Author: blass uri
Date: 10:44:13 02/10/99
Go up one level in this thread
On February 10, 1999 at 13:17:45, Bruce Moreland wrote: > >On February 09, 1999 at 05:36:27, blass uri wrote: > >> >>On February 09, 1999 at 01:58:51, blass uri wrote: >> >>> >>>On February 08, 1999 at 17:20:31, Albert Silver wrote: >>> >>>>Hi, >>>> >>>>I thought I'd submit the following combination from Marcio Miranda-GM Felix >>>>Izeta, Vince Toth Memorial, 6th round. >>>> >>>>6rk/3b1p2/1p1p3p/1Nb1p1r1/2P1Pp2/P2B2P1/1P3PKP/R4R2 b - - 0 26 >>>> >>>>The winning move is 26...Bh3+!! How long does it take for your computer to play >>>>this and how long before it gives Black the advantage with this move. >>>> >>>>Just for your info: Black has two mating patterns here: 26...Bh3+ 27.Kxh3 f3 >>>>28.Kh4? d5!! with Be7 and mate to follow, or the main line I have goes 26...Bh3+ >>>>27.Kxh3 f3 28.g4! Rxg4 29.Nc3!! (best defense I've seen so far as Black is >>>>threatening a second mating pattern with R8g5, Rf4, and Rh5 mate) R8g5 30.Ne2 >>>>fxe2 31.Bxe2 Rxe4 and Black is winning. >>> >>>I gave Junior5 3 options to analyze in the position after Bh3+ and at depth 21 >>>the result is Kf3 -0.17, Kxh3 -1.15,Kh1 -1.22 >>> >>>The main line after 27.Kf3 is Bxf1 28.Rxf1fxg3 29.hxg3 Ra8 >>> >>>Uri >>Fritz5 sees 27.Kf3 Bg4+ 28.Kg2 Bh3+ >>I am not sure if Bh3+ gives black more than a draw. >> >>Junior5 can find 27.Kf3 in less than 54 minutes in my pentium200MMX computer. >> >>The commercial Junior5 found it after looking at 415301 Knodes. >>The latest Junior is a little faster in this position and needs only 406731 >>Knodes. >> >>It is also faster at the Rxd5+ position and needs 90128 Knodes to find Rxd5 with >>a positive evaluation when the commercial Junior5 needs 107974 Knodes. >> >>I do not want to give exact times because I did other applications at the same >>time and maybe it slowed down the computer. > >Superficially it seems that Bh3+ Kxh3 leads to a game where black is up a pawn >and white's kind of cracked up. I'd be confident if my program achieved that >position. > >After Bh3+ Kf3 Bxf1 Rxf1 I'd be doing a lot better I think. > >Mine sees Bh3+ at about three minutes, expecting Kxh3 in response, but expects >Kf3 in response at the next ply (15 minutes or so), all times on a P2/300. > >bruce To be exact about Fritz5(16 bit) Fritz5(16 bit) evaluated the position after Kf3 as a draw but it did not find Kf3 in a reasonable time(It saw Kxh3 is an advantage for black after 20 minutes or so but after more than 2 hours on pentium200MMX it still wanted to play Kxh3 with evaluation -0.06 and considered Kf3 for a long time). Chessmaster6000 found Bh3+ in less than 3 minutes on pentium200MMX expecting Kh1 and in the next ply(20 minutes or so) expected Kf3. 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.