Author: Joshua Lee
Date: 05:25:41 06/07/01
Go up one level in this thread
>>>From a recently completed game in the 24th world correspondence chess >>>championship semi-finals: >>> >>>[D] r1qr2k1/pb2bpp1/1p2pn1p/2p4P/3P1B2/2PB1N2/PP2QPP1/1K1R3R w - - >>> >>>Here the move 15.Rhg1 is very strong, as it gets the rook off the long diagonal >>>and threatens to charge the g-pawn down the board. But computers don't often >>>like to give up material .... in this case the h pawn and after 15.Rhg1! Nxh5 >>>16.Bxh6! also the bishop. Can any program find 15.Rhg1? I believe it may win >>>by force, and if not it certainly leaves black in big trouble, but all programs >>>I tested don't find it, even over night. If that is too hard, try the position >>>after 15.Rhg1 Nxh5. >> >>Yace (which is especially good at difficult tactical positions) was unable to >>solve after 4 hours: > >I don't think 4 hours will be near enough time. You could try the position >after the main line of 15.Rhg1 Nxh5 15.Bxh6 gh and see how long it takes for >Yace to find a + score for white. It should then steadily climb for as long as >you let it crunch on the position. > >Robin Smith Here is what fritzy has with 10 hours: Fritz 6 - Lee,J r1qr2k1/pb2bpp1/1p2pn1p/2p4P/3P1B2/2PB1N2/PP2QPP1/1K1R3R w - - 0 1 Analysis by Fritz 6: 1.Bxh6-- ³ (-0.28) Depth: 1/4 00:00:00 1.Bxh6-- gxh6 -+ (-1.91) Depth: 1/5 00:00:00 1.dxc5! -+ (-1.53) Depth: 1/5 00:00:00 1.dxc5! Qxc5 = (0.00) Depth: 1/5 00:00:00 1.Be5! = (-0.03) Depth: 1/9 00:00:00 1.Be5! cxd4 2.Bxd4 = (0.00) Depth: 1/9 00:00:00 1.Be5 cxd4 2.Bxd4 = (0.00) Depth: 2/9 00:00:00 1.Be5 cxd4 2.Bxd4 Qc6 = (-0.12) Depth: 3/14 00:00:00 1kN 1.dxc5! = (-0.09) Depth: 3/14 00:00:00 3kN 1.dxc5 Qxc5 2.Qe3 Qa5 3.Bc7 = (-0.12) Depth: 4/15 00:00:00 6kN 1.Be5! = (-0.09) Depth: 4/15 00:00:00 7kN 1.Be5 cxd4 2.Bxd4 Qc6 3.Bb5 Qc7 = (0.00) Depth: 5/16 00:00:00 24kN 1.Be5 cxd4 2.Bxd4 Qc6 3.Bb5 Qc7 4.Qe3 Be4+ 5.Bd3 Bxd3+ = (-0.19) Depth: 6/20 00:00:00 62kN 1.Be5 cxd4 2.Bxd4 Qc7 3.Qe5 Qxe5 4.Bxe5 Bxf3 5.gxf3 Rac8 = (-0.25) Depth: 7/25 00:00:00 184kN 1.Be5 cxd4 2.Bxd4 Qc6 3.Rh4 Rac8 4.Bxf6 Bxf6 5.Rc4 Qd7 = (-0.22) Depth: 8/23 00:00:00 408kN 1.Be5 cxd4 2.Bxd4 Qc7 3.Be5 Qc5 4.Bd4 = (-0.19) Depth: 9/28 00:00:02 1445kN 1.Be5 cxd4 2.Bxd4 Qc7 3.Be5 Qc5 ³ (-0.28) Depth: 10/30 00:00:04 2566kN 1.Be5 cxd4 2.Nxd4 Bd5 3.Rhe1 Qc5 4.g4 Nd7 5.Bf4 Rac8 = (-0.22) Depth: 11/31 00:00:18 10792kN 1.Be5 cxd4 2.Nxd4 Nd7 3.f4 Qc5 4.Bc2 Bf6 5.Rhe1 Nxe5 = (-0.22) Depth: 12/34 00:00:56 33141kN 1.Be5 cxd4 2.Nxd4 Nd7 3.f4 Qc5 4.Rhf1 Rac8 5.g4 Bd5 = (-0.25) Depth: 13/33 00:01:59 72471kN 1.Be5 cxd4 2.Nxd4 Nd7 3.f4 Qc5 4.Rhf1 Bd5 5.Be4 Bxe4+ = (-0.25) Depth: 14/35 00:04:38 173172kN 1.Be5 cxd4 2.Nxd4 Nd7 3.f4 Nxe5 4.fxe5 Bc5 5.Be4 a6 = (-0.19) Depth: 15/40 00:21:34 806654kN 1.Be5 Ng4 2.Bg3 Nf6 3.Be5 = (0.00) Depth: 16/44 01:42:21 3839610kN 1.Be5 cxd4 2.Nxd4 Nd7 3.Bg3 Nc5 4.Bc2 Bf6 5.Rhg1 a6 = (-0.19) Depth: 17/44 02:26:52 5553718kN 1.Be5 a6 2.dxc5 bxc5 3.Rhg1 Bd5 4.c4 Bc6 5.g4 Qb7 = (-0.19) Depth: 18/48 10:21:32 23614941kN (Lee, Pensacola,Fl 07.06.2001) I claim that it wins in all variations so does fritz but this one is either solved at this ply or the next it is really deep and may take up to 30 or 90 hours on my computer 800Mhz , i am using 256MB for hash tables i don't know if 286 or 384MB would be any better nobody responded to a post i had about that. All i know is that the f5 manual suggests powers of two which would be 256 and next 2 ^9 512 2^10 1024 2^11 2048 2^12 4096 i think 1GB would be better than 512MB and 2GB but 4GB would be better than all the others i don't have much proof but at 384MB for hash table this problem and others doesn't seem to make a difference. I will keep you updated as to when and if this gets solved.
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.