FreeCell Two Decks: 104 cards in plain sight, and not one lost cause found

FreeCell Two Decks takes everything people like about FreeCell and doubles it: two full decks, ten columns, eight free cells and eight foundations. Every card is dealt face up, so there is nothing hidden and no luck of the draw. If a game is winnable, the path is on the table from the first second.

How high does the ceiling go? We ran a solver on 400 deals using this site's rules. It found a full solution for 88.0% of them and proved none impossible. The remaining 12.0% were left undecided when the solver reached its search limit. So the honest answer is that between 88.0% and 100% of deals can be won, and we did not find a single one that cannot.

Play it here in your browser for free, with nothing to download and no sign-up.

The ceiling at a glance

We found no published winnability figure for two-deck FreeCell, so there is nothing external to compare with. For single-deck FreeCell, the published figure is 99.998881% winnable, from Blake & Gent (2026), and our own solver found 99.68% on this site's FreeCell.

Why "undecided" is not the same as "impossible"

A solver works by exploring positions. We gave it a budget of 150,000 positions per deal. When it found a solution, the deal was solved. When it had searched every reachable possibility without success, the deal was proven impossible. When it ran out of budget first, we called the deal undecided.

With 104 cards, the number of possible positions is enormous, and 12.0% of deals needed more searching than we allowed. None of them was proven impossible. Our best reading is that FreeCell Two Decks is a very winnable game, but we will not round "88.0 to 100" up to a single number we have not measured.

The rules as this site plays them

104 cards in ten columns

Two decks, 104 cards, all face up, dealt into ten columns: four columns of 11 cards and six columns of 10.

Cells and foundations

• Eight free cells, each holding one card.
• Eight foundations, a pair per suit, each rising from Ace to King.

What the ten columns allow

• Build down the tableau in alternating colors: a black 6 on a red 7.
• Any card can go into an empty free cell or an empty column.
• You can drag a run of several cards if you have enough free space to move it one card at a time. The game works out the limit for you.

End of the game

All 104 cards on the foundations.

Getting closer to the ceiling

Settle in for about 150 moves

The solver's first solutions averaged about 148 moves, and a person will usually take more. Settle in.

Start with the Aces you can already see

In 55.0% of deals, at least one Ace sits at the bottom of a column and can go home on the first move. Clear those at once, then look for the columns hiding the others.

Eight cells is a lot, and it is still not enough

With twice the cards, eight cells fill up just as fast as four do in the single-deck game. Park cards only when the move that follows gains you something.

Two piles per suit, so check both

Each suit has two piles. When one 4 of spades is buried, check whether the other is free to go up.

Columns are worth more than cells

A column can hold a whole run, while a cell holds one card. Empty a short column early and protect it.

How FreeCell Two Decks differs from the one-deck game

If you already play FreeCell, most of what you know carries over. A few things change enough to matter.

• Every card has a twin. When you need a red 8 to hold a black 7, there are four red 8s in play instead of two. That gives you more choices, and more ways to paint yourself into a corner.
• Columns are longer. Four columns start with 11 cards, and six with 10. Low cards can be buried deep, so it takes more planning to reach them.
• Cells are proportionally the same. Eight cells for 104 cards is the same ratio as four for 52. They do not feel more generous in practice.
• Two foundation piles per suit. You can start the second Ace of a suit as soon as it appears, so both piles can grow side by side.

The practical effect is that single mistakes are rarely fatal. The solver proved no deal impossible in our test. But a game can drift into a tangle that takes a long time to unpick. Short, deliberate plans work best: pick one buried card, clear a way to it, then choose the next target.

Inside the test, one step at a time

The solitaire-online.us team ran this test in September 2026 to find the ceiling for FreeCell Two Decks on this site.

1. Reading the rules from our game code. Ten columns, eight cells, eight foundations, alternating-color building and any card into a space, all taken from the game files behind this page.
2. Rebuilding the game. We extended our FreeCell program to two decks with those rules.
3. Dealing the cards. Seeds 101 to 500. We used seeds 1 to 8 earlier to tune one setting of the solver, so those were kept out of the test.
4. Choosing the player. A solver that searches for a complete solution. Every card is face up in this game, so the solver knows exactly what you know.
5. Running it. Up to 150,000 positions per deal. Solved, proven impossible or undecided.
6. Checking against published research. We found no published figure for two-deck FreeCell. Our single-deck FreeCell solver does agree with the published results.
7. Working out the margin of error. A 95% range means the true value would fall inside it in 19 of 20 repeats. For deals solved: 84.4–90.8%.
8. What the numbers cannot tell you. 400 deals is a smaller sample than our single-deck tests, so the ranges are wider. And "solved" means a solution exists, not that a person will find it.

The code behind the ceiling

The study is written in Python 3.11, standard library only. FreeCell Two Decks shares a module, freecell2.py, with Eight Off, and a shared runner spreads the deals over processor cores with Python's multiprocessing.

The solver is a best-first search: at each step it expands the position that looks closest to a win and keeps track of positions it has already seen. It only sends a card to the foundations automatically when that can never hurt: for low cards, or when every opposite-color card that could need it is already home. We calculate every 95% range with the Wilson score method.

Claude, Anthropic's AI assistant, helped us draft and debug the solver and write a first version of this page. None of the figures came from the AI. They all come from running the solver, and the team checked the method against published results for single-deck FreeCell.

By Admin, solitaire-online.us (Solitairen OOD). Study date: September 19, 2026.

Ceiling questions from players

Is every FreeCell Two Decks deal winnable?

We cannot say every one. We proved none of 400 deals impossible, solved 88.0%, and left 12.0% undecided.

Is it harder than regular FreeCell?

It is longer and needs more planning, and our solver found it harder to crack. We did not find it less winnable.

Is there a limit on dragging runs?

As many as your empty cells and columns allow. The game calculates it. You can always move cards one at a time instead.

How many moves does a win take?

The solver's routes averaged about 148 moves. Expect a relaxed session rather than a quick break.

What else is like it?

Try single-deck FreeCell, or Eight Off, which has eight cells and in-suit building. Solitaire Double is the two-deck version of Klondike.