The idea to beat FreeCell sprang from a message posted by sci.math regular Russell Turpin on July 24, 1994, with the definitive title "FreeCell is NOT always winnable." In his message, Turpin provided an example of an unwinnable FreeCell deal and asked fellow group members to help prove or disprove his gripe...Turpin’s proof-of-concept hand was indeed unwinnable, but it also wasn’t one of the 32,000 included in FreeCell—possibly for that reason...Ring had made up his mind. They would try to solve FreeCell. All 32,000 hands of it.

  1. In the late 90s I was interested enough in FreeCell to start playing through all the hands in order — if I recall, the version included with Windows 95 kept win/loss statistics for you, and that was incentive enough — but I don’t remember how many I played before eventually needing to buy another computer. (Apparently it was fewer than 11,982 of them.) And I never noticed until now that the version of XP I have installed since then doesn’t seem to include the game at all.

    But at the time, at least, FreeCell seemed like a revelation, a solitaire game that was a true puzzle game, with no hidden information and thus no luck in play beyond the initial deal.

  2. This has always amazed be about FreeCell. I assume the game number is just the seed for the pseudo-random number generator that produces the starting deal. I’d expect that either all starting configurations are winnable, or else some significant percentage aren’t. But instead, there’s just one out of tens of thousands that isn’t. That seems so contrary to how the universe in general tends to work.

    There’s another way of looking at the problem, too, which is how many free cells do you need to win each hand? The game gives you 4, and many hands are winnable with fewer. I think the single unwinnable hand becomes winnable with 5 free cells. If you look at all possible starting hands (not just the 32,000 the game can generate), how many need 5 free cells to win? Do any need 6 or more?

    And what is it about the arrangement of the cards that makes this hand more difficult than all the others? Is it possible to then compute the most difficult FreeCell hand that there could ever be?

Sorry, comments are closed for this post.

Skip to toolbar

What Is Gamefilter?

Gamefilter is a community weblog based in spirit and intent on the great work of Matt Haughey and his staff at Metafilter. It's all about sharing links to interesting stuff on the web, but with a narrower focus on gaming-related links -- good stuff that other people might not have already seen -- and talking about them.

Anyone can join for the low low price of zero dollars, and once you're a member, you have the senses-shattering power to post new stuff right from the front end of the site, and talk about that stuff in the comment threads. You don't need an account to read GFi, but as well as giving you The Power To Post, membership also lets you

  • Thank and Favorite all the goodness
  • Build a profile page with all your stuff in one convenient place
  • Send private messages and friend/follow your favorite GFites and their unstoppable Gfiltering style
  • Feel the warm glow of satisfaction in helping grow the community
  • Do all sorts of other neat things, both now and coming in the future!

Gamefilter is part of the growing MefightClub Network of sites, including our forum home, MefightClub itself, FullGlassEmptyClip, our group blog, and Ludic Research Labs, who are a bunch of bastards.

Your genial host is stavrosthewonderchicken, the miraculous poultry who built and administers the MFC Network sites, and a bunch of other web stuff as well. He has a minor addiction to building websites. He loves you all.