Klondike Solitaire: The Complete Mastery Guide

Klondike is the game most people mean when they say "solitaire." It is the default that shipped with Windows for a generation, the patience that travelers played on flights before phones had browsers, and the one layout every card-curious child has dealt out at least once. That familiarity does something strange to the game: most players treat Klondike as luck with some card-moving on top. Win or lose, they shrug. They do not read up on it. They do not analyze the deal. They do not see a place where skill bites.

Part of that resignation comes from the game's own reputation. Klondike ships with every computer in the world and sits in every airport terminal. A thing that universal stops looking like a thing that rewards study. Chess rewards study. Klondike, the thinking goes, rewards the cards you were dealt. We hear the same claim about poker from people who do not play poker. In both cases the claim is wrong for the same reason: the short-term noise of random cards hides a signal that only shows up across a long series of hands.

We think that reading is wrong. Klondike has a real skill ceiling, and the gap between a casual player at 25–35 percent and a serious player at 60–70 percent is not narrowed by luck. It is narrowed by a set of specific habits: reading the tableau before moving, planning a few ply ahead, knowing when to cycle the stock and when to pause, protecting the foundations from premature sends, and recognizing dead positions a turn or two before they lock. The ceiling above serious play, toward 79–82 percent in Draw 1, is the solver ceiling. Humans do not reach it. But the climb from 30 to 70 is available to anyone willing to treat the game like a game with structure.

This guide is the long version of what we teach. It covers the layout logic, the opening priorities, the strategic difference between Draw 1 and Draw 3, the midgame choices that quietly decide most games, the specific problem of managing the waste pile in Draw 3, the endgame technique that separates 60-percent players from 90-percent players, and the ten mistakes we watch players make over and over. Nothing here is deep simulation theory — we save that for our Klondike probability page. This is the tactical playbook.

Klondike deals 28 cards into seven tableau columns of increasing length. Column 1 holds one face-up card. Column 2 holds two: one face-down, one face-up. Column 7 holds seven: six face-down, one face-up. The remaining 24 cards form the stock, and the four foundation slots above the tableau start empty. The foundations want suited ascending sequences from Ace to King. The tableau wants descending alternating-color sequences from King down to Ace. Empty columns accept Kings only.

That architecture is not arbitrary. It encodes the central puzzle of the game: you can only work with cards that are face-up, but most of the cards you need are face-down in the tableau or buried in the stock. The job of the game is to convert face-down cards into face-up cards while building toward foundations. Every move either reveals a new card, rearranges the face-up layer, or sends a card permanently home to the foundations. Moves that do none of these things are suspect.

The stock is the second half of the puzzle. It holds 24 cards you have not seen and cannot access freely. In Draw 1, you turn them over one at a time and access each card individually. In Draw 3, you turn them in groups of three and can play only the top card of each group, with the other two visible but locked. That single-rule difference is enormous, and we will come back to it repeatedly. The stock cycles as many times as you want, but each cycle has costs: in timed and scored modes it subtracts points, and in every mode it gives the opponent — the deal itself — another chance to see which cards you are keeping stranded.

The foundations reward patience. They are a one-way street. Every card you send to a foundation is gone from the tableau forever, and if you sent it too early you can no longer use it to build a sequence or to unblock a column. The classic Klondike mistake is a premature foundation send that locks up the tableau two moves later. We treat the foundations as the end state, not a running score — a place cards arrive when they are no longer useful below.

The first minute of a Klondike game does more work than any other minute. You are choosing which face-down cards to reveal first, which columns to attack, which sequences to begin building, and where on the board the early Kings will land. A strong opening is not a clever one; it is an unhurried one. Before moving, we read the whole row of seven face-up cards and the waste pile top. We ask four questions in order.

First, are any aces exposed? An ace is the only card we send to a foundation without hesitation. It never helps the tableau, and the sooner it leaves the sooner the foundation opens for the twos. Second, are any twos exposed and is the matching ace already up top? If yes, they go up as well. Third, what moves reveal a face-down card? A move that flips a new card is almost always better than a move that rearranges face-up cards without flipping anything. Fourth, which column has the most face-down cards? That column is the most expensive to open, so every move that digs into it is worth more than a move into a column already mostly exposed.

We read left to right, because the leftmost columns are the shallowest and typically open first. Column 1 has only the single face-up card beneath it — zero face-down blockers — so it becomes an empty column quickly if the card is playable. That empty column is a King slot. We prefer to fill empty columns with Kings that unlock a lot of face-down cards in a different column. A King sitting in the stock waste that has a red Queen already on the tableau, built by alternating color below, will bring the whole Queen-Jack-Ten run with it when we move the King into the empty slot. That is the move that quietly wins games.

