Author: Uri Blass
Date: 14:15:39 07/14/04
Go up one level in this thread
On July 14, 2004 at 16:35:30, Will Singleton wrote: >On July 14, 2004 at 05:11:07, Uri Blass wrote: > >>I posted yesterday some test positions and it seems that people missed them >> >>see http://www.talkchess.com/forums/1/message.html?376470 >> >>Uri > >It looks to me like all positions have problems with one exception. So the test >doesn't look that useful. Or maybe I'm wrong (see comments). > >r4bk1/1b4p1/p3p3/q3n1B1/Np1NnQ2/1P3p1R/P1P5/2KR4 w - - bm Nxe6; " too hard "; Falcon found it during the game so it is not too hard Movei needs to improve to find it faster and cannot find it in a reasonable time: Analysis by movei on A3000(that A3000 crashed in bar-ilan again and again but work for more than 48 hours without a single crash here) Analysis so far by movei: depth=12 -0.83 f4h2 e5f7 g5e3 f8d6 h2g1 a5e5 c1b1 e4g5 h3h5 d6e7 c2c4 a8f8 Nodes: 180228316 NPS: 576767 Time: 00:05:12.48 depth=13 -0.70 f4h2 e5f7 g5f4 e4c3 f4c7 a5g5 c7f4 g5e7 a4c3 b4c3 c1b1 e6e5 d4f5 e7a3 Nodes: 244101559 NPS: 570997 Time: 00:07:07.50 depth=13 -0.70 f4h2 e5f7 g5f4 e4c3 f4c7 a5g5 c7f4 g5e7 a4c3 b4c3 c1b1 e6e5 d4f5 e7a3 Nodes: 324662522 NPS: 578545 Time: 00:09:21.17 depth=14 -0.97 f4h2 e5f7 g5f4 e4c3 d1h1 f3f2 h3f3 b7e4 h2f2 c3a4 b3a4 f8c5 f2e2 e4f3 e2f3 Nodes: 422498103 NPS: 575101 Time: 00:12:14.65 >r4bkR/1b4p1/p3N1n1/q5B1/Np2nQ2/1P3p2/P1P5/2KR4 b - - bm Nxh8; " too hard "; Shredder found it during the game and I checked that movei could find it in few minutes so it is not too hard. >r4bk1/1b1R2p1/p3NBn1/q7/Np1Q4/1P3p2/P1P5/2K5 b - - bm Qf5; " too easy "; I checked that it is not too easy for movei and first it prefered a different move. I consider too easy only positions that movei can solve in less than 2 seconds on fast hardware and it was not one of them. >2r3k1/p4rpp/4bp2/4p3/QR2P3/P2qB1P1/5P1P/5RK1 b - - bm a5; " after a5 Rb6, what >then? "; Bd7 and black wins a pawn. >2r3k1/1Q1b1rpp/5p2/4p3/8/P3BqP1/5P1P/1R3RK1 b - - bm e4; " too easy "; Again latest movei needs 5 seconds to fail high on fast hardware. I hope to do it too easy for movei in the future. >2r4k/r5q1/3Pp3/1R4p1/1Pp1Q2p/7P/4B1PK/8 b - - am c3; " too easy "; Cannot be too easy because movei did that mistake in WCCC Another ply was enough to avoid it. >3r4/3P1k2/8/1PR2BK1/6P1/2p5/1r6/8 b - - am Rb3; " bm c2 I don't get why c2 is >better than Ke7? "; You may be right and I will check it later. >1r6/4kpp1/b1pp3p/q1n1pP2/Pr2P3/1BN3Q1/P1P3PP/K2RR3 w - - am Qxg7; " bm Re3 this >is good "; >r3r1k1/p4ppp/b2p4/B1p5/3bPPn1/3P4/PP4PP/2RNNR1K b - - bm f5; " some reason not >to play Rxe4? "; Rxe4 lead to equality and it seems that f5 leads to advantage for black. Movei found it after some hours of analysis later. Deep Sjeng expected f5 and failed low. >r3r3/1p3pk1/p1p2p2/2Pp3p/1P4bP/P1KBP1P1/2R2P2/6R1 b - - am f5; " i don't get >this one "; Movei played it in the first round against Sos. It helps to activate the king but it block the bishop and put it out of game so I believe that it is a bad move. Uri
This page took 0.02 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.