FreeCell Two Decks: 104 open cards and a very long, very fair puzzle
FreeCell Two Decks keeps everything that makes FreeCell the thinking player's solitaire — every card face up from the first second, free cells to park cards in, no stock to rescue you — and doubles it. Two full decks go into ten columns, and you get eight free cells and eight foundations to match. It is the longest game on the site, and one of the fairest.
Our solver, which sees the same open table you do, found a winning line in 88.0% of 400 deals and did not prove a single deal impossible. The rest simply ran past its search limit. Put simply, none of the 400 deals was shown to be a lost cause, so the real share of winnable deals sits in the band from 88% to 100%. On our Spider-based difficulty ladder that places FreeCell Two Decks among the games where the cards almost never beat you; if you lose, it is the plan, not the deal.
Two-deck FreeCell on the ladder
| Game | How we measured it | Winnable deals |
|---|---|---|
| FreeCell Two Decks | Our solver, 400 deals, all cards visible | 88.0% proven; 0% proven impossible; 88%–100% overall |
| FreeCell (one deck) | Our solver, 10,000 deals | 99.68% proven; none proven impossible |
| Spider, 4 suits | Published solver that sees every card | 98.5% |
| Spider, 4 suits | Our no-peeking player, 2,000 deals | 0.2% won |
The Spider rows show why the ladder needs both kinds of figure. Four-suit Spider is almost always winnable in theory yet rarely won in practice, because so many cards are hidden. FreeCell Two Decks hides nothing, so what a solver can do and what a careful person can do are much closer together.
The table on spidersolitaire.us
- Cards: two standard decks, 104 cards, all dealt face up.
- Tableau: ten columns, four of eleven cards and six of ten.
- Free cells: eight, each holding one card.
- Foundations: eight, two per suit, built up from Ace to King.
- Building: down in alternating colors, as in ordinary FreeCell.
- Empty columns: any card, or any legal run, can go in.
- No stock: everything you will ever play with is on the table at the start.
What the solver's games tell you
Plan for a marathon
The first winning line our solver found averaged 148 moves, and those are efficient computer solutions that were not trying to be short or long. A person will usually take more. Treat a game as something to come back to, not a five-minute break.
Half the time, an Ace is waiting
Just over half of the deals we checked, 55%, had an Ace lying at the foot of some column where it could go home on move one. When none is exposed, the first job is to find the Ace with the fewest cards on top of it and dig there.
Eight cells are not eight spare pockets
Doubling the free cells sounds generous, but you also have twice as many cards to sort. The same rule as one-deck FreeCell applies: every card you park weakens you until something comes along to take it. Fill cells to unlock a column or reach an Ace, not to tidy up.
Use the twins
With two of every card, a black 9 has four possible red 10s to land on and two foundation piles to reach. When you are stuck, look for the second copy of the card you need; it is often less buried than the first.
Strategy in brief
- Spend the first minute scanning for all eight Aces and the 2s just above them.
- Target the ten-card columns for your first empty column; they are shorter.
- Keep at least two cells free for as long as you can.
- Do not race one foundation of a suit ahead of its twin while low cards of the other color are still buried.
- When you do clear a column, fill it with a long ordered run, not a single stray card.
Where two-deck games go wrong
Because none of our 400 deals was proven impossible, a lost game of FreeCell Two Decks is almost always a planning problem. These are the usual culprits.
- Filling cells in the opening. Eight empty cells feel like plenty. After a dozen moves they can all be full, and from then on every move is one card at a time.
- Building long mixed runs on the wrong copy. A run built on the "wrong" black 10 can bury its twin's column. Check which copy sits on more useful cards before you commit.
- Chasing one Ace at a time. With eight Aces, the best plan is often the one that frees two at once.
- Not saving progress mentally. In a game this long it is easy to lose track. Undo is unlimited, so when a line fails, step back to the last position you liked rather than restarting the deal.
Settling in for 104 open cards
A table this crowded is where fullscreen and the zoom control prove their worth; the card faces are drawn large so ranks stay legible. For long games, Undo has no cap, so you can back out of a dead-end line many moves later. Hint, Restart and Auto Play round out the toolkit, and you can drag or tap. A small counter tracks wins, time and moves. It is all free, browser-based on computer, tablet or phone, with no account, and sound and felt color are adjustable.
How we measured FreeCell Two Decks, step by step
This test was run by the spidersolitaire.us team on September 19, 2026.
- Took the rules from the two-deck FreeCell branch of our engine: 104 cards in ten columns of eleven and ten, eight cells, eight foundations, alternating-color building and any card in an empty column.
- Built the game as a program that makes only the moves the site allows.
- Dealt numbered hands. We tuned one setting of the solver's scoring on deals 1 to 8, then measured it on deals 101 to 500, which played no part in that tuning.
- Used a solver. Every card is visible in this game, so a solver that searches for a winning line is the right tool, and its verdicts apply to a person too. Each deal got up to 150,000 positions.
- Ran 400 deals and sorted each into solved, proven impossible or undecided.
- Looked for published figures. Blake and Gent (2026) report one-deck FreeCell but not the two-deck game, so this result has nothing direct to be checked against.
- Worked out the margin of error. For 88.0% solved the 95% range is 84.4% to 90.8%; for proven impossible it is 0% to 0.95%.
- Named the limits. 400 deals is a smaller sample than most of our tests, and 12% of deals were left undecided, so we quote a range rather than one number.
Searching a 104-card table
The two-deck solver lives in freecell2.py, the same Python 3.11 module (standard library only) that handles Eight Off. It ranks every reachable position by how promising it looks and always expands the best one, while a lookup table stops it revisiting a position twice. Cards go home automatically only when no lower card of the opposite color could still need them. On 19 processes the 400 deals took about ten and a half minutes. Ranges are Wilson score 95% intervals. The AI assistant Claude, from Anthropic, helped write and review this solver and draft the text; the numbers came out of the solver itself.
FreeCell Two Decks questions
Is every deal winnable?
We did not find one that was proven impossible in 400 deals. At least 88% were solved outright.
Is it harder than ordinary FreeCell?
It is longer, and our solver needed more search to finish it, but the deals themselves look just as fair.
Why are there eight free cells?
Twice the cards needs twice the working space. Eight cells keep the game roughly in balance with one-deck FreeCell.
Where should I start if I have never played FreeCell?
With the one-deck FreeCell, then come back. For another two-deck game that plays very differently, try Solitaire Double or Forty Thieves.
How does this compare with Spider?
Both use 104 cards, but Spider hides most of them. The four-suit Spider page shows how much that hidden information costs.
By Admin, spidersolitaire.us (Solitairen OOD). Research run September 2026.

