Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Approximate # of Unique chess positions (FEN working)

Author: Uri Blass

Date: 06:58:59 01/19/02

Go up one level in this thread


On January 19, 2002 at 06:55:11, Sune Fischer wrote:

>On January 19, 2002 at 06:26:17, Uri Blass wrote:
>
>>>Oh, so that's what you mean, well I'm not concerned with counting *too many*
>>>squares since is was supposed to be an upper bound anyway. That it was easier
>>>when the capturings began.
>>
>>I believe that a better upper bound can be achieved by giving every pawn 48
>>squares.
>
>That isn't possible with 32 pieces, and with 31 I assumed a promotion and 64
>squares.
>My plan was to get a descending sequence, but the promotion factor does blow
>things up beyond my wildest dreams.
>
>>The advantage is that the order of pawns is not important and when you assume
>>only 24 squares for the d or the e pawns the order of pawns is important.
>>
>>I remember that the upper bound of my program(the last number that was posted)
>>was better than the upper bound by your calculation.
>
>What is your lowest upper bound?
>I get 10^46 when running your code.

I got 3.701063012120722292782714774208595e46 but no doubt that it can be
improved because there are a lot of impossible pawn structure and even some
impossible material strucutures

Here is an example for impossible material structure(unchanged means the same as
the initial position)

White:2 queens,3 rooks and 6 pawns,unchanged
black:0 queens,unchanged

It means that white captured the black queen and it means that white can promote
only one pawns so having 2 queens and 2 rooks is impossible.

My program counted it as possible because white can get 2 promotions by one
capture but did not consider the fact that it cannot be achieved without
capturing a pawn.

It is possible to change my program by assuming that capturing a piece that is
not a pawn can lead only to one promotion for both sides and I cannot think of
example when the number of promoted white pawns is bigger than the number of
captures+the number of captured black pawns.

we need to prove this theorem mathematically because impression is not enough
and I did not try to write a formal proof.

The number of promoted white pawns is a variable that can be increased by more
than 2 with one capture

The number of good pawns by my previous definition also can be increased by more
than 1 by capturing a black piece that is not a pawn so they are not enough for
a mathematical proof.

I do not remember if there was a formal proof by retko v tomic in the posts
about this subject that were posted about a year ago(I do not remember the exact
date).

I remember that retko v.tomic got a better number but still bigger than 10^46.

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.