Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hard4 test set analysis by Rybka at 720 seconds per position

Author: Vincent Lejeune

Date: 05:33:28 01/13/06

Go up one level in this thread


Hello Dann, thanks to run this test suite with 12 min/moves, it will certainly
help de debug this quite hard positions.
I've some remarks about some positions and other about Rybka :

1) 8/5p1p/8/6k1/8/6P1/5PP1/7K w - - bm Kh2
Here engine have to find the main line 1. Kh2 f6 2. Kh1! f5 3. Kg1 Kf6 4. Kf1
Ke5 5. Ke1 Kd5 6. Kd1 Ke5 7.Kc2! Kd4 8. Kd2 Ke4 9. Ke2 h6 10. Kf1! , or at least
the first 5-6 moves or the position is not consider as solved.

10) 4k1rr/1R1R3p/8/8/8/2K5/7P/8 w k - bm Rd2;
Rybka still have big troubles to solve zugzwang theme.

12) 8/2B2pPb/7k/4K3/3p2P1/8/8/8 w - - bm g8=N+;
Do you use the first Rybka release who doesn't support under promotion ? ;-)

13) 2r1r3/p3bk1p/1pnqpppB/3n4/3P2Q1/PB3N2/1P3PPP/3RR1K1 w - - bm Rxe6;
I will probably remove this one, it's not clear for me that Rxe6 is clearly
better than Qh3, it's far deep an too complicated to have a final evaluation, I
think.

14) 8/8/p7/n1p5/PpPp4/1P1P4/2K1k1B1/8 b - - bm Nxb3
Only +1.12 but Rybka seems to have understood the position

15)3r2k1/pp3ppp/5q2/8/1n2Q3/7P/PPP2PK1/R1B5 w - - am Qxb4??;
I will remove this one too, Qxb4 is a clear draw but it's not clear white have
better than equal here ...

22) 8/3k3p/1p4p1/p2p1p2/PP1K1P1P/2P3P1/8/8 b - - bm b5!!
b5 is easy to find but the winning line is very very long

24) 4q1kr/p6p/1prQPppB/4n3/4P3/2P5/PP2B2P/R5K1 w - - 1. Qxe5 fxe5 2. Rf1 *
This one will be in the impossible category (or solvable in 2015 by comp :-) )
of the final set.

31) 1knB4/8/7p/5R2/7K/8/3r4/7B w - - bm Rb5+;
Rybka doesn't understand this one (1.Rf5b5+ Nc8b6 2.Rb5xb6+ is bad)
The main line is 1. Rb5+ Nb6 2. Bxb6 Rh2+ 3. Kg4!!(Rybka have trouble to find
this one) Rxh1 4. Bg1+!!(the rook is trapped) Kc7 5. Rb1 Kd7 6. Re1!!(a sequence
of zugzwang to but the black king on a black square!) Kc6 7. Rd1 Kb7 8. Rc1 Ka6
9. Rb1!(it's the end the rook is lost) Ka5 10. Bb6+ Ka6 11. Rxh1 Kxb6 *

33) 3k4/2Rnn3/1p2pbp1/pP1p1pBp/P2P1P1P/4KPP1/4B3/8 w - - bm Rxd7+; c0 "draw !?";
Rxd7+ is a move to avoid and not a best move, after Rxd7+ white cannot win
anymore.

34) r7/6k1/5R2/p3p3/Pp1pPp2/1PpP1Pp1/2P3P1/1K6 w - - bm Rh6; c0 "stalemate
theme";
Great to see thie one solved ! Congrats to Vasik !

38) 8/2b2k1K/1pPp1p2/1P1P1P2/5B2/8/8/8 w - - bm Be3; id "The Pawnbishop";
May be Be3 is not the only solution more general is to find the sacrifice Bxg6
(after 2 white Bishop moves) in the line and white king go to c7 take the pawn
and bishop and go queening.


In the other hand, the Hard5 is already in a good way to go out ... ;-)
I hope to make shortly a beta version of the "CCC-hard-positions-set-2006" with
latest refutations.

Thanks all !

On January 12, 2006 at 14:22:26, Dann Corbit wrote:

