New Tactics and Endgames EPD collections

[ivoryknight]

Wonderful! Thanks for sharing!

In the tactics file, the toughest 10% seem to be:

r3kb1r/pp3ppp/2n5/q3p3/5B2/2N4P/P1P1QPP1/3R1RK1 w kq - 0 1 bm Rd5;
r3qrk1/1b3p2/1p1npnp1/2b1N1N1/p1P5/P1B5/2B1QP1P/R2R2K1 w - - 0 1 bm Ng4;
3r1rk1/p1q2ppp/3bb3/3pB3/1p6/2P2Q1P/PPB2PP1/3RR1K1 w - - 0 1 bm Bxh7+;
2rq1rk1/pp3ppp/3p4/6P1/b3pP2/1N4Q1/PPP3P1/2KR3R w - - 0 1 bm Rxh7;
r1b1qrk1/pp1p1ppp/8/n3p1BQ/2B1P3/8/P1P2PPP/3R1RK1 w - - 0 1 bm Bf6;
r1bq1rk1/ppp3p1/7p/3P2n1/2PQ1p2/1N5P/PPP2PPK/R1B2R2 b - - 0 1 bm Nf3+;
r3r1k1/pbq1ppbp/1pp2np1/4N3/3P4/2P5/PPB1QPPP/R1B1R1K1 w - - 0 1 bm Nxf7;
r1br2k1/p1q2pp1/4p1np/2ppP2Q/2n5/2PB1N2/2P2PPP/R1B1R1K1 w - - 0 1 bm Bxh6;
r2qr3/2p1b1pk/p5pp/1p2p3/nP2P1P1/1BP2RP1/P3QPK1/R1B5 w - - 0 1 bm Bxh6;
1q1r3k/3P1pp1/ppBR1n1p/4Q2P/P4P2/8/5PK1/8 w - - 0 1 bm Rxf6;
rqbn1rk1/1p3ppp/p3p3/8/4NP2/8/PPP1BQPP/1K1R3R w - - 0 1 bm Nf6+;
r1b1kb1r/1p1n1ppp/p2ppn2/6BB/2qNP3/2N5/PPP2PPP/R2Q1RK1 w kq - 0 1 bm Nxe6;
rqb1k2r/1p1nbp1p/p4pp1/8/1PBN1P2/P1N1P3/7P/2RQ1RK1 w kq - 0 1 bm Bxf7+;
r1b2rk1/2q1bppp/p1n1p3/1p1np1P1/5P2/1NNBBQ2/PPP4P/R4RK1 w - - 0 1 bm Bxh7+;
r3r1k1/p3bppp/q1b2n2/5Q2/1p1B4/1BNR4/PPP3PP/2K2R2 w - - 0 1 bm Rg3;
r2qrnk1/4bppp/b1p5/1p1p2P1/p2P1N1P/2NBP3/PPQ2P2/2K3RR w - - 0 1 bm Bxh7+;
2rq1rk1/pp1brpbp/4p1p1/3pP1N1/3P2Q1/2PB4/P4PPP/R1B1R1K1 w - - 0 1 bm Nxh7;
2r3rk/5p1p/3pb3/pP1Np3/P1BNPp1q/3Q1P2/6PP/R5K1 b - - 0 1 bm Rxg2+;
rn1q1rk1/pb1p1ppp/4p3/2pnP3/1bp5/2N2N2/PP3PPP/RBBQK2R w KQ - 0 1 bm Bxh7+;
3r2k1/pb1rqpb1/1pp1p1p1/4P1Np/3P2nP/P7/BPQ1NPP1/3R1RK1 w - - 0 1 bm Bxe6;
r4rk1/p1p2pp1/bp1p1q1p/n1nPpN2/2P1P2P/2PB4/P2BQPP1/R3K2R w KQ - 0 1 bm Bg5;
2r4r/1q1kb1p1/4p2p/1n1pP3/1p6/3RBP2/NPP2Q1P/1K2R3 b - - 0 1 bm Qa6;
r1b3k1/3nqp1p/p1n1p1pP/3pP1N1/Pp3QN1/1Pr3P1/2PR1PK1/4RB2 w - - 0 1 bm Rxd5;
r1r3kb/1pqbpp2/p2p1npB/n7/3NP3/1BN2P2/PPPQ2P1/2KR3R w - - 0 1 bm Bf8;
r4rk1/pp1n1ppp/3qp3/3nN1P1/b2P4/P2B1Q2/3B1P1P/1R2R1K1 w - - 0 1 bm Bxh7+;

[Norm Pollock]

With the new links, you now have duplicates at lines 211 and 241 in ah_endgames-250.epd.

