By Online Solitaire Free Editorial Team · Editorial Standards
Every solitaire game begins with a shuffle — a random permutation of a standard deck of 52 cards or, in double-deck games, 104 cards. The number of possible distinct orderings of a 52-card deck is 52 factorial (52!), a number so astronomically large that every shuffled deck in history is practically unique. Understanding the probability behind solitaire deals reveals which outcomes depend on luck and which on strategy.
Every solitaire game begins with a shuffle — a random permutation of a standard deck of 52 cards or, in double-deck games, 104 cards. The number of possible distinct orderings of a 52-card deck is 52 factorial (written 52!), a number with 68 digits. To put that in perspective: if every person who has ever lived had been shuffling a deck every second since the beginning of the universe, they would not have come close to exhausting the possible arrangements. This means that every solitaire game is, in a practical sense, a unique deal — the probability that any two shuffled games in history have begun with identical card arrangements is negligibly small.
This astronomical number of possible deals is the foundation of all solitaire probability analysis. The core question in solitaire probability is not what any individual card will be — once a deck is shuffled and dealt, every card's position is fixed — but rather: given the random process of shuffling, what proportion of all possible deals are winnable, and what proportion are unwinnable regardless of strategy? The answer to this question, which differs significantly across variants, is the starting point for understanding why win rates differ so dramatically between games and what the maximum achievable win rate for any given variant actually is.
Solitaire probability analysis uses three main mathematical tools. Combinatorics counts the number of ways cards can be arranged to produce specific outcomes — for example, the number of deals in which all four Aces are buried under face-down cards in the deepest tableau positions, creating an immediately stuck game. Monte Carlo simulation runs millions of random deals through automated solvers to estimate win rates empirically when exact combinatorial calculation is intractable. Game tree analysis maps all possible legal move sequences from a starting position to determine whether at least one path leads to a win — the computational approach used to establish FreeCell's near-universal solvability. Each tool is suited to different questions and different game structures.
The relationship between probability and outcomes in solitaire has two components that are worth separating carefully: the probability of receiving a winnable deal, and the probability of winning a winnable deal given the player's strategy. These two probabilities multiply to produce the observed win rate, but they are driven by completely different factors and respond to completely different interventions.
The probability of a winnable deal is determined entirely by the game's mathematical structure — the rules, the layout, and the number of cards. No player action affects this probability. In FreeCell, approximately 99.999% of all possible deals are winnable — only deal numbers 11982, 146692, and fewer than ten others in the first million deals are mathematically unwinnable. In Klondike Turn 1, the estimated unwinnable deal rate is 9–21% — for every five games played, roughly one will be unwinnable regardless of strategy. In Forty Thieves with one stock pass, the estimated unwinnable rate is 40–60% — the majority of sessions end in losses that no strategy could have prevented. These unwinnable deal rates are properties of the game design and are not adjustable by the player.
The probability of winning a winnable deal is determined by the player's strategy. In FreeCell, where nearly all deals are winnable, a player who never pre-plans their moves might win 50–60% of games, while a strategic player wins 80–90%. The gap between these rates — 30 percentage points — is entirely strategy-driven and represents the maximum improvement available through better play. In Klondike Turn 1, where the winnable deal rate is roughly 80–91%, a casual player might win 15–25% of all games while a strategic player might win 35–45%. The difference between the casual player's rate and the strategic ceiling reflects both the strategy gap and the unwinnable deal floor.
The practical implication: improving strategy improves win rate up to the limit set by the winnable deal proportion, but cannot exceed it. A Klondike Turn 1 player whose win rate is 40% is already achieving a substantial fraction of the theoretically possible maximum given that perhaps 80–91% of deals are winnable. A FreeCell player whose win rate is 50% has large room for improvement before approaching the 99.999% solvability ceiling. Understanding which component of the win rate gap is strategy-driven and which is deal-luck-driven is the most important use of solitaire probability analysis for practical play improvement.
Online solitaire platforms use pseudorandom number generators (PRNGs) to simulate card shuffling. A PRNG produces a deterministic sequence of numbers that is statistically indistinguishable from true randomness for practical purposes — the sequence is long enough and irregular enough that the deal produced by any single PRNG output looks random. The PRNG is seeded with a value (the seed) that determines the specific deal generated; two games with the same seed produce the same deal. This is how FreeCell's numbered deal system works: deal number 1 always produces the same 52-card arrangement, deal number 2 another specific arrangement, and so on.
