Author: Thorsten Czub
Date: 07:54:08 09/16/99
Go up one level in this thread
On September 16, 1999 at 05:34:15, blass uri wrote: >>You don't know them. Not my problem. >I read about cases like the case when shredder3 did a stupid blunder against >Junior5 in a drawn endgame and that this blunder could not be reproduced but >this is not a proof that something is wrong with the autoplayer because I saw >cases of stupid blunders that cannot be reproduced in games without the >autoplayer. you can never be 100% sure. But who matters at all ? Nobody knows exactly. and nobody ever is interested in exact things. we are human beeings. i don't care if president clinton had sex with this girl. i don't need to know 100%. i was not there in this office. although: i can have my thoughts about it. i don't need to be 100% sure to know that he had sex with her. the same with many other things. nothing in life is 100% sure. you only guess it is. with chess programs it is similar. you very often know later. after a match what has happened. or you don't know after it. is it really important to know 100% if we know that you cannot know anything 100% ? I don't trust in this autoplayer, because my results playing manually are very very different to the cb-autoplayer results. so i don't use it. if somebody is in my house, and i don't trust him (i don't need to have EVIDENCE) than i say: please leave me and go out of my house. >I am interested to know if there is a clear proof that there are more stupid >blunders with the chessbase autoplayer relative to games that people play >manually or with different autoplayer. the best proof is that the results betwenn played WITH and played WITHOUT differ much. you don't need no better proof. Also we know that this company does not WANT to give same chances to all (example:paderborn championship). >I also saw strange results in the ssdf games against chessbase. > >I saw that Hiarcs7.01 won Junior5 28.5:11.5 and I did not see results like it >with hiarcs7.32. understand. >Many people found positive results for Junior5 against Hiarcs7.32 >The best result of Hiarcs7.32 against Junior5 was of Mark young in engine-engine >games (12.5:7.5) with time control 90/60(ssdf games were 120/40 on slower >hardware) but Hiarcs knows better how to use time controls of x minute/y moves >when y is big(For example if I try 120 minutes /200 moves Junior thinks less for >the first moves relative to 99 minutes/ game when 120 minutes/200 moves is >clearly slower than 99 minutes/game) i do not trust any engine-engine matches on 1 machine. i have enough machines. i don't need to test these silly engine-engine matches. if we have enough cars, why shall we emulate a car race one ONE machine ? shall schumacher and others share ONE car ?? pah -
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.