Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: O(1) garbage

Author: Graham Laight

Date: 04:14:44 05/19/01

Go up one level in this thread


On May 18, 2001 at 00:27:00, Robert Hyatt wrote:

>I am reminded of a story:
>
>A young kid walks in after school one day and says "Dad, can you explain the
>difference between theory and reality?  We were discussing this in school today
>and I simply didn't get the difference."
>
>The dad thinks for a few minutes, then says "Son, go into the kitchen and ask
>your mother if she would have sex with the mailman for $1,000,000.00."
>
>The young kid walks away puzzled, but does as his dad suggested, and asks
>his mother the question.
>
>She thinks for a couple of minutes and says "Yes I would."
>
>The young kid returns to his dad and says "Dad, I asked her, and she said 'yes',
>but I don't see what this has to do with theory and reality..."
>
>The dad explains, "Son, based on that answer I can determine two things:  (1) in
>theory, we are millionaires;  (2) in reality, your mother is a slut."
>
>:)
>
>Computers are real, not theory.  Alpha/Beta is O(W^D) not O(1).
>
>Let the theorists spend their money and develop their O(1) algorithms.  :)

I could use this same story to support the case that chess is already solved
"beyond reasonable doubt", and is, with best play from both sides, a draw.

Clearly, as chess play gets better, players (human and computer) are finding it
more and more difficult to obtain a win against opposition of equal strength. To
me, it looks quite likely that around 3500 Elo (USCF scale), players will be
"unbeatable" - so there can never be a 4000 Elo player.

If you are going to say that chess's ultimate result cannot be "proven" without
crunching out every possible game, then yes - it's difficult to argue with that.

But, as a pragmatist who believes in "reality" (like the gentleman in your
story), I have to say that the chances of chess being a forced win for white are
diminishing rapidly.

It's becoming more and more difficult for top players to find a way to beat
their opponents.

I'd also use the "aliens" argument about the forced wins - "if they exist, why
can't we see them?".

Everyone who's read books on chaos/complexity, or who's studied encryption, will
know that it's extremely difficult to design a complex system where the
complexity, at some level, doesn't self negate. Chess wasn't designed for the
purpose of being incredibly complex - it's quite fortunate for the game that
it's survived this far without the results converging nearly 100% onto either a
win or a draw.

It's "possible" that chess is a win for white - in which case, the graph of Elo
level against draws, which from 2200 (USCF) upwards rises very rapidly, will
level off and eventually start to fall again.

I think that chess is a draw beyond reasonable doubt, and I'd be happy to bet
anyone at 11/1 odds, that chess is a draw in the following way:

You send me a bottle of champagne. If, over a period of 3 years, there is a
sustained and continuing decline in the proportion of draws at the top level of
chess, I'll send you a CRATE of champagne.

Happy drinking!

-g



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.