Why Almost Every FreeCell Deal Is Winnable (And the One That Is Not)
FreeCell has a reputation that no other patience game has managed: people will tell you, flatly, that every deal is winnable. It is very nearly true, and the exception is famous enough to have a number.
Microsoft shipped 32,000 numbered deals with Windows FreeCell. Players worked through them methodically for years, and eventually all of them fell except one. Deal 11982 is unsolvable. Not hard - impossible, provably so once solvers got strong enough to exhaust it. A handful of others in that set are brutal, but 11982 is the one that cannot be done.
That single fact tells you almost everything about what kind of game this is.
Why almost everything is winnable
Look at what the deal actually gives you. All 52 cards are face up from the first move. Eight columns, four free cells, four foundations. Nothing is hidden, nothing is shuffled again, and no card arrives from anywhere you cannot see.
Compare that with the game most people mean by "solitaire". When you play Solitaire online in its Klondike form, a large part of the tableau is face down and the stock order is unknown. Some Klondike deals cannot be won by any sequence of legal moves, and you have no way to know which ones until you are stuck. That is genuine, irreducible luck.
FreeCell removes it. The randomness happens exactly once, before you have made a decision, and then the game becomes a pure search problem. Which is why the solvability rate is not 80% or 90% but something like 99.999%.
The move formula nobody teaches you
Here is the thing that separates players who win most deals from players who win some. FreeCell lets you move one card at a time. Every interface will happily move a run of five for you, but that is a convenience animation on top of a sequence of single moves - and it only works if you have the space.
The number of cards you can shift in one go is:
(free cells + 1) x 2(empty columns)
Four empty cells and no empty columns: five cards. Four empty cells and one empty column: ten. Two empty columns: twenty. The formula is multiplicative in empty columns and merely additive in free cells, and that asymmetry is the strategic heart of the game.
It means an empty column is worth vastly more than a free cell. Players who fill columns to clear a cell have the trade exactly backwards.
Four habits that do most of the work
- Empty a column early and defend it. It doubles your manoeuvring capacity. Do not fill it with a stray king because the king had nowhere else to go.
- Do not send cards to the foundations too soon. A two on the foundation is a two that can no longer accept a red ace... or rather, can no longer help you park an awkward low card in the tableau. Keep low cards available until their job is done.
- Free cells are parking, not storage. A card in a cell is doing nothing. Four occupied cells is not a full hand, it is a seized engine.
- Plan the whole dig before the first move of it. Count the cards you must relocate and name a destination for each one. If you cannot, the dig fails halfway and you have spent your cells for nothing.
That last habit is the transferable one. It is the discipline of not creating a dependency you cannot resolve, and it shows up in every good game. Anyone who has learned not to bury their own well when they play Tetris online is doing the same accounting against a clock, and anyone choosing which of four identical tiles to remove when they play Mahjong online is answering the same question about future reachability.
What the variants change
The family on this site is really a set of experiments in adjusting exactly one parameter:
| Variant | What changes | Effect |
|---|---|---|
| Baker's Game | Build down in suit, not alternating colour | Far harder - halves your legal placements |
| Eight Off | Eight cells, same-suit building | Easier despite the suit rule |
| Seahaven Towers | Ten columns, four cells, same suit | Tight and unforgiving |
| ForeCell | Cells pre-filled at the deal | You start already constrained |
| Relaxed | Only kings may fill empty columns is lifted | Gentler, good for learning |
Play Baker's Game for one evening and you will understand suit restrictions in your bones - it is the same board with one rule changed and a dramatically different difficulty.
The honest part
Because FreeCell hides nothing, it also offers no excuses. You cannot blame the deal. The deal was, with overwhelming probability, winnable.
That is uncomfortable and it is also the reason the game improves you faster than its relatives. Games with fully visible state give legible feedback: every loss points at a specific decision. It is the same reason a board game with nothing hidden is such good calibration - take Reversi online: there is no deal to blame at all, and the discipline of owning the result carries straight back here.
If you want the opposite experience occasionally - a game where the randomness is genuinely the engine and the correct play still loses sometimes - play Yahtzee online for half an hour. It is a useful reminder of how unusual FreeCell's contract really is.
Try it on a real board
Play the standard game until the move formula is instinctive, then take on Baker's Game to feel what a suit restriction costs. The rules hub has the exact move legality, the glossary covers the terms, and the move-limit explainer works through the formula with examples.
And if you want to know why the deal you just lost was almost certainly winnable, the companion piece covers what the solvers actually found.
💡 Put it into practice: deal a free game of FreeCell, read the rules of every variant, or try today's daily challenge.