Author: J.Dufek
Date: 10:12:18 09/12/05
Go up one level in this thread
On September 12, 2005 at 08:43:27, Alex Shalamanov wrote: >On September 12, 2005 at 07:56:53, Uri Blass wrote: > >>Why? >You know, game positions are the best judge to me; if it comes that when >analyzing with shredder or fruit the evaluation tends to be high first but as it >goes deeper and deeper in search the evaluation falls more and more, it means >they overestimate the stronger side chances too much. As to DF, his evaluation >is usually more objective and restrained. I don't know if it's true only for my >positions, but it's more helping to me to use him as a consultant. I may agree >or disagree with him, but his search is closer to the truth than that of other >engines. I like his style too: it's very objective and restrained, more suitable >to CC. It's not for nothing that most of the CC players in Russia use Fritz as a >consultant to say nothing of OTB players who have always been his loyal fans. > Fritz is useless for correspondence games. He didn't find any deeper positional ideas, his evalution have tendency be to close 0,00, he tends play with king to corner or play with rook pawn, because null-move etc.... >>I do not want to help my opponents so I may report later about the tournament >that I play today. > >Right, so will I. ;) > >>This is something different. >>I dislike shredder because of preprocessing. > >Actually, I meant DF8. His evaluation is never too far from the truth. >>If you want to build a tree and use the score of the engine to decide about your >>move then I agree that shredder is a bad choice because it can give a score of >>0.5 pawn for white regardless of the depth and after a move of black say >>advantage for black regardless of depth. >> >>I clearly hate to see it. >>I think that Fruit and Fritz are better than Shredder for this task. > >I believe, DF is really better. Here I side with you. Fruit 2.1 is more >objective than Shredder but even he can't find the correct move sequence. DF >can, luckily. Fruit WCCC might be not bad, though. We will hardly ever know it, >though, as, I believe, Fruit WCCC will never be open to the public. > > >>I do not think that you are right. >>My experience in the past is that there were cases when I won correspondence >>games simply for the fact that my computer outsearched the opponent computer(now >>it is going to be harder because the opponents that I play against them are >>stronger but we are going to see what is going to happen and I think that >>sometimes the human factor can cause inferior moves). > >There're positions when programmes are weaker, say in closed or semi-closed >positions, endgame etc., then the human factor becomes more decisive. It's only >a matter of skill how to develop them. I'm very inspired by GM Mikhail Umansky's >or GM Grigory Sanakoev's games. > >>I do not know about the real high level but I believe computers with no help can >>easily get ICCF rating of more than 2500. > >You may be right here. Say, GM Umansky is IM in OTB and it's a fact he didn't >use much computer assistance even in his peak success years. I think, if GM >Garry Kasparov ever decided to try CC, he'd make much progress. > >>I see no way how one match can prove something about the future. >>We also know nothing about the strength of hydra in correspondence games and it >>is possible that fruit can also beat it in these conditions. > >It's very doubtful. I don't believe it's true. Again, you underestimate the >human factor. To beat GM Nickel in a CC match one has to be a grandmaster >himself. Whichever programme you take, it'll never be strong enough for that. > >Best regards, >Alex
This page took 0.01 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.