The practical consequence of PRNG shuffling for players is that online solitaire deals are fair — the probability of receiving any specific card arrangement is effectively uniform across all possible arrangements, and the distribution of winnable and unwinnable deals in online play matches the theoretical distribution calculated from the game's mathematical structure. There is no mechanism by which an online platform produces systematically harder or easier deals than the mathematical baseline; the PRNG produces the same statistical profile as a fair physical shuffle.
Card distribution probabilities in a freshly shuffled deck are worth knowing for specific planning situations. The probability that a specific card (say, the Ace of spades) occupies any given position in a 52-card deck is 1/52, approximately 1.9%. The probability that all four Aces are among the first 28 cards dealt into the Klondike tableau (rather than in the 24-card stock) is approximately 15% — meaning in about one in seven games, all four Aces are immediately accessible somewhere in the tableau, creating a structurally favourable opening. The probability that the top card of the Klondike stock in a Turn 1 game is the specific card needed for a foundation play is roughly 1/24 for each individual draw — low enough that stock discipline (not drawing until the tableau is exhausted) is the correct policy even when the player "feels" that the needed card should come soon.
Monte Carlo simulation for win rate estimation. The most widely used mathematical approach to solitaire probability is Monte Carlo simulation: generating millions of random deals, solving each with an automated strategy (typically either the best available heuristic solver or an exhaustive solver for simpler games), and recording the win rate across the sample. This approach produced the most reliable estimates of Klondike's win rate — various simulation studies have produced estimates ranging from 79–82% for optimal play across all winnable deals, with the full game win rate (including unwinnable deals) typically estimated at 35–45% under optimal strategy. The wide published range for Klondike reflects the genuine difficulty of determining whether a given deal is unwinnable versus not yet solved by the heuristic used — Klondike's hidden information makes exhaustive solving computationally expensive in a way that FreeCell's complete information does not.
Exhaustive game tree analysis for complete-information games. FreeCell's complete information makes exhaustive game tree analysis feasible for individual deals: a computer program can explore all possible legal move sequences from a given starting position and determine with certainty whether at least one sequence leads to a win. This analysis, applied systematically to FreeCell's numbered deals, established that deals 11982 and 146692 are provably unwinnable (among the first 32,000 deals) and that the remaining deals are all solvable. No comparable exhaustive analysis has been completed for Klondike, because the hidden information and the possibility of multiple stock passes produce a game tree too large for current computational resources to search completely.
Combinatorial analysis of specific deal properties. Combinatorics can calculate the exact probability of specific deal properties without simulation. For example: the probability that the first three cards dealt to the Klondike stock include at least one Ace is approximately 1 − (48/52 × 47/51 × 46/50) ≈ 21%. The probability that a Pyramid deal places all four Jacks, Queens, and Kings in the bottom row (creating maximum blocking of lower cards) is (8/52) × (7/51) × (6/50) × (5/49) × (4/48) × (3/47) × (2/46) ≈ 0.0009% — effectively impossible, but the general principle that high-card clustering in the bottom rows creates hard deals is probabilistically significant across the full deal distribution. These specific probability calculations are useful for understanding which deal properties are rare versus common, and therefore which stuck positions should trigger deal resignation versus continued effort.
Win rate tables across the mainstream catalogue. The combined result of Monte Carlo simulation and exhaustive analysis across the patience family produces the win rate table that is the most practically useful output of solitaire probability research. The most reliable estimates for the major variants with strategic play are: FreeCell 80–90% (99.999% of deals winnable); TriPeaks 75–85%; Spider 1-Suit 60–70%; Scorpion 45–55%; Klondike Turn 1 40–45%; Pyramid 40–60%; Spider 2-Suit 40–50%; Yukon 35–45%; Klondike Turn 3 (strict) 20–25%; Spider 4-Suit 30–40%; Forty Thieves 20–35%. These ranges reflect both the strategy component and the deal luck component; the width of each range reflects the difficulty of separating the two precisely in games with hidden information.
Adjust play style to match the winnable deal proportion. The most important strategy adjustment driven by solitaire probability analysis is calibrating how much effort to invest in each game based on the proportion of unwinnable deals. In FreeCell, where nearly all deals are winnable, spending five minutes planning an apparently stuck position is justified — the stuck appearance is almost certainly a planning error rather than an unwinnable deal. In Forty Thieves, where 40–60% of deals are unwinnable, spending five minutes on an apparently stuck position is less often justified — the probability that the situation reflects an unwinnable deal rather than a planning error is meaningfully higher. Experienced players in high-unwinnable-rate games develop faster recognition of "deal loss patterns" — configurations that appear repeatedly in unwinnable deals and distinguish them from recoverable stuck positions.
