Solitaire Strategy That's Actually Been Checked

That paragraph is the short version. The rest of this page is the long version, and it comes with a disclaimer you won't find on other strategy pages: this advice is checked against a computer solver, the same one that powers this site's hints and its winnable-deal mode. The solver checks whether a game can still be won from the current board. When a tip below says a move throws games away, that's not folklore. We can watch deals flip from winnable to dead the moment the move is made.

Solitaire strategy articles are usually a game of telephone, the same ten tips recycled between sites, with no way to test any of them. Some of the classic tips survive solver contact. A couple need caveats nobody mentions. Here's what holds up.

First, the frame: you're not unlucky

Exhaustive research puts turn 1 winnability at about 90%, and our own solver confirms winning lines in at least 47.8% of the deals it samples even under a tight search budget. Meanwhile, published data from millions of real games shows players winning about 33% of turn 1 games. That gap is the whole reason strategy is worth reading. The cards allow far more wins than players collect, which means the difference between you and a much better win rate is decisions, not shuffles. (Full numbers and sources are on the odds of winning page.)

Internalizing that changes how losses feel. A loss stops being bad luck to shrug off and becomes a puzzle with an answer: somewhere in that game, there was probably a moment where a different move kept the win alive.

The principles, in order

Order matters here. When two principles collide, and they will, the one higher on this list usually wins.

1. Aces and twos go up immediately

When an ace or a two becomes available, play it to the foundation. Don't overthink it.

Aces are obvious: they can't hold any card on the tableau (nothing builds onto an ace downward), so they do nothing in play. Twos are nearly as clear. A two can only hold an ace, and aces don't need holding, since they're always playable to the foundation.

This is the one foundation-greed exception. Every card three and up deserves a moment of thought before promotion. Aces and twos don't.

2. Flip face-down cards before anything else

A flip is the only move in Klondike that gives you new information and new material at once. When you're choosing between a move that flips a hidden card and one that just rearranges face-up cards, the flip wins almost every time.

The face-down cards are the game. All 21 of them have to turn over for you to win, and the ones you can't reach are the only thing that can make a game unwinnable. Face-up cards are never the problem; they can always be moved eventually. So measure every candidate move by one yardstick: does this bring a hidden card closer to flipping?

3. Attack the long columns first

Not all flips are equal. Column seven starts with six face-down cards; column one starts with none. Given a choice between flipping in a short column or a long one, work on the long one.

The reasoning is about time. The tall stacks need many moves across the whole game to excavate, so every turn you delay starting is a turn the game gets harder to finish. Short columns can wait; they're quick to clear whenever you get around to them. Solver traces of lost-but-winnable games show this pattern constantly: the player mined the easy shallow columns, sent some cards up, and hit a wall because a deep column never got touched until the resources to dig it out were gone.

One caveat the usual advice skips: an early flip in a short column that unblocks a long column beats a slow grind on the long column directly. The target is the deep dirt. Any route there counts.

4. Never empty a column without a king ready

An empty column is only worth what you can put in it, and the only thing you can put in it is a king. Clearing a column while no king is available wastes the moves you spent clearing it, and worse, the empty space sits there taunting you while kings stay buried.

Before you make the move that empties a column, ask: which king takes this spot, and can it get here now? If the answer is "none yet," leave the column alone and spend those moves elsewhere. The column isn't going anywhere.

The exception: if emptying the column is itself the flip that exposes cards you need (say, it's the only way to keep the game moving), take it. An imperfect empty column beats a dead game.

5. Choose your kings by their queens

When you do have an empty column and a choice of kings, don't grab whichever is closest. Look one step ahead: a king is only as good as the queen of the opposite color that will land on it.

A black king wants a red queen. If both red queens are hopelessly buried, that black king becomes a dead end that clogs your hard-won empty column. Check the queens, then the jacks if the game is tight. Thirty seconds of looking ahead here is worth more than anywhere else on the board, because empty columns are the scarcest resource in the game.

6. Hold foundation moves while the card can still work

Here's the tip that separates intermediate players from beginners, and the one the solver validates most dramatically. Past the twos, a card sent to the foundation stops working. A red five on the foundation can't catch the black four you flip two minutes later. That black four now needs the other red five, which might be under three face-down cards.

