Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: number of pawn positions in chess...

Author: Dan Ellwein

Date: 09:26:03 01/14/00

Go up one level in this thread


On January 14, 2000 at 04:10:33, Michael Neish wrote:

>On January 13, 2000 at 02:54:31, Timothy J. Frohlick wrote:
>
>>I don't think the "a" pawn is going to end up on the e,f,g,or h files and don't
>>forget that the black pawns prevent occupation except by capture of the white
>>pawns. You also have to account for en passant which usually occurs only once
>>per game at most.
>>
>>[D] 8/pppppppp/8/8/8/8/PPPPPPPP/8
>>
>>So the progression should be 20x19x18....etc and be extended out for a total of
>>sixteen pawns.  20x19x18x17x16x15x14x13x12x11x10x9x8x7x6x5x4= 8.109673360589e+17
>>which is a pretty big number of possibilities.  Even if we limited  the combos
>>to (Whitemoves x 10x9x8x7x6x5x4x3) + (Blackmoves x 10x9x8x7x6x5x4x3)=3,628,800
>>possibilities you can see that we will not run out of moves soon.  If the
>>average game is 75 moves or 150 ply we would have to play at least 24,192 games
>>to exhaust the possibilities.  Again, since not every move is a pawn move we
>>would have a variable number of pawn moves and positions.  So if only one
>>quarter of our moves are pawn moves we would have to play about 100,000 games to
>>come to a correct estimation of the number of different pawn structures.
>>
>>This is just my opinion and is probably wrong.
>>
>>Tim Frohlick
>
>Very interesting comments.  I never thought of that.  The best I can
>do -- for sixteen Pawns, which seems to be the one giving the most
>possibilities -- is 15^8, which is 2.563e9, which is a whole lot less
>than the "naive" calculation I slogged through in another post of
>about 2.9e16 (for 16 Pawns).  For sixteen Pawns my assumption is
>that all Pawns are on their original files, although there are
>surely many more possiblities if pieces have been captured.
>
>Okay, let's look at some possibilities.  If White has captured just
>one piece, creating doubled Pawns on one file, the number of
>combinations becomes 3.35e10, by treating each file individually
>and multiplying the combinations for each of them.  For one doubled
>Black Pawn, it's the same, so just multiply this figure by two to
>get the combinations for White and Black.
>
>For one set of doubled Pawns for each side, well, get the computer out.
>
>For doubled White A-Pawns, Black can have Doubled Pawns on files
>A to F, and in addition, for files B-G the file to the left or
>right of the doubled Pawns can be empty.
>
>For the case of doubled White A-Pawns, we get
>
>White Doubled A, Black Doubled A = 1.025e9
>White Doubled A, Black Doubled B, no B-Pawn on A-file = 2.567e9
>White Doubled A, Black Doubled B, no B-Pawn on C-file = 4.101e9
>White Doubled A, Black Doubled C, no B-Pawn on B-file = 2.734e9
>White Doubled A, Black Doubled C, no B-Pawn on D-file = 6.561e9
>White Doubled A, Black Doubled D, no B-Pawn on C-file = 6.561e9
>White Doubled A, Black Doubled D, no B-Pawn on E-file = 6.561e9
>White Doubled A, Black Doubled E, no B-Pawn on D-file = 6.561e9
>White Doubled A, Black Doubled E, no B-Pawn on F-file = 6.561e9
>White Doubled A, Black Doubled F, no B-Pawn on E-file = 6.561e9
>White Doubled A, Black Doubled F, no B-Pawn on G-file = 6.561e9
>White Doubled A, Black Doubled G, no B-Pawn on F-file = 6.561e9
>White Doubled A, Black Doubled G, no B-Pawn on H-file = 6.561e9
>White Doubled A, Black Doubled H, no B-Pawn on G-file = 6.561e9
>
>Which comes to 7.5012e10.
>
>And this is for a doubled White A-Pawns only.  I couldn't
>be bothered to write a program to do it for all possibilities
>for one set of doubled Pawns for each colour, so if someone
>does it could you post the results on here?  Thanks.
>
>By the way, why are we doing this?  Academic reasons or is
>someone thinking of setting up a Pawn database?

Michael... I was doing this just to get a 'feel' for what's involved in chess...

obviously... chess is not for the faint of heart... i see... :)

and yes... setting up a pawn database is in the works...

PilgrimDan
>
>This is a tricky business!
>
>Cheers,
>
>Mike.



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.