Author: Uri Blass
Date: 11:39:40 12/16/00
Go up one level in this thread
On December 16, 2000 at 13:40:41, Enrique Irazoqui wrote: >On December 16, 2000 at 13:18:25, Uri Blass wrote: > >>On December 16, 2000 at 12:45:33, Ed Schröder wrote: >> >>>On December 16, 2000 at 11:18:04, Uri Blass wrote: >>> >>>>On December 16, 2000 at 11:00:12, Andreas St. wrote: >>>> >>>>>On December 16, 2000 at 10:11:31, Uri Blass wrote: >>>>> >>>>>>[D]4rrk1/5ppp/p2b4/1p1Q1b2/3P4/2P3Pq/PP1B1P1P/RN2R1K1 b - - 0 1 >>>>>> >>>>>>The problem is to find Bf4 >>>>>> >>>>>>I tried 2 programs Chessmaster6000(ss=10) and Gambittiger and both of them found >>>>>>Bf4 in less than an hour on my pIII450 >>>>>> >>>>>>Chessmaster6000 needed 18 minutes and 31 seconds when Gambittiger needed more >>>>>>time but less than an hour(I do not remember the exact time). >>>>>> >>>>>>Is there a program that is better than the old program chessmaster6000 in >>>>>>finding Bf4? >>>>>> >>>>>>I am interested in the results of Deep Fritz and Chessmaster8000 >>>>>> >>>>>>Uri >>>>> >>>>> >>>>>Hi, >>>>> >>>>>again i see as always default settings in Chessmaster 6000/7000 arent very good. >>>> >>>>ss=10 is not the default setting for chessmaster(ss=6 is the default). >>>> >>>>I know that tests did not prove significant difference between >>>>chessmaster6000(ss=10) and other personalities. >>>> >>>>It is possible that chessmaster6555 is better at test positions(I do not know) >>>>but there is no clear evidence which personality is better in games. >>>> >>>>Uri >>> >>>I am sure I am missing something. Bf4 looks like a horrible positional >>>blunder (1..Bxf4). What is the point of Bf4? >>> >>>Ed >> >>You missed the fact that black is the side to move and not white. >> >>Uri > >I don't think it's a convincing test position. The programs I tried evaluate Bf4 >0.1 or 0.2 better than next best. Gambit Tiger says -0.30 for Bf4 and -0.48 for >Rd8. Century 3 also shows tiny differences. > >What's the complete line? After Bf4 RxR RxR Qg2 black looks better to me, but so >it did from the start position. > >Enrique I do not have a winning line for black after Bf4 RxR RxR Qg2 I posted this position as test position because of the fact that yoav dothan posted a correspondence game when this position happened and he claimed that even modern chess programs cannot predict this move. He claimed that it demonstrates the advantage of humans relative to computers. I do not agree with him and it was proved that at least 2 programs found Bf4. The game that was published from the correspondence olympiad in 1960 proves the advantage of computers relative to humans because white probably could not find the best defence after 18...Bf4 and prefered 19.Qg2 that is the second best move based on chess programs. 19.Qg2 Rxe1+ 20.Bxe1 Bc1 21.Qxh3 Bxh3 22.Nd2 Bxb2 23.Rb1(I think that 23.Rd1 is better) Bxc3 24.Nb3 Rc8 25.f3?(25.Bxc3 or 25.Rc1 were better) Bf5 26.Rc1 Bxd4+ 27.Bf2 Rxc1+ 28.Nxc1 Bb2 0-1 was the human-human game that was published in the chess newspaper in Israel. I think that I can give better hard test positions based on my games. Here is one that could happen in my correspondence game. The target is to find Bh8 [D]rqr3k1/pp1bppbp/2np1npB/8/2BNP1P1/2N2P2/PPPQ4/2KR3R b - - 0 1 The alternatives are bad. 14...Nxd4 15.Bxg7 is winning for white and white has also a clear advantage after 14...Nxe4 15.Nxe4 Bxd4 16.Qg5(for example 16...Bh8 17.Rxd6 Qxd6 18.Nxd6 exd6 19.Qh4) I remember that Gandalf needed a long time to find 14...Bh8(more than an hour on my PIII450 if my memory is correct). 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.