Author: Uri Blass
Date: 14:53:54 02/02/02
Go up one level in this thread
On February 02, 2002 at 14:51:32, Oliver Roese wrote: >On February 02, 2002 at 13:48:28, Otello Gnaramori wrote: > >>On February 02, 2002 at 10:36:24, Oliver Roese wrote: >> >>>I knew that a good human (> 2350 ELO) can easily calculate 20 moves (=40 plies) >>>and more within some seconds in some pawnendings. Recently i posted such a >>>position here. Such a depth is out of reach for moderate current hardware. >> >>The above statement doesn't convince me ... >>Don't get me wrong, but the calculus is the computer strongest weapon. >> >>w.b.r. >>Otello > >This is the position i mentioned. Load the fen-string into an engine >and see for yourself. >http://www.it.ro/ccc_search/ccc.php?find_thread=113841 > >Oliver Roese It is impossible to calculate 40 ply forward in few seconds in that endgame The guy was simply lucky to be right and in another position he could be wrong after few seconds. human may suspect that Qxc4 is a draw but it is not responsible to decide that it is a draw in a few seconds. There are also programs that have no problem to find Qxc4 Saying that programs cannot find something based on an old post from 2000 is not convincing. software and hardware got better since 2000. Junior7 is one of the worst programs in pawn endgames but it has no problem to find Qxc4 in less than 1 second even on p200. New position [D]8/4k1pp/5p2/P1p1p3/2Qq4/1P4P1/5PKP/8 b - - 0 1 Analysis by Junior 7: 1...e4 2.a6 Qxc4 ² (0.50) Depth: 3 00:00:00 1...h5 2.a6 Qxc4 ² (0.48) Depth: 3 00:00:00 1...Kd6 2.a6 Qxc4 ² (0.42) Depth: 3 00:00:00 1...Kd6 2.Qa6+ Kc7 3.Qa7+ Kd6 4.Qxg7 Qe4+ 5.f3 ± (0.86) Depth: 6 00:00:00 3kN 1...Qxc4! = (0.12) Depth: 6 00:00:00 3kN 1...Qxc4 2.bxc4 Kd6 3.Kf3 Kc7 4.Ke4 Kd6 5.a6 ² (0.45) Depth: 9 00:00:01 65kN 1...Qxc4 2.bxc4 Kd7 3.Kf3 f5 4.h4 Kc6 5.h5 Kb7 ² (0.38) Depth: 12 00:00:03 347kN Uri
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.