Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The total number of possible chess positions? WT

Author: Robert Hyatt

Date: 06:42:06 09/03/00

Go up one level in this thread


On September 02, 2000 at 02:00:33, Uri Blass wrote:

>On September 01, 2000 at 23:41:29, Robert Hyatt wrote:
>
>>On August 31, 2000 at 15:54:16, Frederic Friedel wrote:
>>
>>>On August 31, 2000 at 15:01:18, Vincent Lejeune wrote:
>>>
>>>>If I remember well I red in a magasin some years ago that there is 10^80
>>>>possible positions and there is 10^120 different playable games (no
>>>>demonstration was given).
>>>
>>>There are 10^112 possible games lasting 40 moves. This is considerably more than
>>>the 10^82 elementary particles in the universe. It is clear that, for principle
>>>reasons, all possible games will not be reconstructed (generated and stored) in
>>>the course of this universe.
>>>
>>>But we don?t need to do that in order to solve chess (in the Thompson endgame
>>>sense). The number of possible legal chess positions is far smaller: between
>>>10^53 and 10^55. So will we (or someone or something) be able to work them out,
>>>effectively retro-analysing the 32-piece endgame chess represents?
>>>
>>>Again the answer is no, but not on principle grounds but for practical reasons.
>>>A computer processing a billion positions per second would require about 10^38
>>>years to solve the game. If you use a billion computers in perfect
>>>multi-processing you will still have to wait 10^29 years for the answer (and if
>>>you are not careful with the program it might simply produce  ?42?).
>>>
>>>But there is still a major problem. You cannot store the tables on CDs or DVDs.
>>>As John Nunn explained to me we will need the matter from many millions of
>>>galaxies to store the information that is generated. So solving the game using
>>>the method of exhaustive analysis is theoretically possible, but one wonders
>>>what Greenpeace would say if we started dismantling galaxies in order to store
>>>chess positions.
>>
>>
>>This is way too simplistic an estimate, unless the 50 move rule and 3-fold
>>repetition rules are discarded.  In reality there are nearly an infinite number
>>of positions, because each position is unique to the path leading to it.  Just
>>because the same piece configuration has been reached in two or more positions,
>>the positions are not guaranteed to be equal.  How many different pathways can
>>be spanned to reach those identical positions?  And how do those pathways
>>affect the 50 move and repetition rules?
>>
>>When you take a single position with the 32 pieces (or fewer) placed on the
>>board in a legal configuration, and then try to enumerate all the pathways
>>from the original position which can lead to these two identical positions,
>>the number of pathways is absolutely enormous.  And each "identical" position
>>would be different because the moves played _after_ each of these positions
>>would inherit different 50 move counters and different repetition lists.
>
>Repetitions are not relevant for the 32 piece tablebases so they are not
>relevant for the complexity of the problem of solving the game.
>The 50 moves rule is usually not relevant and a program that has the 32 piece
>tablebases without considering the 50 move rule probably can solve the game(It
>can solve the game without the 50 move rule and if it say that the game is drawn
>then you can be sure that the game without the 50 move rule is also drawn).
>

Yes... but if the tabelbase says "win" you can't be sure it is a win.  It
could be a 50 move rule draw just as easily.  This _already_ happens in today's
tablebases.

However the original question was just "how many unique positions are there?"
and the path information _must_ be considered in that...  because I can reach
the "same" position with different sequences of moves, which limits what can
happen after that position with respect to 50 moves and repetititions.



>>
>>I think you could safely say that the game is nearly infinite...  or at least
>>so large that the difference between the real number and infinity is not
>>easy to interpret.
>
>I agree that you can safely say that the number is so big that the game will not
>be practically solved.
>
>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.