts: 4846
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: STS re-re-re-re-re-visited

Post by Ferdy »

We are doing an analysis of each position, to improve the suite. Currently I am evaluating positions to replace or not based from the current analysis. Position evaluation info is in google sheet sts_positions. The analysis of each position is in csv folder.

project folder: https://drive.google.com/drive/folders/ ... bTuSwMbmFT

Sample entries from sts_positions sheet.

Image

I am reading your feedback here.

If you want to contribute, I need a reviewer to "replace or not the current sts positions" Send your gmail by PM, and will give you a write access to sts_positions sheet. Typical procedure is setup the epd on the board, remember the old bm, and open/read the current analysis. The analyses are mostly searched to a depth of 40 with top 10 move evaluation.

* If there is a new bm based from the current analysis, and this new bm is different from the current theme (undermine, etc.) then put yes under replace column. This position can still be useful in other theme.
* If there is a new bm based from the current analysis, and this new bm is similar to the current theme (undermine, etc.) then do not replace the position.

This is just the first phase. Later for further improvement, we will review again based on the difficulty, evaluation distribution, etc.

If you encounter a nice position with a theme, similar or not to the current STS theme, post it here. We will expand the theme of current STS.

If you want to help on the analysis, send me your gmail will give you write access to the csv folder where you can upload the analysis file.

csv file format:

Code: Select all

index_<index number of google sheet sts_positions>_d<depth of analysis>_<your username>_sf15.csv
Example:
index_1_d26_ferdy_sf15.csv
Top

Return to “Computer Chess Club: General Topics”