Left-to-right is also how we think about sequence starts. An exposed black 7 opposite a red 8 in a neighboring column is a candidate join. We do not always make the join immediately. If joining them buries a card we will need later, or commits a 7 we might want on a different 8, we wait. In the first few moves we prefer joins that reveal face-down cards, not joins that merely tidy the face-up layer.

King placement deserves its own paragraph. Kings are the only card that can fill an empty column, which makes them both powerful and dangerous. A King placed too early in the wrong empty column can lock a sequence away from its natural home. We try not to commit a King to an empty column until we know what is riding on it. If we have a Queen of the opposite color ready to build, the King is productive. If we have no Queen yet and the other empty column is about to open, we wait. The worst King placement in Klondike is the reflex one, made because a column is empty and the King happened to be on top of the waste.

One more opening principle: the first stock cycle is a survey. In Draw 1 we go through the stock once without making aggressive plays, reading what is there. In Draw 3 we turn the stock and note which cards are accessible, which are buried, and whether the accessible ones are cards we actually want. Only after that survey do we make a plan. Players who do not survey end up reacting to the stock instead of using it.

One caution on the survey pass: do not play a card just because it is legal. In Draw 1 the temptation is to slap every playable card onto the tableau immediately and feel productive. That habit produces a cluttered tableau with no face-down reveals and a chewed-up stock. The survey pass is a reading exercise. We play only the moves we are certain of — aces, two-to-foundation sends when the ace is already up, and a single King that unlocks a clear chain. Everything else we hold until we have seen the full stock and can plan the next few moves together.

Draw 1 and Draw 3 are different games. That is not rhetorical. The solvability bounds differ by a few percentage points — roughly 82 percent in Draw 1 versus 78–82 percent in Draw 3 depending on the analysis — but the human win-rate gap is much larger. Good players win 60–70 percent of their Draw 1 games and 15–20 percent of their Draw 3 games. Expert players stretch Draw 1 toward 79–82 percent and Draw 3 toward 25–33 percent. The same deal, played under the same hands, wins in Draw 1 and loses in Draw 3 more often than not.

The gap is not about the cards. It is about stock access. In Draw 1, every card in the stock is reachable on every pass. The stock is essentially a queue you can search. In Draw 3, only one-third of the stock is reachable per pass, and every play shifts the cycle, rearranging which cards are reachable on the next pass. Draw 1 pacing rewards methodical planning: you see the problem, you plan the sequence, you execute. Draw 3 pacing rewards adaptive memory: you hold the cycle in your head, you know where the key cards are sitting, you wait for the window when the cycle puts the card you need at the top of its group.

Strategy changes accordingly. In Draw 1, we build foundations a little more aggressively because we know we can fetch a replacement if the tableau needs one. In Draw 3, we build foundations conservatively, because the replacement card might be trapped at position 1 or 2 in its group for three more passes. In Draw 1, empty columns are useful; in Draw 3, they are precious. In Draw 1, cycling the stock is cheap; in Draw 3, cycling without playing anything is a warning that our plan is not working.

There is a second, quieter difference between the modes: the psychology of cycling. In Draw 1, players feel in control because every card is reachable. In Draw 3, players feel at the mercy of the cycle because so much of the stock is locked. That psychology leaks into decisions. Draw 3 players sometimes rush the tableau to feel productive, sending cards to foundations or joining sequences early, because they cannot fix anything in the stock. Draw 1 players sometimes overplan, holding every card for a better home and never committing. We teach the same remedy to both: notice the feeling, name it, and then ignore it. The right move is the right move whether or not you feel in control.

Cycle management is the Draw 3 skill. Every stock play shifts the positions of every subsequent card. A good Draw 3 player knows this and exploits it: plays a blocker even when the landing spot is imperfect, because the shifted cycle will expose a more valuable card on the next pass. A bad Draw 3 player plays whatever is on top of the stock when it is useful and does not think about how the cycle will look after. We cover the full cycle-shift technique on our Draw 1 vs Draw 3 comparison. If you take only one thing from this section, take this: Draw 1 is a planning game, Draw 3 is a memory game, and the habits that win them are different.

The midgame is the stretch between the first few stock cycles and the endgame where the foundations start filling. Most games are decided here, and most players treat this stretch as the boring middle. We think the opposite: the midgame is where the game is won or thrown away. The key choices are about when to burn stock cards, how to protect the foundations, when to rebuild columns, and how to recover from positions that feel stuck.

