Author: Robert Hyatt
Date: 18:51:57 11/08/00
Go up one level in this thread
On November 08, 2000 at 21:21:35, Thorsten Czub wrote: >On November 08, 2000 at 19:09:04, Bruce Moreland wrote: >>I don't know what you are attributing to Bob. I've heard Bob say that he >>doesn't want to see wild and wrong stuff. I don't think he is against material >>sacrifices in principle, he just wants to be sure that they are right. > >and this is the paradoxon, or contradiction. you can not know. >you don't know. nobody knows. >if it is a real sac, you can't see in the moment it happens. >only if it is a kind of bednorz-toennissen-test-suite-position >you KNOW because there is a certain key move and the position >is not unclear when the key-move comes. > >you cannot be sure. that the move is accurate. >in the moment he believes: it's not forbidden, only we have to make >the thing accurate, he somehow classifies himself to be old paradigm. > >there is no move that is accurate. Thorsten, these discussions are often fun, but _please_ don't make statements that are utter nonsense. An example: take position fine 70. I take white. I play Kb1. I win no matter what you do. _that_ is accurate. I will be happy to construct as many positions as you care to try, where there is a correct move to win, even though I don't feel like calculating to the ultimate mate. I am also willing to construct positions where there is more than one winning move, so long as you recognize the winning _idea_. That is _also_ accuracy. So drop this "drive into the fog" nonsense. It only gets you involved in a massive wreck. >only a mate in 5, or moves from tablebases are accurate. >but not normal chess positions. see above. this is simply incorrect. > >if you are an assurance company, and somebody buys an insurance policy, >e.g. he wants to get money when he dies, for his children... >you don't know when he dies. all we know is: he will die ! >same with computerchess. we know it will be 1-0 / 0-1 or draw. >but we don't know when. if he dies you search WHY he dies, but >how can you be sure exactly. I can think of several ways this can be done. Not legally, perhaps, but _still_ done. :) >these are fuzzy things. we trust. each day in our life we believe that >the next day we wake up we will still be alive. because the way >we live is normally not the way as if we would know WHEN we die. >we forget about death. we have to. to generate the power to do it again, >and again , and again, despite death. > >a chess program that considers anything, trying to be accurate, will >do nothing. a human beeing who behaves like an accurate chess program >would be paranoid ! a person who wants to think anything in forward, >and does not want to say something before it has thought any consequences >in forward is not normal but ill. call this guy autistic or catatonic or >whatever, but he will not be normal. > >old paradigm believes computers are computers and programs have to >take this into account and computer programs should be accurate. Correction (important correction). Not "accurate". "accurate as possible". There is a huge difference in the two terms. > >i remember my grandpa always liked when i had short hair-cut, because >short hair was arrucate hair ! he believed (really) that persons with >long hair are unable to be correct and you should not trust in >young people with long hair. >somehow i think bob believes in chess programs with short hair. >he cannot understand that chess programs of the future >do not compute about accuracy anymore. > >i am a lasker fan. i like his ideas because they fit in my point of view >about the world. i don't believe there is a world outside and we only >have to find out HOW this world is build. IMO WE build the world with >our interpretations. >so for me it is wasted time to try to make a program playing accurate, >when nobody knows what this is. only mate in X is accurate. >but those positions are rare in a game of chess. > >bob tries to find out the BEST move in a position (Tarrasch), he believes this >move exists. he is right. the move exists. but not for humans and computers. >only for gods. Care to take my position test? I'll make 10 positions, you either find the right move or you lose/draw. Humans do this _all_ the time. > >chris and i do not believe in best moves. we believe the best move >is relative. and that it is not important to play accurate moves. >all you have to do is to play moves that lead into things YOU know. If you don't want accuracy, _how_ can you know where the moves lead? That is a 100% circular argument. If it leads to a position you know you can win, then yes, I like that move. If it leads to a position where I am material down, and I can't see winning or losing, then _no_ I don't like that move. > >when americans go into foreign countries, the first thing they >do when they want to go eat is: McDonalds ! They drink coke and >all this fast-food because it's part of their roots. >they eat and drink what they