Author: Thorsten Czub
Date: 12:28:05 11/13/97
Go up one level in this thread
>Let me get this right. You do reasonably well in *one* tournament, and >suddenly the "fast programs" are left behind in the dust? What about >the >"fast programs" last year in Jakarta? There was no Mchess, and a tough playing Virtual in Jakarta, and shredder was weaker in Jakarta, and no junior, and no chess-system-tal and no ... The fast programs lost this championship. And if you look which programs lead or dominated the tournament (I am not speaking of CSTal now !) you see no fast program. The same programs that do dominate the ssdf-list, will dominate the championships. And there is Hiarcs, Rebel, Mchess, Shredder, Virtual, Junior and some more, and no fast searcher in the near. And than come some other program. And 100 Elo points behind, we see the fast searchers. Decreasing Fritz-versions. Decreasing Genius-versions. So what do you talk about ? We know that Shredder is strong. I know that shredder is strong. But despite of shredder and virtual - how many intelligent programs were in Jakarta ??? Be a good loser and acceot that Paris was a WATERLOO for the fast-search-paradigm. > >What about Rebel leading the SSDF for a good while, yet clearly falling >in >the "fast program" category based on search speed? Rebel 9 (and 8) is/was no fast-searcher. It is an intelligent program you and any fast-searcher could learn much from. It leads the ssdf-list. When Hiarcs came, he overtook Rebel8. Since Rebel9 is on the market I a think this will overtake Hiarcs6. I don't see your point. Rebel is no fast-search program. > >No one's said that the fast programs will "dominate" everyone else. I >have >said repeatedly that going faster is always better, so long as going >faster >doesn't mean excluding knowledge that is important. Right. I agree. > >You predict "our" demise far too soon. Repeat your result two or three >times >in a row, *then* you can make your claim to fame. We have many years of ssdf-list experience. A massive database of ssdf-games played 40/120. >So don't read too much into Paris just yet. Time will reveal the truth, >as >always... I seem to recall Thorsten being very depressed about how >CSTal >was doing against a "fast program" right before Paris. Don't forget - I changed my testing SETTING because I upgraded from p5/120 into k6/200. This changed parameters of the setting. It took chris a while to change the search effectively. Arround 3 days before Paris he was able to implement SEX-TENSIONS and now we are near. Only some tuning of the search and we are near. Easy job for chris, maybe 3 joints and a nice evening with red-wine ?? >That hasn't >changed >at all, other than the author of that "fast program" did some things >that >turned out to be obviously bad after the fact. Seems so... > >Of course, if you want to believe the "fast" guys are being "left >behind", >feel free to do so. Take me in the group... I am Mchess-fan ! > But be sure that in addition to watching your >rear-view >mirror, you occasionally look out the front window too. Never know >who/what >you might see up there as well. :) > Evidence. Please concentrate on facts :-) >I would also suggest that *any* program that searches at 80K nps on a >P6/200 >must fall into the "fast" category, because that is how fast *I* am >going. Right. As we said: we do 4000 NPS. You are 20 times faster Bob. So you are fast. But you are not strong enough the factor would imply, or ? >When you take that benchmark, you might be surprised to evaluate the >programs >that finished at the top in Paris. I would be hard-pressed to call >*any* of >them "slow and knowledge-based" when considering their speed... Maybe some of them cheat with their NPS to impress the fast-searchers: I can do this too ?? :-)
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.