The Mathematics of Solitaire: Probability, Odds, and the Numbers Behind Every Deal

A Number Larger Than the Universe

Shuffle a standard 52-card deck and deal a game of Klondike Solitaire. The specific arrangement of cards you are looking at has almost certainly never existed before in the history of humanity, and almost certainly never will again.

That sounds like an exaggeration, but it is a mathematical fact. The number of ways to arrange 52 cards is 52 factorial, written as 52!, which means 52 × 51 × 50 × 49 × ... × 2 × 1. The result is approximately 8.07 × 10^67 — an 8 followed by 67 zeros.

To put that in perspective: scientists estimate the observable universe contains roughly 10^80 atoms. If every atom in the universe dealt one unique solitaire game every second since the Big Bang (about 13.8 billion years ago), they still would not have covered even a tiny fraction of all possible arrangements. The total number of seconds since the Big Bang is only about 4.3 × 10^17. Multiply that by the number of atoms, and you get roughly 4.3 × 10^97 deals — impressively large, but the number of possible deals is so astronomically bigger that each deal you play is, for all practical purposes, one of a kind.

Conditional Probability: How Solitaire Trains Your Brain

Every time you flip a card in solitaire, you are doing probability calculations in your head — whether you realize it or not. This is conditional probability at work: updating what you know based on new information.

Say you are playing Klondike and you need a red 7 to place on a black 8. There are four 7s in a standard deck — two red (Hearts and Diamonds) and two black. If you have already seen one red 7 in the tableau, there is exactly one left somewhere in the remaining unseen cards. The probability of it being the next card you draw from the stock depends entirely on how many unseen cards remain.

If there are 20 unseen cards left and one red 7 among them, your probability of drawing it next is 1 in 20, or 5%. If you draw three times without seeing it, the probability updates: now it is 1 in 17, or about 5.9%. This constant mental updating — adjusting probabilities as you gain information — is the same mathematical framework used in fields from medical diagnosis to weather forecasting.

Expert solitaire players do this instinctively. They track which cards have appeared, estimate the likelihood of useful cards appearing, and make decisions based on those evolving probabilities. It is Bayesian reasoning applied to a card game.

The Winnability Problem: An Embarrassment of Mathematics

One of the most fascinating mathematical questions about solitaire is deceptively simple: what percentage of Klondike deals are actually winnable?

Despite decades of research and millions of computer-simulated games, mathematicians still do not have a precise answer. The problem has been called "one of the embarrassments of applied mathematics" because the game is so well-known yet so resistant to exact analysis.

The best estimates come from a variant called Thoughtful Klondike, where all cards are visible from the start (face-up in their dealt positions). This removes the hidden-information problem and turns the game into a pure logic puzzle. Large-scale computer analyses of Thoughtful Klondike with Draw Three rules have produced a winnability estimate of approximately 82%, with a margin of error under 0.1%.

Standard Klondike, where tableau cards start face-down, is harder to analyze because the hidden information creates a partially observable system. The best estimates suggest about 79-82% of deals are theoretically winnable with perfect play for Draw One, and roughly 75-79% for Draw Three. But "perfect play" is an important caveat — it assumes you always make the optimal move, sometimes requiring dozens of look-ahead steps that no human player can realistically compute.

Win Probability by Variant: A Mathematical Spectrum

Different solitaire variants span an extraordinary range of mathematical winnability, from near-certainty to near-impossibility.

FreeCell sits at one extreme. Because all 52 cards are visible from the start, the game is one of complete information — like chess, every relevant fact is visible to the player. Of the original 32,000 numbered FreeCell deals first cataloged by Microsoft, only one (deal #11982) was proven unsolvable. Across the full space of random deals, the winnability rate is approximately 99.999%. In mathematical terms, FreeCell is almost a pure puzzle, and losing almost always means the player made a mistake.

