FreeCell Two Decks: 104 open cards, and where to look first
FreeCell Two Decks takes everything that makes FreeCell a thinking game, with every card face up and no luck of the draw, and doubles it. Two packs are dealt into ten columns, and you have eight free cells and eight foundations to work with.
Our solver found a way to win 88.0% of the 400 deals we tested and could not prove a single one impossible. The remaining 12.0% were still unfinished when the search reached its limit, so the true share of winnable deals lies somewhere between 88.0% and 100%. On a deal this open, reading the table well matters more than anything else, and that is what this page is about.
What a two-deck deal looks like
| What we measured | Result | Deals | How |
|---|---|---|---|
| An Ace sits at the bottom of a column, ready to play | 55.0% (95% range 50.1–59.8%) | 400 | Deal inspection |
| Solved | 88.0% (84.4–90.8%) | 400 | Solver that sees every card, 150,000-position limit |
| Proven impossible | 0 deals (0–0.95%) | 400 | Same solver |
| Unfinished when the search limit was reached | 12.0% | 400 | Same solver |
| Length of the solver's first solution | about 148 moves | 352 solved deals | Not necessarily the shortest |
| Single-deck FreeCell, for comparison | 99.68% solved, 0 impossible | 10,000 | Our FreeCell study |
We know of no published figure for the two-deck game, so these are our own results without an outside check.
Two-deck FreeCell rules on klondike.nz
The layout
- Two standard packs, 104 cards, all dealt face up.
- Ten tableau columns: four of 11 cards and six of 10.
- Eight free cells, each holding one card.
- Eight foundations, two for each suit.
Moves
- A card may go on a card one rank higher of the opposite colour.
- Any available card can be parked in an empty free cell.
- Any card may go into an empty column.
- Foundations are built up in suit from Ace to King; you win when all 104 cards are home.
Reading the deal
About half the time, an Ace is waiting
In 55.0% of our deals, at least one of the eight Aces sat at the bottom of a column, playable on the very first move. In the rest you will need to dig. Find the Ace with the fewest cards on top of it and plan a route to it before moving anything.
Count what is on top of each Ace
With eleven cards in the longest columns, an Ace near the top of a column can have ten cards over it. List the eight Aces and how deep each one is; that list is your plan for the first part of the game.
Look for twins
Every card has an identical twin. When a red 7 needs a home, there are two black 8s it could go on, and two red 7s competing for them. Deciding which twin to use for what is the main new skill this game asks of a FreeCell player.
Spot the short columns
The six columns of ten are your best chance of an early empty column, and with any card allowed into a space, an empty column is the strongest tool on the table.
Playing the long game
Treat free cells as working space
Eight cells feel generous, but a 104-card deal fills them quickly. Keep several empty for as long as you can; in single-deck FreeCell our solver's success fell from about 99.7% with four cells to 74–80% with two, and cells matter here too.
Settle in for 150-odd moves
The solver's first solutions averaged about 148 moves. They are not the shortest possible, but they give an idea of scale: well over twice a typical single-deck FreeCell game.
Build both foundations of a suit evenly
Sending both 2s of a suit home before racing one pile up to 8 keeps lower cards moving and leaves more places for the twins.
Try lines, then take them back
With everything face up, there is no hidden information to blame. If a line of play goes wrong, step back and try another. That is how you learn to see the long routes.
A first-five-minutes routine
- List the eight Aces and how many cards sit on each. Start with the shallowest.
- Pick one of the six ten-card columns as your first target to empty, ideally one full of cards that already have homes.
- Keep at least half your cells free while you do it. Park cards only when the next move clearly gains something.
- When a column empties, fill it with a card that unlocks a buried Ace or a long run, not just the first card that fits.
- Send low cards to the foundations early, both copies where you can, so the 2s and 3s stop getting in the way.
How it compares
If you like the open-card style, single-deck FreeCell is the classic, and Eight Off adds eight cells but builds in suit. For a two-pack game with a hidden stock, try Solitaire Double.
Playing FreeCell Two Decks on klondike.nz
The game is free, needs no download or sign-up, and runs in any modern browser on a computer, tablet or phone. The table is big, so the zoom control and fullscreen mode help, and you can tap a card and then its destination instead of dragging. There is a Hint button, unlimited Undo, Restart and Auto Play, and the game tracks wins, time and moves.
Where these numbers come from
The klondike.nz team measured this game in September 2026 using this site's own rules.
- Rules from our code. Ten columns (four of 11, six of 10), eight cells, eight foundations, alternating-colour building and any card into an empty column, as set in the game's two-deck settings.
- A program copy of the game. We built the two-deck version on top of our single-deck FreeCell solver.
- Numbered deals. Deals 101 to 500. Deals 1 to 8 were used to tune the solver beforehand, so they were kept out of the test.
- Reading each deal. We checked whether an Ace was playable at the start.
- The solver. It sees every card (as you do in this game) and tries up to 150,000 positions per deal.
- Comparison. We found no published result for the two-deck game.
- Error bars. The 95% range for "solved" is 84.4–90.8%; with 12.0% unfinished, the true winnable share is somewhere between 88.0% and 100%.
- Limits. 400 deals is a small sample, chosen to keep the run to a reasonable time, and many deals needed more search than we could give them.
Under the bonnet
The solver is written in Python 3.11 with only the standard library, in freecell2.py, which also handles Eight Off. It is a best-first search: it always expands the most promising position next, keeps a record of positions already seen so none is searched twice, and makes only foundation moves that are provably safe. A shared runner uses multiprocessing to spread deals over processor cores, and ranges are Wilson score 95% intervals. An AI assistant, Claude by Anthropic, helped write and review the code and draft this page; every figure comes from running the code.
Questions players ask
Is FreeCell Two Decks harder than regular FreeCell?
It is longer and asks more of your planning. Our solver finished 88.0% of two-deck deals within its limit, against 99.68% for single-deck (where it was also given a bigger search limit on a second pass), but it never found a two-deck deal to be impossible.
Can a two-deck deal be impossible?
We did not find one in 400 deals. That does not prove none exist, only that they are uncommon.
Why double the cells as well as the cards?
To match the doubled pack. With 104 cards, four cells would leave far too little room to work.
What should I do first?
Find the eight Aces, count how many cards sit on each, and start with the shallowest.
Who ran this study?
The klondike.nz team. See our about page.
By Admin, klondike.nz. Study date: 19 September 2026.