Almost. A solver playing perfectly wins 99.9989% of random FreeCell deals, which works out at roughly one impossible deal in ninety thousand. If you have ever lost a FreeCell game, the deal was almost certainly not the problem.
That is a very different answer from Klondike, where about one deal in five cannot be won no matter how well you play. The reason fits in one line: FreeCell deals every card face up.
The Microsoft 32,000
The original Windows FreeCell shipped with 32,000 numbered deals. Because the numbering was reproducible, players could compare notes on the same deal — and in the mid-1990s they did, at scale. The deals here are numbered the same way: deal #1 in FreeCell is the same board as Game #1 in Windows, and the number of the deal you are playing is shown at the top.
The Internet FreeCell Project, led by Dave Ring, ran from August 1994 to April 1995 and set out to solve every one of the 32,000 by hand, with volunteers sharing their solutions online. They got all of them except one.
Deal #11982 has no solution. Not "no solution anyone found" — it was later confirmed by computer that no sequence of legal moves finishes it. It is probably the most famous single hand of patience in existence, and it is famous for being impossible. You can play deal #11982 here and find out for yourself where it runs out.
Later Windows versions expanded the pool past a million deals, and in that larger set eight are unsolvable. Microsoft also added deliberately impossible hands as an easter egg, numbered negatively, which do not count toward your statistics.
What the solvers say now
The Solvitaire project at the University of St Andrews ran a general solver over ten million random FreeCell deals and reported winnability at 99.998881%, give or take 0.000207%.
That is thoughtful play — perfect information, every line searched. In FreeCell that caveat barely matters, because perfect information is what the game gives you anyway. Unlike Klondike, the solver is playing almost exactly the game you are.
The free cells are the whole game
The most interesting thing in that research is not the headline number. It is what happens when you change how many free cells there are.
| Free cells | Deals that can be won |
|---|---|
| 4 (standard) | 99.9989% |
| 3 | 99.36% |
| 2 | 79.5% |
| 1 | 19.3% |
| 0 | 0.21% |
Read that bottom-up. With no cells, about one deal in five hundred is winnable. Each cell you add multiplies what is possible, and the jump from two to three is the difference between a hard game and a solved one.
This is worth carrying into how you play. Every card sitting in a free cell takes a chunk of the game's room away for as long as it sits there. Filling three cells to make one clever move is usually how a FreeCell deal is lost — not by a bad deal, but by a self-inflicted one.
What an impossible deal looks like
There is no pattern you can spot at a glance. Unwinnable deals tend to bury the low cards you need early under long stretches of high ones, in columns that cannot be emptied without more room than four cells give — but plenty of winnable deals look exactly like that too.
That is the practical point. Because unwinnable deals are so rare, the right assumption at the table is always that the deal is winnable and the line exists. If you are stuck, the answer is in the board, not in the shuffle.
So why do people lose FreeCell?
Because a solvable deal is not an easy one. Knowing a solution exists tells you nothing about whether you will find it, and FreeCell punishes the moves that feel most natural — filling cells early and emptying them late, breaking a column you will need intact, chasing an ace before the column above it is clear.
Our own hint shows how far the game is from luck. It searches for a complete solution, for a fraction of a second, before it suggests a move. Left to play the first 2,000 Microsoft deals on its own, it won 97.2% of them — against 45.6% of Klondike deals at Turn 1 and 12.4% at Turn 3, played the same way. Nearly every FreeCell deal has an answer, and with every card in view, even a quick search finds it.
What to do when a deal looks lost
Undo, and go back further than feels necessary. In Klondike, stepping back is a gamble because the cards you have not seen may not cooperate. In FreeCell every card was visible the whole time, which means the mistake is findable — it is usually a few moves before the point where you noticed the trouble.
Try it on a deal in FreeCell, or see how the same question plays out in the game where the answer is genuinely "no" — is every game of solitaire winnable. If you are deciding which of the two to play, there is Klondike vs FreeCell.