Author: blass uri
Date: 01:08:59 06/30/00
Go up one level in this thread
On June 30, 2000 at 02:31:50, stuart taylor wrote: >On June 29, 2000 at 16:28:54, Dann Corbit wrote: > >>On June 29, 2000 at 04:04:49, stuart taylor wrote: >>>On June 28, 2000 at 01:56:53, David Blackman wrote: >>>>On June 27, 2000 at 00:32:15, stuart taylor wrote: >>>>> Is it not true that human ratings are lower than computer ratings relative to >>>>>true standard of play due to the fact that humans make many blunders of the >>>>>nature that computers do not? >>>> >>>>Correct. >>>> >>>>It is also true that I make many blunders of the nature that Kasparov does not. >>>>This is the main reason that his rating is 1200 points higher than mine, IHMO, >>>>and if truly deserved ratings were used our ratings would actually be much >>>>closer :-) >>> >>>Yes. I know! It can get a little bit complicated. But still, human vs. human >>>is very different to human vs. calculator. The calculator simply does NOT make >>>any mistakes which it is not programed to make. All humans DO- in abundance! >> >>You are wrong. Programs are full of bugs. Opening books are full of bugs. >>Algorithms are deficient. Eval functions are deficient. >> >>Look at some of the funny gaffes like immobilizing your own pieces that have >>been recently demonstrated. >> >>Computers make plenty of mistakes, *especially* positional ones. Some of them >>are simply hilarious. > >I'm talking about the type of mistake that after making, you have to just resign >immeadiately. I don't think the top commercial programmes ever do such mistakes. >e.g. to throw away a peice. Or even something near to that, is usually much more >serious (for the end result) than computer errors. >S.Taylor I saw cases when computers made stupid tactical mistakes. This is rare but it can happen. One example:I know that one of the top programs(Junior6a) can say mate against itself when there is a simple stalemate combination. The programmer knows about the problem. I told Amir about it when I tested Junior5.x versions but he did not fix it in Junior6a(I have not Junior6a but I asked people who have it). Here is the relevant position when Junior6a plays Qd3+ with evaluation of mate against itself. [D]6r1/7k/Q7/8/7K/7B/8/6qq w - - 0 1 I guess that other programs have different bugs. If they happen in less than 1 of 1000 games then many programmers do not care much about them even if they know about them and in many cases they even do not know about them. Uri
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.