are used to. >normal, isn't it ?! Not me. Haven't been inside a McDonald's in 25 years or longer. No future plans to visit one either. > >you can argue/discuss about the cultural reasons and advantages to do so, >but... > >IMO chess programs should do the same. >they should try to make their game. > >bob is IMO not trying to do so. he wants to visit paris, as an american, >unable to speak french, and wants to order the accurate meal ! > >But HOW bruce ? how to order french cuisine meals if he does not understand >the language ?! > >this is what i very often see with stupid programs. they play moves >that don't lose directly, but that lose on bigger distance. >and these programs don't see this. they have no idea why they lost the game. >for them, suddenly their position is shit, and they cannot defend anymore, >and thats it. > >i would call the old paradigm the materialistic paradigm. they do not trust, >they do not believe , they want to proof. > >i would call the other paradigm the idealist. they don't want to proof anything. >they sit behind their machines, drink and eat and have fun and get enthusiastic >when their programs plays a strange move, and they get even happier >if the move wins. especially they have much fun when they play >in tournaments, and ther opponent program says : 0.00 ! and new paradigm >program says : +3.56 and sacs another piece. > >if their program wins, they make a big party and drink much. or they be nice to >their family. or talk years later about it. and when they lost the game, >they forget it very soon. :-))) > >the interesting confrontation comes from the different (complementary) >point of views. > >i find this cool. and i like that christophe has such a cool program too. > >i have great respect for programs like wchess/superconny, socrates, >mephistoIII, mchess, diep, ... > >and i am bored to see the other programs play. >i like cool programs. like cool (old) cars. >i understand that bob find accurate programs cool. >ok - no problem for me. >but what i don't understand: if he knows that NOBODY will find out if the move >is accurate, why trying to do it ? > >and HOW to call this effort MORE scientific then other approaches ?! > >hm. > > >>We all can make up our own minds without feeling like we are enslaved to Bob. >>If mine sacrifices a piece and loses, big deal. > >detect the fuzzyness positions nobody knows. drive directly in. >why solving the position ? and HOW? > >when i drive in the night with my car, >and there comes a curve, i do not stop driving and went out of the car >to see if there is something out of the light where i could drive into. > >now comes bob. or materialists like him. they went out of the car, >take a flashlight, look into the new direction, go back into the car, >drive a few meters and again use the flashlight to light >the direction the car wants to drive into. > >they want to know in forward if there is a problem. and they waant to >be sure. maybe it could be risky to drive into ! > >maybe people like bob will construct car lights that light into the curves. >brilliant. i will buy this car. >citroen ds / sm / c5 lights the curve. brilliant ! >i buy it :-))) >but --- >new paradigm chess program does it different. if there is darkness somewhere, >drive into it ! then wait until it crashes... >if you are happy, you will stand the crash. so be prepared ! airbags. >whatever you need. > > >>I've seen it a few times for sure in the past 24 hours. It plays well. I have >>seen it try very hard to press attacks, but it hasn't done any long-term >>sacrifices yet. > >what about rc6, isn't this long-term ?! > >its so long that we do not know for sure. >and ? IMO it is not important to know in forward. let the program >find a way out. trust it. >don't try to find out in forward. >the program WILL make it. >like with you son.- you cannot make anything at the best for your son. >ONE day he has to live alone, trust in him, and in yourself, because you >tried to teach him, trust. if you try to be accurate with his future, >he will die because you do not let him live. same with a chess program. > >that is new paradigm. give the program the freedom to make inaccurate >moves. >trust. > >"if you love somebody - set them free" or how was the song by police? > > >> Another thing it seems to do is highly prize well-posted >>pieces, wherever they occur. > >right. > > >>It doesn't play like other programs, > >thank you. at least ONE person subscibes my observations. partly. >whatever. you have seen this too. its easy, isn't it. > > > >>and it is very stiff on defense even when >>it's getting whacked in the endgame. > >right. > >>Here are some games against a Pentium III >>600 (mine on it's quad 450 Xeon): > >oh - i need time to replay. its 3h a.m. in germany. have to go to bed now. >will replay the games tomorrow. > >thanks for sending them. >first some hours sleep, and then i will take a look into the games...
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.