Burning the stock means playing a stock card you do not need onto the tableau to expose the card beneath it (Draw 3) or to keep the stock from stalling (Draw 1). Burning is a trade: you accept a worse tableau in exchange for better stock access. We burn when the cost is small — the card lands on a column we are already planning to reorganize, or on a temporary sequence we can rebuild later. We do not burn when the burn buries a card we actively need. A burn that commits a red 4 onto a black 5 we were saving for a different red 4 is not a burn, it is a mistake.

Protecting the foundations is the discipline of not sending cards too early. The standard guidance is "keep foundations within two ranks of each other," and it is reasonable as a rule of thumb. The deeper principle is that every foundation send removes a card from the tableau's vocabulary. A red 6 on the foundation cannot host a black 5. A 4 that went up cannot carry a 3-2-A sequence back down. We hold cards that might be needed below. We send cards when their usefulness below has expired — when both opposite-color successors are already placed or the tableau no longer has a need for the sequence the card could anchor.

Column rebuilding is the midgame habit of taking a half-built column apart to rescue a buried card. This is painful and often correct. If the 10 of spades is buried under a 9 of hearts and 8 of clubs, and we need the 10 to host a red 9 that unlocks a much bigger chain, we may need to move the 9-and-8 elsewhere to reach the 10. The rebuilt column is worth the cost if the chain that comes free is long enough. We count moves before starting a rebuild: three moves to dismantle and one to reassemble only makes sense if the chain it unlocks pays more than four moves of progress.

One more midgame habit: protect the empty columns. Once a column empties, we try not to fill it casually. An empty column is a temporary storage bay that lets us dismantle other columns safely. If we fill it with a King at the wrong moment, we lose that workspace. The rule we teach is: do not fill an empty column unless the King coming in drags a productive chain behind it, or unless leaving the column empty would mean losing the next move anyway. Holding an empty column for one extra turn, even when a King is sitting on the waste, is often correct.

Recovering from stuck positions is the midgame test. A position feels stuck when no legal move reveals a new card and the stock has no playable top. Often the move exists, and we just have not found it. The recovery routine is: survey every face-up card in every column, check the waste top, check if any foundation card could temporarily come back down (if rules permit), and check if a tableau-to-tableau move would unlock a downstream move two ply ahead. If none of those find a move, we cycle the stock one more pass with a specific card in mind. If that fails, we concede the deal and reset. A hand spent trying to force a dead position is a hand spent not learning why it died.

The waste pile is the graveyard of played stock cards, and in Draw 3 it becomes the pivotal structure of the game. Managing the waste in Draw 3 is a distinct skill from managing the tableau. We think about it as pattern recognition over the stock cycle, not as card-by-card play.

The cycle shifts every time a card is played from the waste to the tableau. If the original stock order had the key card at position 2 of group 4 — meaning it was the second of the three cards in the fourth turn — then after playing a single stock card the key card shifts forward by one. After playing two, it shifts by two. Experienced Draw 3 players track this in their head. They know that if they play one stock card, the previously buried card will come to the accessible position on the next pass. That is not luck; that is arithmetic.

The mistake pattern is playing stock cards without thinking about what they do to the cycle. A player sees the 5 of hearts on top of the stock, has nowhere useful to place it, and plays it anyway because it is legal. The 5 lands on the tableau, burying a card, and the next stock turn brings up a 9 of clubs they did not need. Two moves later the 5 is in the way and the cycle has shifted in a direction that buries the ace they were chasing.

The correct pattern is to read the stock as a sequence of three-card windows, identify which windows contain the cards we need, and plan the cycle shifts that will bring those cards to the top of their windows. If the 7 of spades is at position 1 of a group (buried under two cards), we need two cards to come off the top of previous groups to shift the 7 forward. That means finding two useful plays from earlier in the cycle before the group containing the 7, even if those plays are suboptimal in themselves. The payoff is that the 7 then sits at the accessible position on the next pass.

In Draw 1, the waste pile is much simpler — every card is eventually reachable — but it still carries a lesson. The order in which stock cards cycle back is the order in which they were buried, so playing them in a different order on the first pass changes which card sits at the top when we need it. We still think about the cycle in Draw 1, but the cost of getting it wrong is small because every card comes back in a single pass.

The endgame in Klondike starts when the foundations hold enough cards that the remaining tableau is a short countdown. The tableau at this point usually has no more face-down cards, the stock is either exhausted or predictable, and the game is effectively a sequencing puzzle: in what order do we send the remaining cards to the foundations so that no column blocks another?