[Aser Huerga]

With the new links, you now have duplicates at lines 211 and 241 in ah_endgames-250.epd.

OMG, I'm a scatterbrain!

The links above are right (thanks Fern), except line 241, which have to be change with:

4k3/5p1p/1p5P/6P1/P7/8/8/4K3 w - - 0 1 bm Kd2;

Thank you very much Norm.

I hope there will be no more changes.

[ivoryknight]

Houdini 2b Pro 6CPU finds the answer to all of the *tactics* problems within 4 seconds except for these two:

1q1r3k/3P1pp1/ppBR1n1p/4Q2P/P4P2/8/5PK1/8 w - - 0 1 bm Rxf6;
r3r1k1/pbq1ppbp/1pp2np1/4N3/3P4/2P5/PPB1QPPP/R1B1R1K1 w - - 0 1 bm Nxf7;

[Aser Huerga]

Houdini 2b Pro 6CPU finds the answer to all of the *tactics* problems within 4 seconds except for these two:

1q1r3k/3P1pp1/ppBR1n1p/4Q2P/P4P2/8/5PK1/8 w - - 0 1 bm Rxf6;
r3r1k1/pbq1ppbp/1pp2np1/4N3/3P4/2P5/PPB1QPPP/R1B1R1K1 w - - 0 1 bm Nxf7;

Thanks for the test Brent. As I said in the initial post, the difficulty is choosed having in mind 1 core, medium hardware, and wide ELO strength engines, but any test is welcome.

Cheers!

[ivoryknight]

Aser,

The answers to these two problems in the endgame EPD are impossible, assuming I've copied everything correctly:

8/8/5K2/Pk2N3/2R5/2n5/4p3/n7 w - - 0 1 bm Kb4;
7K/7p/1P6/8/6rB/8/4k3/5N2 w - - 0 1 bm Nf3;

[ivoryknight]

Houdini 2b Pro 6CPU finds the answer to all of the *tactics* problems within 4 seconds except for these two:

1q1r3k/3P1pp1/ppBR1n1p/4Q2P/P4P2/8/5PK1/8 w - - 0 1 bm Rxf6;
r3r1k1/pbq1ppbp/1pp2np1/4N3/3P4/2P5/PPB1QPPP/R1B1R1K1 w - - 0 1 bm Nxf7;

Thanks for the test Brent. As I said in the initial post, the difficulty is choosed having in mind 1 core, medium hardware, and wide ELO strength engines, but any test is welcome.

Cheers!

That's how I first started testing these problems, on my old, slow laptop: 2 CPUs @ 1.6GHz. In 4 seconds, Houdini 1.5a missed 37, Rybka missed 64, Critter missed 52, and Komodo I quit testing after it missed half of them by problem 128 (unfair as it only uses 1 CPU).

[jdart]

IMO to validate test positions you really need to run them in multi-variation mode, because they are no good if there a 2nd or 3rd solution that is close in score to the supposed "best" move.

[Aser Huerga]

Aser,

The answers to these two problems in the endgame EPD are impossible, assuming I've copied everything correctly:

8/8/5K2/Pk2N3/2R5/2n5/4p3/n7 w - - 0 1 bm Kb4;
7K/7p/1P6/8/6rB/8/4k3/5N2 w - - 0 1 bm Nf3;

It has to be:

8/8/5K2/Pk2N3/2R5/2n5/4p3/n7 w - - 0 1 bm Rd4; (pos 132)
7K/7p/1P6/8/6rB/8/4k3/5N2 w - - 0 1 bm Ne3; (pos 181)

I'm a bit ashamed, I was think all was checked, sorry.

[ivoryknight]

Aser,

The answers to these two problems in the endgame EPD are impossible, assuming I've copied everything correctly:

8/8/5K2/Pk2N3/2R5/2n5/4p3/n7 w - - 0 1 bm Kb4;
7K/7p/1P6/8/6rB/8/4k3/5N2 w - - 0 1 bm Nf3;

It has to be:

8/8/5K2/Pk2N3/2R5/2n5/4p3/n7 w - - 0 1 bm Rb4;
7K/7p/1P6/8/6rB/8/4k3/5N2 w - - 0 1 bm Ne3;

I'm a bit ashamed, I think all was checked, sorry.

I'm sorry you feel ashamed. From my perspective, I think you've done *terrific* work here! Engine creators can use what you've put together for testing and improvement. 2 typos out of 250 problems--that's less than 1% error. And the bigger picture is all you've done and gotten right. Keep up the good work, friend!