Why FreeCell Is Almost Always Solvable

Not all solitaire games are created equal when it comes to winnability. Most popular solitaire variants have solvability rates far below FreeCell's:

The gap between FreeCell and everything else is enormous. What makes it so special? Three interlocking design principles.

FreeCell deals all 52 cards face-up into eight tableau columns. There's no stock pile, no hidden cards, no draw mechanism. You see the entire game state before making your first move.

In game theory, this is called perfect information — the same property that chess and Go have. Every player (or solver) can analyze the complete state of the game at every point. This is fundamentally different from Klondike, where hidden tableau cards and an uncontrollable draw pile inject randomness into every decision.

Perfect information means that if a solution exists, it can in principle always be found through analysis. There are no "blind alleys" caused by face-down cards or unknown draw order. The only question is whether the specific arrangement of cards permits a winning sequence of moves — and in FreeCell, the answer is almost always yes.

The four free cells are the engineering marvel at the heart of FreeCell's design. They serve as temporary storage spaces where you can park cards that are blocking your progress.

Without free cells, card-blocking situations in an 8-column layout would frequently be impossible to resolve. Card A blocks Card B, which blocks Card C, creating chains of dependencies. Free cells break these chains by letting you temporarily remove cards from the tableau, creating the space needed to rearrange the remaining cards.

The mathematics are revealing. With 0 free cells, you can move only 1 card at a time. With 4 free cells and no empty columns, you can effectively move up to 5 cards in a sequence. Add an empty column, and that jumps to 10. Two empty columns: 20. The supermove formula shows how free cells and empty columns create exponential movement capacity:

Max cards = (1 + free cells) × 2^{empty columns}

Four free cells are the sweet spot: enough to prevent most deadlocks, but not so many that the game becomes trivially easy. Baker's Game — which uses the same layout but with same-suit building — has only 4 free cells and drops to ~75% solvability, proving that free cells alone aren't sufficient. The building rule matters too.

FreeCell's tableau building rule is deceptively important. By allowing any card to be placed on any card of opposite color and one rank higher, FreeCell gives each card two possible destination cards at any time.

Compare this to Baker's Game, which requires same-suit building. There, each card has only one possible target — the next higher card of the same suit. This cuts available moves by roughly 75% and drops solvability from ~99.999% to ~75%.

The alternating-color rule creates a much denser "graph" of possible moves. In graph theory terms, each game state has more outgoing edges (legal moves), which means the search space is better connected and there are more paths from any starting position to a winning state. Deadlocks require a much more specific (and therefore rarer) card configuration to be truly unresolvable.

This is why FreeCell's design is so elegant: perfect information eliminates hidden randomness, free cells prevent hard deadlocks, and alternating-color building maximizes the number of available moves. Together, these three principles make the game almost always solvable while remaining genuinely challenging.

FreeCell's solvability statistics come from decades of computational research:

• The 32,000 Microsoft deals: The original Microsoft FreeCell game shipped with 32,000 numbered deals. By 1995, internet communities had solved all of them except Deal #11982, which was eventually proven unsolvable through exhaustive computer search. Later, Deal #11982 was confirmed by multiple independent solvers.
• The 1,000,000 deal analysis: When the deal set was extended to 1,000,000, eight additional unsolvable deals were found: #146692, #186216, #455889, #495505, #512118, #517776, #781948, and #875865. That's 8 out of 968,000 additional deals — an unsolvability rate of roughly 0.0008%.
• Billions of random deals: Large-scale random sampling studies have tested billions of deals, consistently finding unsolvability rates around 1 in 78,000 (approximately 0.0013%). The variation from the Microsoft set rate likely reflects sampling differences in the deal generation algorithm.

A FreeCell game can be modeled as a directed graph where each node is a game state (the arrangement of all cards) and each edge is a legal move. A deal is solvable if there exists any path from the initial state to the solved state (all cards on foundations).

FreeCell's state space is enormous but well-connected. The combination of perfect information, free cells, and alternating-color building creates a graph where most nodes have many outgoing edges. The game's reversibility (you can undo most moves) also contributes: states that seem stuck can often be backed out of and approached differently.

Unsolvable deals occur when the initial card arrangement creates an unavoidable circular dependency — a set of cards that mutually block each other with no way to break the cycle even using all four free cells. Because alternating-color building provides so many valid placements and free cells provide escape routes, these circular deadlocks are extremely rare.

FreeCell is a masterclass in game design balance. Paul Alfille's 1978 modification of Baker's Game achieved something remarkable: a game that is almost always winnable yet never trivial. The player always has agency (no luck), almost always has a path to victory (high solvability), and still needs to think carefully (genuine difficulty).

This balance is why FreeCell has endured for nearly 50 years. Losing feels fair because you can trace your mistakes. Winning feels earned because it required real thought. And the tiny chance of an unsolvable deal adds just enough uncertainty to keep things interesting — you can never be 100% sure a deal is impossible until you've tried everything.