Re: How long would it take a computer to completely "solve" chess?

From: David Bandel (dwb1729_at_yahoo.com)
Date: 09/14/04 

Date: 14 Sep 2004 09:14:20 -0700

Theo van der Storm <thstorm@wanadoo.nl> wrote in message news:<4146218A.1050605@wanadoo.nl>...

Now I'll deal with YOU one by one. And I'd like to start by saying
that every single point you made was wrong. It shows you're either
just trying to be argumentative or you know NOTHING about chess.

> Right. One by one:
> >> there are too many directions the solution is being approached from
> Obviously itīs the quality, not the number of directions.

wrong. it is also the quantity. I am not proposing that ONE of the
directions will solve the game singly. i am saying (and this is well
known by any chess programmer) that all the solutions are approaching
the game from different angles to meet at the CENTER. brute force is
starting from the beginning and heading towards the end.. table bases
start from the end and work their way backwards. if they meet
anywhere.. game is instantly solved. any moron can realize that.. i
guess you're not just your "typical moron"

> >> endgame tablebases
> nonsense. There arenīt enough atoms to store the tablebases,
> so you would have to go for another technique, e.g. depth first search.

see my point above. even if table bases aren't up to the 32-man level
(something i would NEVER purport) they are up to the 6 man level.. and
they will be up to the 8 man level within this century surely as you
can see by the numers posted by macon. that is a huge step forward..
now the brute force calculations need only reach positions with 8 men
or less left.. see the significance? if not, stop right here... ur
useless to this conversation.

> >> depth of opening research
> irrelevant. Opening theory gets you nowhere with regard to proving
> things, because you are missing the lines overlooked by flawed humans.
> Yes, we all have something in common, although shall we say: "some are
> more common than others". At the other end of the spectrum IGM Bronstein
> himself says it: there is no such thing as opening theory. It should be
> called opening practice.

opening theory may not rigorously prove that there are perfect lines
to follow.. but they can help you avoid "pitfalls" there are tons of
common openings that lead to well shown to be "inferior" positions for
either side. that in itself aids programs in their endeavors to
"prune" get it?

> >> awesome pruning capabvility of the top chess engines
> Already taken into account with branching factor 3.16 in my calculation.
> For my argument it doesnīt matter much if it would be improved radically
> to say 2.1.

well i've already shown that your argument is a load of crap. there's
no need for the pruning factor to get any better than 3. but if it
reached 2.. that would drastically increase the chance that it would
delve deep enough to reach the "endgame tablebase" stage.

> >> and the brute processing strength computers are gaining
> >> (yes they are approaching some physical limits..
> >> but in the laboratory..
> What were you going to say?:
> "In the laboratory physical limits donīt apply"?
> "Quantum computers can be any size"?
> Get real.

no i was saying that in the computers at home we have not reached the
physical limits yet. and that is what we are basing the "current"
speed of chess computers on. because chess computers are run on
processors available to the public which have not APPROACHED the
physical limits of the uncertainty principle and discreteness of
charge and minimum size and seperation distance between transistors..
and my super conductor point still stands.. though you apparently
overlooked that.

> >> chess will certainly be solved before the year 2100
> You havenīt given any substantial arguments and no calculation.

these are all substantial arguments.. and they all amount to a type of
calculation showing that in all likelihood.. chess will be perfectly
played by a machine "before" 2100. it's an inequality. i'm sorry i
couldn't predict the exact year.. i'm not hari seldon.
 
> Theo
(who has no idea what he's talking about)

david
(who does)

Relevant Pages

  • Re: How long would it take a computer to completely "solve" chess?
    ... >>about chess that he does. ... there is no such thing as opening theory. ... Already taken into account with branching factor 3.16 in my calculation. ... "Quantum computers can be any size"? ...
    (sci.math)
  • Re: with perfect play, should it be 1-0, 0-1, or a draw?
    ... from the point of view of an omniscient chess computer, ... This is like saying that a robot can never run faster ...
    (rec.games.chess.misc)
  • Re: Herculean Chess: Possible large board future of chess
    ... chess though - they see the real problem in chess as being too much ... much opening theory - and, in fact, its opening theory went right ... You can't really win a chess game on points, ... White can move the game into an area where both players get the same ...
    (rec.games.chess.misc)
  • Re: Why is there no Rec.Games.Chess.Variants
    ... > I disagree with the statement that the opening theory is just ... > "deferred" in Gothic Chess. ... "Orthodox" Chess at one time... ... > The game is deceptive in its apparent depths. ...
    (rec.games.chess.misc)
  • Re: Theoretical Computer Science and Disassemblers
    ... As you are fond of saying in your posts, ... they could not beat a human chess player. ... The game itself aside, however, programming a computer to play chess is ... such as arbitrarily pruning search trees when it appears ...
    (alt.lang.asm)