Foundation ordering is the skill. A naive endgame sends every card to the first foundation that accepts it. A careful endgame holds cards when sending them would break a tableau sequence we still need. If we have a black 5 on a red 6, and the red 6 is the only place the black 5 can sit, sending the 5 to a foundation is fine — it is done with its tableau job. If the 5 is still carrying a 4 and a 3, we leave it in place and finish the 4 and 3 first, so the 5 goes up with nothing riding on it.

Move-counting matters in the endgame. We count the shortest sequence of moves that finishes the game and compare it to the sequence we are playing. If our plan takes 20 moves and a shorter plan exists, we adopt the shorter plan — not because the game cares about the number, but because the shorter plan usually has fewer branch points where we could make a mistake. Shorter plans are safer plans.

Another endgame habit is holding the last few tableau cards in a known order. As cards go up, the tableau becomes a countdown and we can often read the entire remaining path. At that point we stop reacting and start executing. If the final eight cards will go up in a specific sequence, we play that sequence without hesitation. If the sequence has a fork — two possible orders — we pick the one with fewer branches and finish. Endgames that drag on are usually endgames where the player stopped planning and started reacting to each card as it came.

Knowing when we have won or lost is the last endgame skill. A game is won the moment every face-up card has a clear path home, regardless of how many moves remain. A game is lost the moment a required card is buried under a card that cannot move anywhere — a dead-card configuration that will not resolve. Recognizing these moments early saves time. In particular, recognizing losses saves the frustration of playing a dead position for another five minutes. The sooner we see a loss, the sooner the next deal starts.

These are the mistakes we see most often when we teach Klondike. Every one of them is fixable, and fixing any of them typically adds a few percentage points to a player's win rate.

• Premature foundation sends. Sending a 4 to the foundation when the opposite-color 3 is still needed below. The foundation is a one-way street; do not hurry.
• Reflex King placement. Dropping the first King you see into the first empty column. The King stays there forever; wait for the King that unlocks the longest chain.
• Chasing visible aces. Building the tableau around revealing an ace that is deep-buried, when two other aces are available at smaller cost. The column with more face-down cards is usually not the best target.
• Stock-cycling without a plan. Cycling the stock and playing whichever cards are legal, instead of deciding first which card you are fishing for.
• Ignoring Draw 3 cycle shifts. Playing stock cards without tracking how the cycle changes. The cycle is predictable; ignoring it is choosing to lose information.
• Joining sequences too early. Dropping a 7 onto an 8 when you might want that 7 on a different 8 two moves later. Early joins can lock in suboptimal structure.
• Leaving face-down cards face-down. Preferring face-up reorganizing to face-down reveals. The game is won by revealing cards, not by tidying.
• Rebuilding for no reason. Dismantling a column to reach a buried card that, once free, does not open anything useful. Count the payoff before you pay the cost.
• Forcing dead positions. Playing a locked position for another five minutes in the hope something shakes loose. It will not. Reset.
• Playing at the wrong tempo. Racing through moves when the position needs thought, or stalling when the position is clear. Tempo is a skill; we return to it below.

We have watched a lot of Klondike hands, and one of the clearest predictors of win rate is tempo. Players who race through moves lose to players who pause and read. Players who freeze on every decision fail to develop pattern recognition. The winning tempo is neither fast nor slow; it is responsive to the position.

Fast is correct when the tableau is simple and the next three or four moves are forced. Stock top is an ace; the ace goes up. Stock top is a red 9 and a black 10 is exposed; the 9 goes to the 10. Fast play here is good play, because there is nothing to think about and thinking would only waste attention. The danger of fast play is when the position stops being simple and the tempo does not reset. Players blow good midgames by continuing to play at opening-tempo speed when the position has become subtle.

Slow is correct at the branch points: the first King placement, the decision to send the first 2 to a foundation, the decision to start a stock cycle that will burn cards. These are the moves that echo. We spend a few seconds reading the full board before committing, and we ask whether the move closes any doors we might want open. That small pause is the single biggest tempo correction we give players.

Timed play changes tempo but not the fundamentals. In a timed session we still pause at branch points; we just pause less. Players who race for a clock time almost always lose points to mistakes they would not have made in untimed play. The experienced approach is to decide up front whether the session is for speed or accuracy and to play the whole hand at one tempo. Switching tempos mid-game is where the real losses hide.

Focus is the other half of tempo. Klondike is a long game for a 52-card deck, and concentration slips. A player who was doing well at move 15 can lose at move 40 simply by losing focus. We treat games as self-contained: sit down, play one deal with full attention, finish, take a short break, start another. Marathon sessions with half-attention produce marathon losses. Short focused sessions produce the improvement graph we want.