Author: Jonathan Parle
Date: 22:48:35 02/03/02
Go up one level in this thread
On February 03, 2002 at 04:28:57, Brian Katz wrote: >On February 03, 2002 at 03:10:09, Jose Menendez wrote: > >>On February 03, 2002 at 00:18:55, Brian Katz wrote: >> >>> >>>Can anyone please explain this? >>> >>>I am a USCF rated player, presently rated at 1920, but usually in the upper >>>1900's to low to mid 2000's. Highest rating acheived was 2070. >>> >>>I just recently played a 5 minute blitz game (no increments for either side or >>>added time, just straight 5 minute blitz) vs. Fritz 7 with the latest upgrade: >>>modification date January 11, 2002. >>>I drew this game quite easily, which I find puzzling. >>>Could it be that the Hash Tables I set were too high for my processor speed? I >>>have plenty of ram. I am using a Pentium II 350 MHz Processor with 384 sdram. >>>I set the Hash Tables for 128 MB by the advice of some on this site. >>>According to Steve Lopez from the Chessbase helpline, the formula for Hash >>>Tables should be: >>> 2 X processor speed X avg. seconds per move. >>> This was based on the Fritz 5 program, which allowed you to set the Hash Tables >>>to an exact number of KB's, such as 16,384. Since Fritz 6 and Fritz 7 have Hash >>>Tables with MB settings, you would have to divide your figure by 1000. >>>So for a 5 minute speed game which is 300 seconds, you would have to divide an >>>average game of 50 moves into that. >>> Since the first 10 moves on an average are played instantly, with practically >>>zero time used by the program, it is more practical to divide 300 seconds by 40 >>>moves rather than 50 moves. So that would be: >>> 2(350)=700 X (300 /40)=7.5 = 5250 / 1000 = 5.250 MB Hash Tables. >>>So you would probably set it at 4-8 MB Hash Tables according to this formula. >>> Chessbase states that if your hard drive doesn't have any disk activity at all >>>when you start the engine, then your Hash Tables settings are too timid. They >>>suggest that for Blitz games, 4-64 MB Hash Tables should be enough. Yet, the >>>Optimize Strength feature in Fritz 7 and Fritz 6 is only available for Blitz or >>>long games. Such is the case with the Fritz 6 interface. It is supposed to be >>>the case with Fritz 7 as well, but it works with all the other levels too. >>>(Another Bug? for Fritz 7?) >>> The Optimize Strength feature will set your Hash Tables at the highest possible >>>setting for your amount of Ram! That definitely conflicts with the above >>>mentioned 4-64 MB Hash Tables suggested for Blitz games. This is confusing to >>>me. >>>Please give your opinion on the Proper Hash settings for a 5 minute speed game >>>with a Pentium II 350 MHz Processor, with 384 MB SDRAM. Also, a general formula >>>for any time control would be appreciated based on processor speed. >>> Perhaps the mention of disk activity was really directed at computers a few >>>years ago, which usually for the most part had 32 to 64 to 128 MB SDRAM. So hard >>>disk activity (swapping) was quite common. Lately, many have 256 MB SDRAM and >>>up! Fritz 7 doesn't swap on my computer until perhaps 192 to 256 MB Hash Tables >>>are set. Usually closer to 256 MB Hash Table. >>> Was the fact that it was set for 128 MB Hash Tables too high for this game? >>>Only about 2-3 minutes of time was used by me and less for Fritz 7. >>> >>> Here is the game for anyone who is interested. I usually get crushed by this >>>program. That is why drawing this game as easily as I did (no take backs >>>involved) seems very odd to me. >>> >>>Opinions would be greatly appreciated. >>> >>>Katzparov,Brian (1920) - Fritz 7 (2676) [D43] >>>5 minute speed game HP Pent II 350 MHz, 02.02.2002 >>> >>>128MB, Fritz7.ctg, Pentium II 350 MHz 384 SDRAM >>> 1.d4 d5 2.c4 c6 3.Nc3 e6 4.Nf3 Nf6 5.Bg5 h6 6.Bxf6 Qxf6 7.e3 Nd7 8.Bd3 dxc4 >>>9.Bxc4 g6 10.0-0 Bg7 11.Rc1 0-0 12.e4 e5 13.d5 Nb6 14.Bb3 Rd8 15.Qe2 Bg4 16.Rfd1 >>>Rac8 17.dxc6 bxc6 18.h3 Bxf3 19.Qxf3 Qxf3 20.gxf3 h5 21.Rxd8+ Rxd8 22.Na4 Rd6 >>>23.Nc5 Bh6 24.Rc2 a5 25.Na4 Nxa4 26.Bxa4 Rd2 27.Rxd2 Bxd2 28.Bxc6 Bc1 29.b3 Kf8 >>>30.Kf1 Ke7 31.Ke2 g5 32.Kd3 Kd6 33.Be8 Ke7 34.Bc6 Kd6 35.Be8 Ke7 36.Bc6 Kd6 >>>37.Be8 ½-½ >>> >>>[D]4B3/5p2/3k4/p3p1pp/4P3/1P1K1P1P/P4P2/2b5 b Is this result really that surprising? I think a very large number of people could claim to have drawn with and defeated the strongest PC programs. Your ELO rating indicates that it is theorectically possible to pull points off the program. What would be strange is if you consistently scored draws or even wins off the program. Let us know your score against the program after 100 serious games!!! If you get more than a few points then either you are possibly under-rated or simply a player that Fritz has difficulty with.
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.