At the opposite extreme sits Pyramid Solitaire, with a win rate estimated at just 2-5%. The card-pairing mechanic (removing pairs that sum to 13) creates rigid constraints that most random deals simply cannot satisfy. A King is worth 13 on its own and can be removed alone, but every other card requires a specific partner. If that partner is buried behind other cards or eliminated too early, the game becomes unwinnable — often without the player realizing it until many moves later.

Spider Solitaire demonstrates how mathematical complexity scales with suit count. One-suit Spider is winnable approximately 99% of the time because every card shares the same suit, making sequence-building straightforward. Two-suit Spider drops to roughly 20-30%. Four-suit Spider, with its full combinatorial complexity, falls below 10%. The progression is not linear — each additional suit multiplies the constraints on valid moves, creating an exponential increase in difficulty.

The Supermove Formula: FreeCell's Hidden Equation

FreeCell contains one of the most elegant mathematical formulas in any card game. While the basic rules only allow moving one card at a time, skilled players know you can move entire sequences by using free cells and empty columns as temporary storage.

The maximum number of cards you can move in a single logical operation — called a supermove — follows a precise formula: (empty free cells + 1) × 2^(empty columns). With all four free cells empty and no empty columns, you can move 5 cards. Add one empty column and the number doubles to 10. Two empty columns: 20. Three empty columns: 40.

This formula emerges from a mathematical property: each empty column doubles your capacity because you can split a sequence, temporarily store half in the empty column, complete part of the move, then retrieve the stored cards. It is a practical application of the mathematical concept of recursion — breaking a large operation into smaller identical sub-operations.

Understanding this formula transforms how you play FreeCell. Protecting empty columns becomes as important as protecting free cells, sometimes more so, because the exponential term (2^columns) grows faster than the linear term (cells + 1).

Card Counting: The Mathematics of Memory

Solitaire rewards card counting — keeping a mental tally of which cards have been played and which remain unseen. This is the same mathematical skill that gives blackjack players an edge in casinos, applied to a single-player game.

In Klondike Draw Three, card counting is especially powerful. When you cycle through the stock pile, cards appear in groups of three. If you memorize the order, you can predict exactly which card will be available after each draw. This turns a game of partial information into something closer to complete information — dramatically improving your odds.

The mathematics is straightforward but the mental effort is significant. In a game with 24 stock cards, there are 24 ÷ 3 = 8 groups of three. Remembering the order of 8 groups and updating that mental model as you play cards from the waste is a working memory exercise that would make any cognitive psychologist nod approvingly.

In Spider Solitaire with two decks, card counting becomes even more relevant because duplicate cards exist. Knowing whether zero, one, or two copies of a specific card remain in the stock changes your strategic calculations entirely. If both 8s of Spades are accounted for, a sequence waiting for that card is permanently stuck.

The Dealing Problem: Why Fair Shuffling Matters

Mathematicians have proven that a standard riffle shuffle needs to be performed at least seven times to produce a statistically random arrangement of a 52-card deck. Fewer shuffles leave detectable patterns — cards that were adjacent before shuffling tend to remain near each other.

This matters for physical solitaire because an insufficiently shuffled deck produces non-random deals. If you collect cards in order after a completed game and only shuffle a few times, you will get deals with clustered sequences that are either much easier or much harder than a truly random game. Computer solitaire avoids this problem entirely by using pseudorandom number generators that produce uniform distributions.

Interestingly, daily challenge modes in online solitaire games use a different mathematical approach: deterministic seeding. A seed value (often derived from the date) is fed into an algorithm that always produces the same sequence of pseudorandom numbers. This guarantees that every player gets the identical deal, making fair competition possible. The deals feel random but are actually perfectly reproducible — a property that pure randomness, by definition, cannot have.

Game Trees and Computational Complexity

Every solitaire game can be represented as a mathematical structure called a game tree. The root node is the initial deal, each branch represents a possible move, and the leaves are either wins or dead ends.