8/5p1p/8/6k1/8/6P1/5PP1/7K w - - bm Kh2;
8/5k2/p7/2pP1Pp1/2P3P1/2N3K1/P4N2/5q2 b - - bm Kf8;
8/1p6/3k4/4n3/p2KP2p/P6N/1P6/8 b - - bm b6;
8/3R1P2/1ppP1p2/3r4/8/K7/p4k2/8 w - - am Kxa2; bm Kb2;
8/2k5/Kp3pp1/7p/P4P2/8/6PP/8 b - - am Kc6; bm h4;
r4rk1/1b1n1pb1/3p2p1/1p1Pq1Bp/2p1P3/2P2RNP/1PBQ2P1/5R1K w - - bm Nf5 Nxh5;
r3n3/3b2kp/1qNp2p1/1p1Pp2n/1P2Pp2/3B1N1P/1Q3PP1/2R3K1 w - - bm Rc5;
r1bq1rk1/1p3ppp/p1pp2n1/3N3Q/B1PPR2b/8/PP3PPP/R1B3K1 w - - bm Rxh4;
3rrk2/2p1n1bQ/1p3pp1/1q2B1N1/p1b1PP2/6PB/P1P5/3R1RK1 w - - bm Be6!;
4k1rr/1R1R3p/8/8/8/2K5/7P/8 w k - bm Rd2;
1q6/k7/3p2nP/pp1Qp3/1P2B3/P7/2P5/1K6 w - - bm h7!; dm 9;
8/2B2pPb/7k/4K3/3p2P1/8/8/8 w - - bm g8=N+;
2r1r3/p3bk1p/1pnqpppB/3n4/3P2Q1/PB3N2/1P3PPP/3RR1K1 w - - bm Rxe6;
r2qr1k1/pbp1bpp1/1p2p2p/3n4/3P1PN1/2PB3R/PP1BQ1PP/R5K1 w - - bm Nxh6!;
2r3k1/2r1ppb1/3p2p1/pq4Pn/1p1BP3/1B3P2/PPPQ4/1K1R3R w - - bm Rxh5;
8/8/p7/n1p5/PpPp4/1P1P4/2K1k1B1/8 b - - bm Nxb3; c0 "score should have >+2";
r1q1rbk1/2pp1ppp/p1b5/1p1NP3/3Q4/3R2B1/PPP2PPP/5RK1 w - - bm Nf6+!;
1r2kb1r/2qb1pp1/p7/3Q2Pn/8/4B2P/PPP3B1/R4RK1 w k - bm Rad1!;
3b4/3B4/4n3/5kp1/8/6P1/6K1/8 b - - am g4??;
rn1qk2r/p2nbp2/1p2p1p1/2ppP1Np/3P1PP1/P1N1B2Q/1PP4P/R3KR2 w Qkq - bm Nxf7!;
3r2k1/pp3ppp/5q2/8/1n2Q3/7P/PPP2PK1/R1B5 w - - am Qxb4??;
8/3k3p/1p4p1/p2p1p2/PP1K1P1P/2P3P1/8/8 b - - bm b5!!; c0 "with >+2 or +3 for
black";
1r1rb1k1/5ppp/4p3/1p1p3P/1q2P2Q/pN3P2/PPP4P/1K1R2R1 w - - bm Rxg7+; c0 "Vescovi
G.,2455 - Gschwendtner S.,2270, Groningen (open) 1994"; id "Aemis Test Suite
No.30"; pv 1. Rxg7+ Kxg7 2. h6+ Kf8 3. Qf6 Bc6 4. c3;
4q1kr/p6p/1prQPppB/4n3/4P3/2P5/PP2B2P/R5K1 w - - bm Qxe5; pv 1. Qxe5 fxe5 2.
Rf1;
8/6p1/P1b1pp2/2p1p3/1k4P1/3PP3/1PK5/5B2 w - - 0 1 bm Bg2; id "Bishop Rivalry";
4B3/5K2/5P1b/2p5/2p5/1k6/8/8 w - - bm Kg6; id "Merry-go-round";
8/1p5k/4pP2/6B1/8/1P6/8/1n4K1 w - - bm f7; id "Good bishop vs Limited knight";
8/p1p2pp1/1pP3p1/3Pb2k/4P3/5P1P/8/3K2B1 w - - bm f4; id "The Winning Arrow";
8/6r1/Pp5p/7k/p7/7K/6P1/R7 w - - bm Ra3; id "The Rebel Rook";
1k3b1q/pP2p1p1/P1K1P1Pp/7P/2B5/8/8/8 w - - bm Kd5; id "Totally Paralysed";
1knB4/8/7p/5R2/7K/8/3r4/7B w - - bm Rb5+; id "The Weekend Nicholas";
4kr2/5p1K/3p1Q2/1p4P1/4P3/1PP5/7b/8 w - - bm Kh6; id "The Giant King"; pv 1. Kh6
Be5 2. Kg7!! Bh2 3. c4! bxc4 4. e5!! Bxe5 5. bxc4 Bxf6+ 6. gxf6 Rh8 7. Kxh8 Kd7
8. Kg8 Ke6 9. Kg7;
3k4/2Rnn3/1p2pbp1/pP1p1pBp/P2P1P1P/4KPP1/4B3/8 w - - bm Rxd7+; c0 "draw !?";
8/1p6/p4p1p/2p1P2k/5P2/6pP/1P4P1/6K1 w - - bm f5; id "Rare Pawn ending"; pv 1.
f5 fxe5 2. h4 c4 3. Kf1 a5 4. Ke2 a4 5. Kf3 c3 6. Kxg3 cxb2 7. Kh3 b1=Q 8. g4+;
r7/6k1/5R2/p3p3/Pp1pPp2/1PpP1Pp1/2P3P1/1K6 w - - bm Rh6; c0 "stalemate threat";
8/8/pp6/3b2K1/1P6/PN2k3/8/8 w - - bm Na5; id "A-pawn and wrong bishop";
k4B1b/p6P/Kp6/1P3P2/4P3/R2P4/2P2p2/6n1 w - - bm Ra1; id "The Arrow's Point"; pv
1. Ra1! Bxa1 2. c3! Bxc3 3. d4! Bxd4 4. e5! Bxe5 5. f6! Bxf6 6. Bg7 Bxg7 7.
h8=Q+ Bxh8;
8/2b2k1K/1pPp1p2/1P1P1P2/5B2/8/8/8 w - - bm Be3; id "The Pawnbishop";
8/5p2/6PR/7P/1k6/p7/7r/1K6 w - - bm g7; id "The Original Mozes";



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.