Most people learned solitaire the same way, by clicking around a green screen at work until something moved. That gets you through an easy deal. It also means a lot of people have played a few thousand hands without ever being told the four or five habits that decide most of them. This post is the rules first, then those habits.
The setup
A standard 52-card deck, no jokers, dealt into three areas:
- The tableau. Seven columns, dealt left to right with one card in the first, two in the second, up to seven in the last. That's 28 cards. Only the last card of each column is face-up; everything beneath it is face-down and hidden.
- The stock. The remaining 24 cards, face-down, that you draw from when the tableau stalls.
- The foundations. Four empty piles, one per suit. This is where you're trying to get everything.
You win when all 52 cards sit on the foundations, each suit stacked ace through king.
The rules
- Foundations build up, by suit, starting with the ace. Ace of hearts, then the two of hearts, and so on to the king. An ace can go up the moment it's available.
- Tableau columns build down, in alternating colours. A red seven takes a black six. A black six takes a red five. Same colour never stacks.
- You can move a whole sequence at once. If a column ends in 9-8-7 correctly alternating, all three travel together onto a black or red ten as appropriate.
- Only a king can move into an empty column. Nothing else. Not a queen, not a sequence headed by anything but a king.
- Flipping is automatic and free. When you move the last face-up card off a column, the face-down card beneath it turns over.
- Cards move from the waste pile too, onto either the tableau or a foundation. Only the top card of the waste is available.
- A card on a foundation stays there. Some versions let you pull one back down into the tableau; this one doesn't, so a foundation move is a one-way door. Worth remembering two sections from now.
How this version is dealt
Variants differ on two points, and they change the game more than people expect. Klondike here draws one card at a time from the stock, and lets you cycle through it as many times as you like, with no penalty for recycling. That's the friendlier of the two common setups. Draw-three, where cards come up in packets and only every third one is reachable, is a substantially harder game.
Scoring is the familiar one: ten points for every card that reaches a foundation, five for flipping a face-down card, five for pulling a card off the waste into the tableau, and nothing at all for shuffling cards between tableau columns or recycling the stock. Hints are unlimited and each one costs a single point, a tiebreaker rather than a punishment. Notice what that scoring rewards, because it's a decent summary of good play: uncovering cards and banking them yourself, not rearranging what's already visible.
Face-down cards are the whole game
Here's the thing that reframes Klondike once you see it. You are not trying to move cards to the foundations. You are trying to turn over face-down cards, and the foundations are mostly a side effect.
Every face-down card is a card you can't plan around. A deal with twenty of them hidden is a deal you're playing half-blind. So when two moves are both legal and you can only make one, the tiebreak is almost always: which one exposes a face-down card? A move that reveals nothing has bought you nothing, however tidy it looks.
The refinement: when you have a choice of columns to dig into, go at the tallest stack of face-down cards first. Column seven starts with six hidden cards under its face-up one. Column two starts with one. Clearing column two is quick and gives you an empty slot; clearing column seven is where the deal is actually won or lost, and the longer you leave it the more likely you run out of moves before you get there.
Don't rush the aces up
This is the mistake that costs more deals than any other, and it feels like progress the whole time you're making it.
A card sitting on a foundation is out of play, permanently, since this version won't let you bring it back down. A three of clubs on the foundation cannot catch a red two. If you've sent every low card upstairs the moment it appeared, you'll hit a point where a red two is sitting in the waste with nowhere to go, and the black three that would have taken it is stacked under six cards of hearts.
Aces and twos are safe to send up more or less immediately, since nothing needs them. Past that, ask a question before every foundation move: could this card still be useful down here? If the opposite-colour card one rank below is still buried or unaccounted for, leave it in the tableau a while longer. It costs you nothing to wait. Foundations don't expire.
Empty columns are not free space
Clearing a whole column feels like an achievement, and then people fill it with the first king they see and wonder why the deal died anyway.
An empty column is the most flexible thing on the board. It can hold a king and the sequence built under it, which is the only way to relocate a king that's blocking a stack. It can also serve as a temporary parking bay while you unpick a sequence. The moment you drop a king into it, all of that is gone, and you've swapped your best tool for one specific card.
So before filling an empty column, look at which king you're committing to. The right king is one that either unblocks a column with face-down cards under it, or has a queen of the opposite colour available to continue the run. A king moved from the waste into an empty column, with nothing to build on it, has just cost you your flexibility for nothing.
Read the stock before you need it
With unlimited redeals it's tempting to treat the stock as infinite. It isn't, in the way that matters: cycling through it doesn't change the order. If a card you need is buried in the waste under cards you can't play, going around again puts it in exactly the same place.
Two habits help. First, on your first pass through the stock, pay attention to what's in there even when you're not playing it, because that's free information about which of your tableau plans are alive. Second, when you draw a card you can play in two places, remember that playing it changes what's underneath it in the waste for every future pass.
And a general one: if you cycle the entire stock without making a single move, the deal is over. Not lost by the rules, but you've proven there's nothing left. That's the moment to take the loss and deal again rather than clicking through another four laps.
The mistakes that lose deals
- Moving cards between columns because you can. A tableau-to-tableau move that reveals nothing and unblocks nothing scores zero points, and the scoring is right.
- Emptying the shortest column first because it's the easiest. Easy is not the same as useful.
- Sending everything to the foundations on sight. See above. This is the big one.
- Filling an empty column with whichever king is handy.
- Breaking up a long tableau sequence to move one card, when the sequence was the only thing keeping several cards accessible.
Every daily deal here is solvable
One thing worth knowing about the daily. Klondike deals are not all winnable, and in the general case a meaningful share of random shuffles cannot be beaten by anyone, however well they play. Losing to one of those and assuming you misplayed is a bad feedback loop for learning the game.
So the daily deal doesn't ship until a solver has actually solved it, and within a deliberately tight search budget rather than an exhaustive one. That second part matters more than the first: a deal that a solver only cracks after exploring an enormous tree is technically winnable and practically miserable. Every daily here has a solution a person can plausibly find.
Extra deals after the daily are random shuffles, unfiltered, which is the honest way to practise. Past dailies stay in the archive if you want to work back through deals that are known to be beatable.
If you'd rather have the guarantee on every single hand, FreeCell is the game to move to: nothing is hidden and virtually every deal is winnable. Spider goes the other way and is considerably harder than either.