From: Jim Pike <jpike@rtis.ray.com>
Newsgroups: sci.math
Subject: Re: puzzle
Date: Mon, 06 Apr 1998 12:36:37 -0500

of the 32000 possible games in the game freecell, 11982 is believed
to be unsolvable - http://physics.tamu.edu/~Cdude/freecell.html

There are also ways to construct an unsolvable hand:
http://spider.cs.ruu.nl/~hansb/d.freecell/node2.html#SECTION0002

And there are also ways to get quick answers to these kinds of
questions: http://www.hotbot.com/  search for freecell and unsolvable

jim pike

LSC wrote:
> 
> Can someone prove that every game of freecell have a winning strategy or
> give a counterexample of a freecell that one cannot win? The freecell here
> is the game in MS Windows 95.
==============================================================================

Newsgroups: sci.math
From: wpclay@arl.mil (William Clay)
Subject: Re: puzzle
Date: Mon, 6 Apr 1998 19:34:55 GMT

On 6 Apr 1998 15:12:20 GMT, "LSC" <liausc@pl.jaring.my> wrote:

>Can someone prove that every game of freecell have a winning strategy or
>give a counterexample of a freecell that one cannot win? The freecell here
>is the game in MS Windows 95.

LSC,

	You might want to check out
http://www.capecod.net/~wcallan/freecell.htm
According to them, all games except one (11982) have been solved.  I
have heard that someone has proved that 11982 is unsolvable, but I
have not seen the proof myself.

                             -- Bill Clay

==============================================================================

Subject: Re: Freecell
Newsgroups: sci.math
From: zeisel@vai.co.at (Zeisel Helmut)
Date: 20 Oct 1999 14:38:11 +0200

In article <380da5b6@webpc1.vai.co.at>, zeisel@vai.co.at (Zeisel Helmut) writes:
>In article <7uk520$1pk3$1@news.cybercity.dk>, "Daniel Molnar" <murmix@post.netlink.se> writes:
>>
>>theodore hwa wrote in message <7ubp48$id7$1@nntp.Stanford.EDU>...
>>>Um, what???  It is easy to construct an unwinnable game.
>>>
>>>In the Windows Freecell program,
>>>
>>>try games -1 and -2 (I'm serious).
>>>try game 11982.
>>
>>
>>While we're at the subject, does anyone have the slightest clue
>>about the algoritm used to calculate the board.
>>The available numbers (1 to 32000) are clearly not enough to
>>all the (52!) combinations. 
>
>I have no exact information,
>but I would assume that the "number" of the game
>is the starting value for some simple random number
>generator (maybe, say, the ANSI C generator),
>and the generated pseudo random numbers are used to "mix" the cards.
>At least, this is how I would do it
>to obtain games that are random enough for a simple game
>and which is simple to implement.
>
>Helmut
>

Maybe next time I should consult yahoo before posting:

Following some links from http://www.freecell.org,
one finds 
http://members.aol.com/wgreview/mshuffle.txt,
where the MS programmer of Freecell
publishes the algorithm.

Helmut