Use probability to make smarter solitaire decisions, improve moves and increase your win rate.
Every strategy decision in solitaire is, at its foundation, a probability decision: the player chooses between alternatives whose outcomes are uncertain because some cards are face-down, stock cards have not yet been drawn, or future deal triggers have not yet occurred. The player who understands solitaire probability does not know what the next card will be — that information is unavailable — but they understand how to use the probability structure of the card distribution to make choices that are statistically superior over a large number of comparable positions. This is the operational definition of solitaire probability strategy: not predicting individual outcomes, but making decisions that maximise the expected win rate across the population of positions that share the same observable features.
Every strategy decision in solitaire is, at its foundation, a probability decision: the player chooses between alternatives whose outcomes are uncertain because some cards are face-down, stock cards have not yet been drawn, or future deal triggers have not yet occurred. The player who understands solitaire probability does not know what the next card will be — that information is unavailable — but they understand how to use the probability structure of the card distribution to make choices that are statistically superior over a large number of comparable positions. This is the operational definition of solitaire probability strategy: not predicting individual outcomes, but making decisions that maximise the expected win rate across the population of positions that share the same observable features.
The distinction between strategic and casual play is largely a distinction in how probability is incorporated into decisions. Casual play uses a single-outcome evaluation: will this move work? If yes, make it. Strategic play uses a probability-weighted evaluation: across all possible states of the hidden information consistent with the current observable board, what is the expected outcome of each available move? The move that maximises expected outcome across that full distribution is the strategically correct choice, even if it sometimes produces worse immediate results than the single-outcome alternative on specific deals. This probabilistic thinking is not complex mathematics — it does not require calculating exact probabilities in real time — but it is a qualitatively different mode of evaluation that produces measurably better decisions across the full range of game states.
This article covers the core probability principles that should govern solitaire strategy, how those principles translate into specific decision habits for tableau management, stock timing, and advanced position evaluation, and which game formats most efficiently develop probabilistic thinking as a strategic skill.
Principle 1: Prefer moves that reveal information over moves that do not. Face-down cards in Klondike, Scorpion, and similar variants are the primary source of uncertainty. Each face-down card is drawn from the set of cards not yet visible, and that set has a specific probability distribution — some positions are more likely to contain high-value cards (Aces, low-ranked cards needed for foundation building) based on what has already been revealed. The strategic principle follows directly: a move that reveals a face-down card is almost always preferable to a move of equivalent tableau value that does not reveal any face-down card, because the revelation eliminates uncertainty and enables better-informed decisions on subsequent moves. The uncovering-first principle is the strategy equivalent of this probability principle: it maximises information gain per move, which maximises decision quality on all subsequent moves.
Principle 2: Evaluate moves by their expected value across hidden card distributions, not by their best-case outcome. The best-case evaluation of a move asks: if the next revealed card is the most helpful card it could be, does this move set up the best position? The expected-value evaluation asks: across the full distribution of possible next revealed cards, weighted by their probability, what is the average quality of the position this move produces? These two evaluations frequently disagree. A move that looks excellent if the next card is an Ace may leave the position worse than the alternative move in the majority of outcomes where the next card is not an Ace. The expected-value evaluation is always the correct one for maximising long-run win rate; the best-case evaluation produces systematic overestimation of the moves that happen to work spectacularly on cooperative deals and underestimation of their cost on the majority of deals where the hoped-for best case does not materialise.
Principle 3: Use the known card distribution to assess conditional probabilities. As a Klondike game progresses and more cards are revealed, the set of unrevealed cards shrinks and the conditional probability of the remaining face-down cards being specific values increases. At the start of a game, each unrevealed card has approximately a 1/52 probability of being any specific card. By mid-game, with 20 or more cards revealed, the conditional probability that a specific face-down card is, say, the Ace of Spades — given that the Ace of Spades has not yet been seen — is 1/(52 minus cards seen). This conditional updating is not usually computed explicitly during play, but the habit of mentally tracking which high-value cards have not yet appeared — and therefore remain possible in unrevealed positions — is one of the clearest markers of experienced play. A player who knows that two Aces have not yet appeared and two remain in the face-down tableau population will correctly prioritise uncovering moves that have the highest probability of reaching those Aces first.
Tableau management decisions have direct probability content that is often not made explicit. The choice between two equivalent-looking tableau moves — both legal, both adding a card to a sequence — is frequently a choice between two different probability profiles for the subsequent game state. Consider a Klondike position where a black 6 can be placed on either of two available red 7s. The two placements produce identical immediate board states except for which column each occupies. The probability-informed choice evaluates which placement leaves the other red 7 in a position where it can receive a black 6 from the face-down cards that are most likely to reveal one — and chooses the placement that maximises the probability of having a valid home for the next black 6 regardless of which column it comes from. This is an application of Principle 2: the expected value of the placement that preserves option flexibility exceeds the expected value of the placement that concentrates option dependency.
