Most people play Klondike by taking whatever move appears. That's why most people win about a third of the time. Here's what the other players are doing.
Every face-down card is information you don't have, and Klondike is fundamentally a game about information. So when you have a choice between two legal moves, ask a single question: which one turns a card over? Take that one. Almost always.
And when both moves turn a card over, take the one from the longest pile — because that's the pile with the most cards still hidden, and the one that will otherwise strangle you at the end.
Beginners rush cards home. It feels like progress — the aces go up, then the twos, and the board looks tidier.
It is often a mistake. A two of hearts on the foundation is a two of hearts that can no longer hold a black ace on the tableau. Cards that have gone home can't come back. So the rule is: send a card home when it has stopped being useful, not when it first becomes eligible.
Aces and twos, fine, send them — they're rarely needed. But think twice before a five goes up while there are still black sixes waiting for a red five to sit on.
This catches everyone. You work hard, you empty a column, and you feel clever — and then you discover you have no king anywhere, and the empty column just sits there doing nothing while your options shrink.
Only a king may enter an empty column. So before you commit to emptying one, look at the board and ask where the king is coming from. If there isn't one, that empty column is a trophy, not a tool.
Moving a card off a nice descending sequence to make one small gain is usually a bad trade. Runs are what let you shift blocks of cards later. Fragment them early and you'll find, twenty moves on, that nothing can move anywhere.
On turn one with unlimited redeals, you can see every card in the stock eventually — so there's no rush, and there's no reason to take a stock card just because it's playable. If the tableau has a move that turns over a card, do that first. The stock isn't going anywhere.
Turn three is a different game entirely, and the discipline reverses: every card you play from the waste changes which cards you'll see on the next pass. Experienced turn-three players will deliberately skip a legal move to keep the rhythm of the pile intact. That's a genuinely different puzzle, and much harder.
Some Klondike deals cannot be won. Not by you, not by a computer with perfect knowledge, not by anyone. The card you need is under the card that needs it, and that's that.
Estimates put around 80% of turn-one deals as theoretically solvable with full knowledge of the face-down cards — which of course you don't have. Real players, playing well, win around a third. So if you're winning one in three, you are not doing badly. You are doing about as well as the game allows.
The table here says so out loud when a deal has genuinely died — because staring at a dead board wondering what you missed is not a game, it's a tax.
Around 80% of turn-one deals are theoretically solvable if you could see the face-down cards — which you can't. In practice, good players win roughly a third of the time. If that's your rate, you're playing well.
No. The foundation is a one-way door — a card sent home can't come back to hold something on the tableau. Aces and twos are safe. Anything higher, ask whether it's still useful where it is.
Because only a king may go into an empty column in Klondike. It's the rule that decides most games, and it means an empty column is worthless unless you have a king to put in it. Check where your king is coming from before you work to empty one.