Author: Mike S.
Date: 10:43:28 04/13/02
Go up one level in this thread
On April 13, 2002 at 13:09:33, Enrique Irazoqui wrote: >With 128MB hash like you and also the 7-1-2002 F7.0.0.6 engine, I get your >results. So there are at least 2 different Fritz 7.0.0.6 and the first version >does better in test positions. I also got identical results in your 2 test positions, for 7.0.0.7 and 7.0.0.6 (Jan. 11th version). First, I could reproduce the 7.0.0.7 results with my P3/700, 128 MB Hash: 7.0.0.7: [D]8/1p6/p4p1p/2p1P2k/5P2/6pP/1P4P1/6K1 w - - 1.e6 = (0.09) Tiefe: 13/22 00:00:00 110kN µ (-1.19) Tiefe: 19/29 00:00:12 3661kN 1.f5 fxe5 2.h4 b5 3.Kf1 c4 4.Ke2 e4 5.Ke3 b4 6.Kxe4 c3 7.bxc3 b3 8.Kd3 µ (-1.16) Tiefe: 19/37 00:00:17 5193kN +- (1.72) Tiefe: 21/41 00:00:41 13775kN [D]2r1r1k1/p4pp1/1p1Pn1p1/6q1/P5P1/BP2Pp1P/5P2/Q1RR2K1 b - - 0 1 1...Txc1 2.Txc1 Td8 3.Tc4 Dh6 4.De5 Sg5 5.Dh2 Sxh3+ 6.Kf1 a6 7.Tc6 Dh4 8.Txb6 a5 = (0.06) Tiefe: 7/20 00:00:00 109kN = (0.22) Tiefe: 14/40 00:02:39 62475kN 1...Tb8 2.d7 Ted8 3.Td3 Tb7 4.Tc8 Tbxd7 5.Txd7 Txc8 6.Dd1 Dh4 7.Dxf3 = (0.19) Tiefe: 14/40 00:04:47 113092kN = (0.12) Tiefe: 15/34 00:07:38 181878kN 1...Sf4! = (0.09) Tiefe: 15/39 00:14:15 339902kN = (0.09) Tiefe: 15/39 00:14:15 339902kN ----- Then, I switched back to the 7.0.0.6 engine (now I know why I keep those backups :o), to see if I get the different results too: ----- 7.0.0.6: 1.e6 = (0.09) Tiefe: 13/22 00:00:00 110kN µ (-1.19) Tiefe: 19/29 00:00:11 3661kN 1.f5 fxe5 2.h4 b5 3.Kf1 c4 4.Ke2 e4 5.Ke3 b4 6.Kxe4 c3 7.bxc3 b3 8.Kd3 µ (-1.16) Tiefe: 19/37 00:00:15 5193kN +- (1.72) Tiefe: 21/41 00:00:38 13775kN 1...Txc1 2.Txc1 Td8 3.Tc4 Dh6 4.De5 Sg5 5.Dh2 Sxh3+ 6.Kf1 a6 7.Tc6 Dh4 8.Txb6 a5 = (0.06) Tiefe: 7/20 00:00:00 109kN = (0.22) Tiefe: 14/40 00:02:39 62475kN 1...Tb8 2.d7 Ted8 3.Td3 Tb7 4.Tc8 Tbxd7 5.Txd7 Txc8 6.Dd1 Dh4 7.Dxf3 = (0.19) Tiefe: 14/40 00:04:45 113092kN = (0.12) Tiefe: 15/34 00:07:35 181878kN 1...Sf4! = (0.09) Tiefe: 15/39 00:14:15 339902kN = (0.09) Tiefe: 15/39 00:14:15 339902kN ----- Evaluations, depths, moves and node rate figures are *identical*. Some times are slightly different, but that was some system speed effect obviously, when the rest is identical (also the long time for Nf4). There were *two* Fritz 7.0.0.6 engines. First, a version was released with file date Jan. 7th 2002, 16:52. This was quickly replaced by the "final" 7.0.0.6 dated Jan. 11th 2002, 22:27. Version number and file size were identical. Sarah will remember that, because she had to stop her usual 1.000 Chessfun lightning games run because of that :o). I took a third look, and switched back to that *old* 7.0.0.6, testing the first position: 1.e6 Kg6 2.f5+ Kg7 3.Kf1 b5 4.Ke2 a5 5.Ke3 a4 6.Kd3 Kf8 7.h4 Ke7 8.h5 = (0.07) Tiefe: 13/22 00:00:00 110kN µ (-0.87) Tiefe: 18/31 00:00:07 2295kN 1.f5 fxe5 2.h4 b5 3.Kf1 c4 4.Ke2 e4 5.Ke3 b4 6.Kxe4 c3 7.bxc3 b3 8.Kd3 µ (-0.84) Tiefe: 18/32 00:00:09 3182kN +- (1.69) Tiefe: 21/38 00:00:34 12951kN ----- 1.f5 at ply 18 (7.0.0.6 Jan. 7th version). Regards, Mike Scheidl P.S. That Nf4 move looks mysterious at first... but after 2.exf4 Qh4, White "suddenly" is in big trouble. Kovacevic did not take the knight. It's not so much for opening the e-file btw., but to close the diagonal e5-g3(h2): If Black plays 1...Qh4 immediatly, then 2.Qe5 defends. This is probably a good test position.
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.