Foundation management has an explicit probability connection through the suit balance principle. The reason to keep all four foundations within two ranks of each other is not aesthetic — it is probabilistic. A foundation advanced far ahead of the others removes high-rank cards of the advanced suit from the tableau before they have served as build bases. The probability consequence: the lagging suits' high-rank cards have fewer available build bases, which reduces the probability that any given tableau move can place those cards productively. Foundation imbalance progressively narrows the probability distribution of useful moves — fewer moves are productive when the build base landscape is sparse — and narrow probability distributions of useful moves are the precursor to stuck positions. Maintaining balance maintains the breadth of the useful-move probability distribution throughout the endgame.
Empty column management is probability management in its most direct form. An empty column is a resource whose value is the set of moves it enables — the set of moves that can only be made with at least one empty column available. The probability-informed empty column habit: before filling an empty column, assess the probability that a higher-value use for it will arise within the next three to five moves. In the opening and mid-game, this probability is typically high — empty columns are needed for uncovering chains, sequence staging, and King placements — so they should be held. In the endgame, this probability decreases as the board thins — the remaining moves are more deterministic — so filling the column becomes less costly. The expert habit of preserving empty columns "as long as possible" is implicitly a probability judgment: the expected future value of the empty column exceeds the immediate value of filling it until the game state is sparse enough that the future value drops below the immediate value.
Stock timing is the most directly probability-sensitive decision in stock-bearing solitaire variants. The stock pile contains a subset of the full deck whose distribution is unknown — but not completely unknown. As cards are drawn from the stock and placed or discarded to the waste pile, the conditional distribution of remaining stock cards updates. A player who has drawn 20 cards from the stock without seeing an Ace knows that the probability of the remaining cards containing an Ace is higher (if Aces are still unaccounted for from the tableau) or confirmably zero (if all Aces are already visible or on the foundation). This conditional updating — mentally tracking which high-value cards remain in the stock — is the most powerful probability tool available in Klondike and similar variants.
The stock discipline principle — exhaust tableau moves before drawing — is a probability rule as much as a discipline rule. The probability content: a stock draw before tableau exhaustion wastes the probability information in the current tableau state. The tableau's current state, fully evaluated, contains moves whose probability of producing useful downstream positions is calculable from the visible cards. The stock draw's value is conditional on the current tableau's state — a stock card that would be highly valuable on a sparse tableau may be unplaceable on a crowded one. Evaluating the tableau fully before drawing ensures that the stock draw is made in the tableau state where the drawn card has the highest probability of being immediately useful, rather than in a partially-evaluated state where it may arrive without a reception position.
In TriPeaks and Golf Solitaire, stock timing has a specific chain-probability content: before each stock draw, the player should evaluate the probability of the current visible tableau cards enabling further chain extensions after the draw. A stock draw that extends the current chain is always superior to one that terminates it. The chain probability evaluation — which rank-adjacent cards are visible, and which stock draw result would continue the chain versus break it — is the exact application of conditional probability to stock timing: given what is currently visible, what is the probability that each possible stock draw continues the chain, and does the tableau's current chain potential justify drawing now or first extending from visible cards?
Expert-level probability application in solitaire operates at two levels that casual and strategic players typically do not reach. The first is conditional move sequencing: evaluating the probability content of move sequences rather than individual moves. A two-move sequence has a probability distribution over outcomes that is not simply the product of the individual moves' probability distributions — because the first move changes the information state that determines the second move's probability content. Expert players who plan three to five moves ahead are implicitly computing the conditional probability distributions of positions at each planning horizon, choosing the sequence whose expected terminal position value is highest across the full distribution of hidden card states. This is the probabilistic version of the sequencing principles described in the card sequencing guide: the correct sequence is not just the one that looks best in the best-case hidden card state, but the one that looks best in expectation across all possible hidden card states.
The second expert-level application is positional probability assessment: estimating the probability that the current position is winnable, given the current observable board state and the known distribution of hidden cards. This assessment is not a precise calculation — exact winnability probability estimation requires solving the game from every possible hidden card arrangement, which is computationally intractable in real time — but experienced players develop calibrated intuitions about position winnability based on the observable structural features of the current board and the proportion of similar positions that have been won in past play. A position with four accessible Aces, two empty columns, and a partially cycled stock containing known-useful cards has a high estimated winnability probability; a position with all Aces face-down, no empty columns, and a depleted stock has a low estimated winnability probability. Acting on these calibrated estimates — investing more evaluation effort in high-probability positions and resigning low-probability positions after the three-pattern diagnostic check — is the expert application of probability that distinguishes position triage from both premature resignation and wasted effort on unwinnable positions. For the framework on distinguishing genuinely unwinnable from difficult positions, see our unwinnable deals guide.
