Author: Jonas Cohonas
Date: 12:10:57 05/12/01
Go up one level in this thread
On May 12, 2001 at 13:34:40, Aaron Tay wrote: >On May 12, 2001 at 12:40:06, Jonas Cohonas wrote: > >>On May 12, 2001 at 09:54:18, Aaron Tay wrote: >> >>>On May 12, 2001 at 09:22:13, Jonas Cohonas wrote: >>> >>>>and we should remember that all the top players of today owns atleast a copy of the above mentioned programs and then some. > > >>>Top GMs own every copy of the top chess programs? Maybe,maybe not.Some GMS i >>>think, might even practice against Fritz only thinking it's the same as Junior >>>or anyother program. Even if they do, they probably use it more as a aid to >>>analysis, blunder checks and stuff, rather than viewing the program as a >>>opponent to beat unlike rival Human GMs. Why would they borther, except for >>the rare one off exhbitition matches? So the 2703 rating you use is probably >>>over-rated once a GM treats the program as a serious opponent. > > >>The 2703 rating i was referring to was actual rating from practical play >against very strong humans. > >How many games was that? How many of those games do you think, the GM was really >prepared? How many serious [with something at stake] long time control games >does a average GM play against a computer per year? The Gms are just starting to >learn...So you are wrong to think that the 2703 rating incorporates >anti-computer play by GM..For those who really borther, just think back about >how deep JUnior got crushed by Kramnik..You think Fritz will do better even with >better hardware? Fritz might be better at beating fellow computers now..but i >doubt it says a thing about how it will handle Kramnik.. How many games i don't remember excactly, but that is besides the point, if you check all the games of DJ6 against GM's you would be surpised at how well it has done, yes i remember how kramnik "stonewalled" DJ6 in Dortmund or frankfurt, i also remember reading how all the leading programmers took notice and made some changes in their programs to avoid positions like that again, thus the term antihuman would apply there aswell, changes like that are made from time to time to accomodate not only program play versus humans, but also against other progs. Infact one of my dad's friends is a GM here in Denmark who uses Fritz 6 as an opponent and a sparring partner. Yes i think that DF7 will do better than DJ6 it is not only better hardware, but also an improved program over DF6, however in my personal opinion i think that DJ7 would have a better chance against Kramnik. > >>And u ask why GM's should bother owning and playing top >programs like a serious opponent? well as technology progresses and programs get >stronger more and more money are to be made from these events, cpu v human, > >Perhaps, but for now, with FIDE banning serious computer participation in normal >chess tournaments, the average GM wouldn't borther. Perhaps only the top 10 >super GMs i.e the BIG names may consider preparing for one-off Man versus >computer matches..But even for them, those are rare. Exhibition matches may be >great, but unless the cash prize is extremely huge, the incentive for >preparation is less, compared to say FIDE rated tournaments where GMs prefer to >spend their energy and time to win elo points to break into the super GM levels >to acquire inivitations to big tournament. I am not talking about big time preparation, like when Kramnik meets, Kasparov or DF7, but they keep up with the programs of today i am convinced of that and that they do that through standard, blitz or whatever they need to polish off, i dont think that if programs where so inferior as they have been made out to be(against humans) they would 1: Get a chance to meet the world champ 2: have such a good record against humans!! > > >>>If we assume the rating of 2750 is >>>>about right then it would be the actual playing strenght, then there is the >>>>matter of perfect endgame which, >>>2 >>>This makes no sense. the 2750 if correct covers every aspect of the game, why >>>the mention of perfect endgame? Because that is one certenty which Kramnik do not have, and in certain endgames, mainly the very difficult one for humans, DF7 will be superior to any human and perfect endgameplay is not mesurable in terms of rating as far as i know, the 2750 i am talking about is the actual play if you removed the endgame tables, but takinng into account that Kramnik would probably try to avoid endgames like those they will probably not occur that often, but occur they will! >> >>Because if they go into an endgame where 3-4-5 men tb's are to be used the cpu >>have the advantage. Actual playing strenght is where the program makes the >move, which is not the case when the program reaches an endgame where it reads >from the endgame tables. > >Yes. I know that. But my question stands, the rating you use already takes into >the account the use of tbs in all ways..So your addtional remark about perfect >endgames is pointless. > >>>>also the antihuman play is almost by definition implemented >>>>in all of todays programs, >>> >>>What is the definition of "anti-human play"? Is there something similar everyone >>>does to their program that I can point to and say there is "anti-human >>>features". Getting the position open? >> >>antihuman play in my would be avoiding anticomputer play :-) and almost all >>programs today have that in their code or in the opening book, the way i >>understand it. > > >How? How? implemented or How they avoid anti cpu play? > >>> >>> >>>all this taken into account makes the match, to me, >>>>very open indeed. >>> >>>>Computers never have a bad day, they don't sweat under pressure, they don't >care about money, they haave no ego >>> >>>This I agree with. This has always being the main strenght of computers. >>> >>>>and they play the board NOT the man!! >>> >>>Well they definitely play the man, if we are going to talk about anti-human play >>>plus preparation against Kramnik...;) >> >>The program's never play the man(i got the ;) ) they are not aware that they >>are progammed to aviod anti computer play :-) > >Not sure what you mean by "aware" i mean that they do not have a conciense and are therefore not aware. > Regards Jonas
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.