Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Perfect Chess Is Approximately ELO 4000

Author: Uri Blass

Date: 10:18:21 02/27/02

Go up one level in this thread


On February 27, 2002 at 12:40:41, Sune Fischer wrote:

>On February 27, 2002 at 12:19:35, Uri Blass wrote:
>>>>If it is deterministic, it will in principle lose all matches after it has lost
>>>>just one, this will not happen to a random program.
>>>
>>>No
>>>It is not going to lose all matches because not everybody tries to repeat wins.
>
>Of cause not, people are not like that, thats why I said "in principle" :)
>
>>>The random player does not try to repeat wins so it is not going to lose matches
>>>against this player.
>>>
>>>There are also humans who do not try to repeat wins against deterministic
>>>programs because they find it boring.
>>>
>>>Uri
>>
>>I can add that even not deterministic players may get 0% if they always fall in
>>the same kind of trap.
>>
>>The opening book and the learning of the chess programs of today may not be
>>enough to score more than 0% against the perfect player because there is finite
>>number of book lines and the computer has finite memory.
>>
>>Uri
>
>They _will_ score more than 0% using only random moves, they can do better than
>that so they will score *much* more, I believe you wrote that yourself
>somewhere.
>
>It's pure probability, it is only because the numbers are so high that people
>believe it is impossible. It will not be impossible to draw against perfect play
>since chess _is_ limited, there are only so many different games that can be
>played, lets say that number is X, and the number of drawn games for black
>against perfect white play is Y, then even with random moves black will score
>Y points (or half points) per X games!
>This means the perfect player has a limited elo, it cannot be any different
>IMHO.
>
>Call me naive, but I believe Super GMs can do much better than random moves ;)
>
>-S.

I also believe that super GM can do better than random moves but it does not
mean that every human can do it and
I believe that part of the weak humans players that can beat easily the random
move generator are going to do practically worse than the random move generator
when they play against the perfect player so even if the perfect player play
against humans then the question against who she plays is important for her
rating.

Uri



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.