TriPeaks and Golf Solitaire are the most efficient formats for developing probabilistic thinking as a strategy habit because their chain structure makes the probability evaluation immediate and binary: does this move continue the chain (good) or terminate it (bad)? Before each stock draw in TriPeaks, the player who asks "which visible card gives me the highest probability of chain continuation after this draw?" is practising exactly the conditional probability evaluation that transfers to stock timing in Klondike and Spider. The feedback is fast — chain continuation or termination is visible within one move — which accelerates the habit formation cycle relative to the longer feedback loops in Klondike and FreeCell.
FreeCell is the best format for practising the information-revelation principle (Principle 1) because its complete information makes the probability of each move's outcome fully calculable — there is no hidden card uncertainty to confound the evaluation. A FreeCell player who consistently asks "which move reveals the most useful information?" is asking the wrong question (all information is already visible in FreeCell) and should instead ask "which move produces the best expected position across the full move tree?" This reframing — from information-revelation to full expected value evaluation — is the progression from strategic to expert-level probability application, and FreeCell's complete information makes the full expected value calculation the only relevant probability question, giving the player direct practice at the most advanced level. For the complete shuffle and deal distribution context that underlies all probability calculations in online solitaire, see our shuffle randomness guide.
What is the best probability-based strategy for solitaire?The three core probability principles — prefer information-revealing moves, evaluate by expected value across hidden card distributions rather than best-case outcome, and use conditional probability updating to track high-value unrevealed cards — apply across all hidden-information solitaire variants and produce the largest improvements on the most common decision types. The single most impactful probability habit for players new to probabilistic thinking is Principle 1: consistently choosing moves that reveal face-down cards over equivalent moves that do not. This habit requires no probability calculations — it is a decision rule with a clear binary application — and it directly improves the information state that all subsequent decisions depend on. Combined with the stock timing discipline (draw only after tableau exhaustion) and the foundation balance principle, the three probability principles cover the main decision points where probabilistic thinking produces the largest win rate improvement relative to single-outcome evaluation.Which solitaire game best develops probability thinking skills?TriPeaks develops chain-probability evaluation fastest because its immediate feedback loop — chain continues or terminates — gives the player a direct signal on each probability decision within a 3–8 minute game. Golf Solitaire develops the same skill in a scored format, adding the probability question of score-target calibration: given the current chain state, what is the expected score, and does the expected score justify drawing now or waiting for a better chain position? FreeCell develops full expected-value evaluation — the most advanced probability application — because its complete information removes the information-uncertainty component of probability assessment and requires the player to evaluate all outcomes deterministically rather than probabilistically. The progression from TriPeaks and Golf (chain probability) through Klondike (conditional updating of hidden card distributions) to FreeCell (full expected value tree evaluation) traces the complete development path from basic to expert probabilistic solitaire thinking.Can every solitaire game be won by applying probability strategy correctly?No. Probability strategy maximises win rate on the winnable deal population by ensuring that each decision is made with the best possible use of available information about the hidden card distribution. It cannot convert unwinnable deals into winnable ones — unwinnable deals by definition have no legal move sequence leading to the win condition, regardless of how well the probability of each move's outcome is evaluated. What probability strategy does is increase the proportion of winnable deals that the player correctly identifies as winnable and successfully navigates to the win condition, and decrease the proportion of winnable deals that the player incorrectly abandons as apparently unwinnable or strategically misplays to stuck positions. The combined effect — more winnable deals correctly won, fewer winnable deals lost to strategy errors — is the full quantitative impact of probability-based strategy on observed win rates.
To improve your decision-making in solitaire using probability, start by assessing the likelihood of drawing specific cards from the stock pile. Keep track of which cards are already in play and which remain hidden. This will help you make informed choices about whether to move cards from the tableau to the foundation or to draw from the stock. Additionally, consider the potential outcomes of each move—if moving a card opens up more options or helps you uncover face-down cards, it may be worth the risk. Regularly practicing these assessments will enhance your strategic thinking over time.
Core strategies based on probability in solitaire include prioritizing moves that maximize your chances of uncovering face-down cards and maintaining a balanced tableau. Focus on moving cards to the foundation when it opens up new moves or reveals hidden cards. Additionally, avoid unnecessary moves that could block potential plays. Always consider the probability of drawing a needed card from the stock pile before making a decision. Lastly, practice patience; sometimes the best move is to wait for a more favorable situation rather than rushing into a play that may limit your options.
The best time to draw from the stock pile in solitaire is when you have exhausted all possible moves on the tableau and foundation. Before drawing, evaluate the current state of your tableau—if there are multiple face-down cards, drawing may help reveal them. Additionally, consider the cards you need; if you know certain cards are still in the stock, drawing can increase your chances of getting them. However, be cautious; if drawing from the stock pile will lead to a situation where you have no moves left, it may be wiser to hold off until you can make a more strategic play.