The game trees for solitaire variants are enormous. A typical Klondike game might involve 100 or more decisions, each with several possible moves. The total number of paths through the tree can easily exceed billions. This is why solving Klondike computationally is so difficult — even modern computers cannot exhaustively search the entire tree for most deals.

Researchers have shown that generalized Klondike (extended to arbitrary deck sizes) is NP-complete, placing it in the same mathematical category as some of the hardest problems in computer science. This means there is no known efficient algorithm that can determine whether an arbitrary Klondike deal is winnable. Every known approach requires, in the worst case, exploring an exponential number of possibilities.

FreeCell, by contrast, is more tractable because perfect information eliminates much of the branching. Solvers can prune the game tree aggressively by recognizing dead-end states early, which is why automated FreeCell solvers exist and perform well.

Expected Value: When to Take Risks

Every move in solitaire has an expected value — the average benefit (or cost) you can expect from making that move, considering all possible outcomes. Good players maximize expected value instinctively.

Consider a common Klondike scenario: you can move a card to reveal a face-down tableau card, or you can play a card to the foundation. Moving to the foundation is safe and makes progress, but revealing the unknown card might open up multiple new options. The expected value of revealing the card depends on how many helpful cards remain unseen and how critical the current board state is.

In mathematical terms, if there are 15 unseen cards and 4 of them would significantly help your position, the probability of a helpful reveal is 4/15, or about 27%. If a helpful card would be worth roughly 3 moves of progress and an unhelpful card costs nothing (you simply did not gain), the expected value of the reveal is 0.27 × 3 = 0.81 moves of progress. Compare that to the certain but fixed value of the foundation play, and you have a rational basis for your decision.

Expert players develop an intuition for these calculations over thousands of games. They may not think in explicit probabilities, but their decision-making patterns align closely with what the mathematics prescribes.

Streaks and the Gambler's Fallacy

Losing five solitaire games in a row can feel like the deck is stacked against you. Winning three in a row can feel like you are on a hot streak. Both feelings are mathematically misleading.

If your true win rate for Klondike is 25%, the probability of losing five games in a row is (0.75)^5 = 23.7%. That means roughly one in four stretches of five games will be all losses. Losing streaks of this length are not just possible — they are expected and mathematically inevitable over any significant number of games.

Similarly, the probability of winning three in a row at a 25% win rate is (0.25)^3 = 1.6%. It happens, just rarely. When it does, it feels special — but it does not mean you have suddenly become a better player or that the random number generator is favoring you.

Each deal is independent. The cards do not remember your previous game. This is the core of the gambler's fallacy: the belief that past outcomes influence future probabilities in independent events. Understanding this helps you approach each new game with a fresh mind rather than the emotional weight of past results.

Open Mathematical Questions

Solitaire still poses unsolved mathematical problems that attract serious researchers.

The exact winnability of standard Klondike Solitaire remains unknown. We have strong estimates, but no proof. The hidden information created by face-down cards makes rigorous analysis extraordinarily difficult.

The optimal strategy for Klondike is also unknown. We know some principles — reveal face-down cards when possible, do not lock Aces under other cards — but a provably optimal algorithm does not exist.

For Spider Solitaire, the relationship between suit count and winnability has not been precisely characterized. We know the general trend (more suits equals lower win rates), but the exact mathematical function describing this relationship remains an open question.

These unsolved problems are not just academic curiosities. They connect to deep questions in computer science about complexity, search algorithms, and the boundary between what computers can and cannot efficiently solve.

---

Test the Odds Yourself

Put the mathematics to work — play for free in your browser and see how the numbers play out in practice:

• Play Klondike Solitaire
• Play Spider Solitaire
• Play FreeCell Solitaire
• Play Pyramid Solitaire
• Play Yukon Solitaire
• Play Golf Solitaire
• Play TriPeaks Solitaire
• Play Forty Thieves Solitaire