The question for every foundation move from three upward: could this card still be useful in play? Useful means it might need to hold a card of the rank below, or slot into a run you're building. When foundations run far ahead of the tableau, you strand cards. Solver checks confirm it regularly: a deal sits at "winnable," a mid-rank card goes up to the foundation, status flips to "lost." The move looked like progress. It was the losing move.

A workable rule of thumb: keep foundations loosely level, and don't push a suit more than a couple of ranks past what the tableau still needs. Most digital autoplay features (ours included) only auto-promote cards that are provably safe, which is a decent guide to what "safe" means.

If this costs you standard-scoring points in the short term, so be it. Points follow wins. The scoring guide breaks down the point values if you're curious what each move pays.

7. In draw 3, manage the stock like a resource

Everything above applies to both draw modes. Draw 3 adds a layer: only every third stock card surfaces, and the pattern only shifts when you play a card from the waste.

Practical habits that come out of this:

• Run through the entire stock once early, playing nothing, just to see what's there. You're allowed to look. Knowing that the two of hearts is sitting fourth from the top changes your whole plan. (In turn 1 this costs you nothing at all.)
• Before playing a waste card, remember it shifts every later group of three. Playing one card can surface cards you need next pass, or bury them. With practice you start noticing which.
• Sometimes play a card you don't want, purely to change the offset. This feels wrong the first dozen times. It wins games.
• If redeals are limited, as in Vegas rules, count your remaining passes and spend them like money. Our Vegas game will teach you this the hard way.

The deeper comparison of the two modes lives at turn 1 vs turn 3.

8. Move runs to release, not to tidy

Moving a run from one column to another achieves exactly nothing unless it flips a card, empties a column you can fill, or frees a card you need. A board can look beautifully organized and be no closer to won.

Partial-run moves are the version of this that most players underuse. You don't have to move a whole sequence. Splitting a run, taking just the seven-six off a black eight, say, to catch a red six elsewhere, is often the move that cracks a stuck board. The full rules for run moves are in the how to play guide.

9. Draw from the stock only when the board is done talking

Every draw is a small commitment, trivially small in turn 1, real in turn 3. Make it after you've exhausted the useful tableau moves, not before. The habit to build: when a new card appears (from a flip or the waste), scan the whole board again. New cards change what's possible everywhere, and the move you dismissed two minutes ago might work now.

Undo is a strategy tool, not a confession

Physical solitaire has no undo, so a lot of players carry a vague guilt about using it online. Drop the guilt. Undo has two legitimate jobs.

The first is tactical: undoing a draw or a move you made a second ago with new information. That's just correcting slips, and nobody's counting.

The second is where improvement actually comes from: post-mortems. When a game dies, rewind it. Walk back to the decision the game hinged on and play the other branch. Sometimes the game was lost from the deal, and that's worth knowing too. On this site you can also toggle winnable-deals-only mode, which guarantees a solution existed; in that mode, every loss is by definition a findable mistake, which makes the rewind habit brutally educational.

Players who rewind losses improve fast. Players who instantly redeal stay at the same win rate for years. The improvement stories we hear are from people who slowed down, not people who played more.

When to abandon a game

Knowing when a game is dead saves time, but calling games dead too early is a bigger leak. Genuinely check before you quit: every waste card against every column, every partial-run split, every foundation pull-down. Draw 3 players, remember that a full extra pass with one changed offset can transform the stock.

If you're in winnable-only mode, the answer is that it's never truly dead; there's a line you're not seeing, and the hint button (also solver-backed, it only suggests moves that keep the win alive) can show you the next step without solving the whole game for you.

In pure random deals, some games are simply lost from the shuffle. Roughly one deal in five can't be won at draw 3 even in theory. Abandoning those isn't weakness. The skill is telling them apart from the four in five that just look that way.

What the solver can't do for you

Honesty section. The solver knows every hidden card when it checks winnability, and you don't. Playing "perfectly" against unknown information is impossible even in principle, so nobody converts 90% winnability into a 90% win rate, and you should distrust any site whose strategy page implies you can. What the solver gives you is a floor for how much better your decisions can get, and proof about which specific habits burn winnable games. That turns out to be a lot. The rest is reps.