Author: Robert Hyatt
Date: 09:18:51 02/18/03
Go up one level in this thread
On February 18, 2003 at 00:49:34, Uri Blass wrote: >On February 17, 2003 at 11:29:59, Robert Hyatt wrote: > >>On February 17, 2003 at 01:54:24, Uri Blass wrote: >> >>>On February 16, 2003 at 21:45:18, Robert Hyatt wrote: >>> >>>>On February 16, 2003 at 21:01:43, Peter McKenzie wrote: >>>> >>>>>>So you _think_ that is why the computer took the pawn? Rather than just >>>>>>"taking a pawn?" BTW most programs would have played that move. Do you think >>>>>>they _all_ understood what was going to come down that file as a result of >>>>>>their _voluntarily_ opening it up to win a pawn??? >>>>>> >>>>>>I don't. At least not mine... >>>>> >>>>>I don't quite see the relevance of your this. >>>>>You gave Nxg4 as an example of a horrible move, I argued that its not a horrible >>>>>move. I guess you still think Nxg4 is horrible? If so, we agree to differ. >>>> >>>> >>>>I think that in general principle, Nxg4 is _bad_. If it _happens_ that it is >>>>the >>>>best move here, so be it, but I'd bet that a program thinks that black is >>>>better, >>>>and that's wrong. >>> >>>I bet that it does not think that black is better. >>> >>>Even an old version(Junior7) gives advantage for white. >>>My program(Movei) also gives a small advantage for white and likes Nxg4. >> >>I am talking about "black is better after Nxg4 than after another move." IE the >>score goes _up_ for taking the pawn. > >The score for Nxg4 is better than the score of other moves but all of the scores >negative for black. > >> >>> >>>Nxg4 does not win a pawn because white takes the h7 pawn so I see no reason to >>>think that programs evaluate black as better. >>> >>>It is possible that the program planned other things against g4 but understood >>>later that they are bad. >>> >>>Saying that the program played bad only because of the fact that it got bad >>>position is wrong. >> >>Eh? So I can play _good_ and still get a bad position? :) >> >>Then I have been playing "good" since I started playing chess at age 7. :) > >you can play like 2500 and get a bad position if the opponent played like 2800; >I do not define 2500 as bad. > >> >> >>> >>>programs are not perfect but against kasparov even GM's can get a bad position >>>in the opening. >>> >>>If the program played the opening like 2500 and the rest of the game like 2900 >>>then I think that it is not wrong to say that it played like a super GM. >> >>Yeah, but do you think it played "the rest of the game like 2900"??? > >I think that it is dependent on the game and it is possible that it even played >the rest of the game stronger than 2900 in game 3. > >> >>I don't. Again, games 1 2 and 3 could have been all losses, easily, and should >>have >>ended 2.5-.5 at least. That's "super-GM" level chess? Particularly after >>looking at >>game 1? > >I do not get conclusions for one game. >Game 1 was the worse game of junior. > >I do not think game 2,3 were so bad. > >> >>Another criteria for super-GM chess (IMHO): In which game did the comp have any >>sort >>of initiative out of the opening? Perhaps in game 5 after the sac, and even >>that is not a clear >>good move as most seem to think it loses. > > >I think that other super GM's will have problems to get the initiative out of >book against kasparov. > >I think that in game 2 Junior got some initiative out of the opening >Junior also got advantage in game 4 and Kasparov was afraid to adnce b5 because >of unclear piece sacrifice of Junior. DJ had an edge in game 2. It promptly tossed its cookies and accepted the exchange sac and nearly lost as a result. Again, getting the edge from the opening does _not_ mean getting an edge from the opening _book_. That is not hard to do. What is hard to do is to build that edge once out of book, and the computer didn't seem particularly well-suited to doing that. IE a lot of pressure and BAM, take the exchange and nearly lose the game. Being unable to tell the difference between an exchange up and a lot of pressure down, vs a significant amount of pressure ahead, equal material. It just made a bad mistake accepting, and I would be willing to bet that Kasparov _knew_ it was going to accept that "gambit" based on experience with it. > > If you look at the 1997 match, DB2 >>played clearly >>strong chess and had an initiative in several of the games. Game 2 comes to >>mind as a game >>with only one flaw, that of Kasparov resigning when he should not have. But >>Kasparov was >>defending the entire game. > >Game 2 and game 6 were the only games that DB97 got initiative after the opening >but kasparov was less prepared for that match so it is better not to compare >between the matches because this is not the subject of this discussion. Deep Blue _could_ have won game 1. I will _always_ believe that game 1 was so very close, and even GM players have agreed. A very slight issue with the two Kasparov pawns turned it into a loss later, but one minor slip and it would have ended. As happened in more than one of the DJ/Kasparov matches. He turned a probable win or two into draws, and a probable win into a draw and then into a loss... > > In which game in _this_ match do you see that >>happening? And I >>don't particularly assess DB2 as "super-GM" stuff myself. Very strong. Very >>consistent. Just >>like Deep Junior. > >The discussion here is not about the level of DB97. > >Uri I'm not sure _what_ the discussion is about, myself...
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.