Use early-deal card distribution to assess game quality. The first few moves of a solitaire game provide information about the deal's quality that probability analysis makes interpretable. In Klondike, a deal where the first several face-down cards uncovered are all of ranks already accessible in the tableau (not helping) can be provisionally assessed as lower quality; a deal where early uncovering reveals Aces or key cards for long buried columns can be provisionally assessed as higher quality. This early-game assessment is probabilistic rather than definitive — a low-quality-looking opening sometimes recovers, and a high-quality-looking opening sometimes collapses — but it informs how cautiously to play and how many stock passes to invest before considering resignation.
Foundation balance is a probability-driven strategy principle. The principle of keeping all four foundation suits within two to three ranks of each other is not arbitrary — it is probability-driven. If one suit's foundation is far ahead of the others, the cards of that suit that would normally contribute to tableau mobility (moving sequences, filling gaps) are instead on the foundation, while the other suits' cards continue accumulating in the tableau. Probabilistically, a balanced foundation state keeps more of the tableau's "useful" cards accessible for the longest portion of the mid-game. This principle applies across Klondike, FreeCell, Spider, and Forty Thieves with variant-appropriate modifications.
Stock discipline is a probability-driven efficiency principle. The instruction to exhaust tableau moves before drawing from the stock has a direct probability-based justification: each tableau move provides information (uncovering a face-down card reveals its identity) or creates flexibility (moving a sequence creates an empty column or a new build opportunity). Drawing from the stock without exhausting tableau options first wastes information-gathering opportunities and uses up the finite stock supply without extracting maximum value from the current board state. In games with limited stock passes (Klondike Turn 3, Forty Thieves), this probability-inefficiency is directly costly — wasted draws reduce the number of future options in a way that cannot be recovered.
For players interested in applying these probability principles through structured practice, our speed records guide covers how probabilistic deal assessment connects to fast play, and our relaxing solitaire guide covers how win rate probabilities map to casual player experience. For the TriPeaks and Golf games where probability and session scoring intersect most directly, the strategy sections of each game's guide cover the specific probability-informed decisions that improve chain identification and score reduction.
Three probability-informed habits improve win rates across all mainstream solitaire variants. First, exhaust all tableau moves before drawing from the stock — this maximises information extraction from the current board state and preserves the finite stock supply for when it is needed. Second, keep all foundation suits within two to three ranks of each other — this maintains tableau mobility for the maximum number of moves. Third, calibrate effort to the game's unwinnable deal rate — invest more sustained effort in high-solvability games like FreeCell (where a stuck position almost certainly reflects a recoverable planning error) and learn to recognise deal-loss patterns faster in high-unwinnable-rate games like Forty Thieves (where a stuck position has a meaningful probability of reflecting an unwinnable deal). The specific probability values that make these habits rational — FreeCell's 99.999% solvability, Forty Thieves' 40–60% unwinnable rate — are the reason the same habit (keep trying vs. recognise and resign) has different optimal forms in different games.
FreeCell offers the highest win probability in the mainstream catalogue on two dimensions: the winnable deal rate (approximately 99.999% of deals are mathematically solvable) and the strategic ceiling (80–90% with optimal play). These two dimensions are related: because nearly all deals are winnable, the gap between casual play and strategic play is mostly skill-driven rather than deal-luck-driven, which means strategy improvement has the clearest direct effect on win rate of any mainstream variant. TriPeaks offers the second-best probability experience for casual players at 75–85% — high win rate driven by a chain mechanic that resolves most of the tableau quickly on cooperative deals. Both games are available free at onlinesolitairefree.com, with win rate tracking that makes probability-based progress measurement immediate and automatic.
No. Every mainstream solitaire variant has a proportion of deals that are mathematically unwinnable — card arrangements from which no legal sequence of moves leads to the win condition, regardless of how perfectly the player plays. This proportion ranges from negligibly small (FreeCell: fewer than 0.001% of deals unwinnable) to majority (Forty Thieves: 40–60% of deals unwinnable). The mathematical proof that specific deals are unwinnable comes from exhaustive game tree analysis — demonstrating that the complete set of legal move sequences from a starting position contains no winning path. For games where exhaustive analysis is computationally intractable (Klondike, due to hidden information and large game trees), win rate estimates from Monte Carlo simulation provide the best available approximation. The practical recommendation: treat any loss in a high-solvability game (FreeCell, TriPeaks) as a likely skill loss and investigate what went wrong; treat any loss in a low-solvability game (Forty Thieves, Klondike Turn 3) as a possible deal loss and reserve extended post-game analysis for sessions where the stuck position appeared to have available moves that might have been missed.