The Odds of Winning Klondike Solitaire
Research shows about 90% of turn 1 Klondike deals and 82% of turn 3 deals are winnable with perfect play and full knowledge of the cards. Real players win far fewer: large-scale player data shows roughly 33% for turn 1 and 11% for turn 3. The gap comes from hidden information and everyday decision mistakes.
Two very different questions hide inside "what are the odds of winning solitaire," and most pages answer one while pretending it's the other. This page keeps them separate: how many deals can be won, and how many are won. The numbers are far apart, and the distance between them is the interesting part.
How many deals are winnable? The research
The strongest published answer comes from a University of St Andrews project called Solvitaire, by Charlie Blake and Ian Gent. Their program exhaustively searched millions of deals to determine, deal by deal, whether a winning line of moves existed. The results, published on arXiv and later in the Journal of Artificial Intelligence Research, put Klondike winnability at:
| Variant (unlimited redeals) | Winnable deals |
|---|---|
| Draw 1 | 90.480% (±0.116%) |
| Draw 3 | 81.945% (±0.084%) |
One huge caveat, which the authors are upfront about: these figures are for "thoughtful" solitaire, where the solver knows the rank and suit of every hidden card from the start. That's the only honest way to define winnability (either a winning sequence exists for a deal or it doesn't), but no human plays that game. You find out what's under a card by earning the flip.
Earlier work had sketched the same territory with wider error bars. A 2009 Oregon State study by Bjarnason, Fern and Tadepalli used Monte Carlo planning to bound thoughtful Klondike winnability years before Solvitaire pinned it down. The consistent finding across all of it: unwinnable deals are real, but they're a minority. Most of the time, the shuffle gives you a game you could win.
How many deals do people actually win?
Solitaired, one of the largest solitaire sites, published win rates from its player base: turn 1 games were won 33.0% of the time across about 2.9 million games, and turn 3 games 11.1% across about 1.4 million games.
Set those against the winnability figures and the picture is stark. At turn 3, perfect play with full information wins 82 games in 100. Actual humans win 11.
This site has no player numbers to add to that pile, on purpose: games on turn 1 and turn 3 keep your statistics in your own browser, on your own device, and nothing is sent to a server. We can cite Solitaired's data. We can't peek at yours.
Our solver's numbers
This site runs its own solver. It's the engine behind the hint button and the winnable-deals-only toggle, and we run it across a large fixed sample of deals to keep our claims honest. Across 1,000 deals per draw mode (seeds 1 through 1,000, so the run is exactly reproducible), it confirmed winning lines in at least 47.8% of draw 1 deals and at least 38.6% of draw 3 deals.
Side by side, the three kinds of numbers look like this:
| Measure | Draw 1 | Draw 3 |
|---|---|---|
| Winnable with perfect play (Solvitaire, near-exact) | 90.48% | 81.95% |
| Confirmed winnable by our solver (1,000 deals, budgeted search, lower bound) | at least 47.8% | at least 38.6% |
| Actually won by players (Solitaired, millions of games) | 33.0% | 11.1% |
Methodology notes, because a stats page without them is just vibes:
- Our solver, like every winnability solver, plays with full knowledge of hidden cards. Its numbers measure the deals, not human skill.
- It runs on a budget: 5,000 search steps per deal, the same budget the in-game hint button gets, so answers come back in a fraction of a second. A deal it can't crack within that budget counts as unproven, not unwinnable. That's why our figures say "at least": they can only undercount true winnability, never overcount it.
- The budget is also the whole reason our floor sits so far below Solvitaire's near-exact figures. Their searches ran for hours per stubborn deal; ours is tuned to answer while you're mid-game. The two sets of numbers don't disagree, they answer at different depths.
- These numbers come from a committed, re-runnable script against our real game engine, not from copying another site. When the engine changes, the numbers get regenerated.
Why the huge gap between winnable and won?
Three reasons, in decreasing order of excusability.
Hidden information is the unavoidable one. You must commit to moves before knowing what's under the face-down cards, and some winning lines require choices no human could justify with what's visible at the time. Part of the gap is simply unclosable, and anyone promising you a 90% win rate is selling something.
Decision mistakes are the closable part. Promoting cards to foundations too early, emptying columns without a king ready, ignoring the deep columns until too late: these show up over and over when winnable games die. The whole strategy guide exists because this slice of the gap is large and fixable.
Abandonment is the sneaky one. Players resign games that still had lines, sometimes several moves' worth of them. Every early redeal quietly converts a winnable deal into a loss in your stats.
Is solitaire ever a guaranteed win?
No version of random-deal Klondike guarantees a win. If you want the guarantee, that's what winnable-deals mode is for: our solver pre-checks the deal, so a winning line always exists, and finding it is on you. Purists consider that a different game. They're not wrong, but it's a genuinely good different game, and it turns every loss into a solvable puzzle.
The daily challenge sits in between: everyone in the world gets the same deal each day, winnable or not, which seems fairer somehow.
FAQ
What percentage of solitaire games are winnable?
For classic Klondike with unlimited redeals, peer-reviewed research from the Solvitaire project found 90.48% of draw 1 deals and 81.95% of draw 3 deals are winnable, assuming perfect play with full knowledge of hidden cards. Our own budgeted solver confirmed winning lines in at least 47.8% of draw 1 deals and 38.6% of draw 3 deals across 1,000 fixed deals per mode; it proves deals winnable, never unwinnable, so those figures are lower bounds.
What percentage of solitaire games do players actually win?
Published player data from millions of real games shows about 33% of turn 1 games and 11% of turn 3 games end in wins. Skilled players do considerably better. This site can't publish player rates of its own: your stats are stored on your device and never reach a server.
Is every game of solitaire winnable?
No. Roughly one in ten draw 1 deals and one in five draw 3 deals cannot be won no matter how well you play, because the cards you need stay mutually locked in every possible line. This is proven by exhaustive computer search, not just experience. Our winnable-deals mode filters those dead deals out.
Why do I lose solitaire games that were winnable?
Two reasons. You can't see the hidden cards, so some winning lines require moves no one could reasonably find. And ordinary decision errors, like promoting cards to foundations too early or wasting empty columns, kill winnable games constantly. The first cause is unavoidable. The second is exactly what the strategy guide targets.
Are the odds different for turn 1 and turn 3?
Yes, on both measures. In theory, 90.5% of turn 1 deals are winnable against 82% for turn 3. In practice the gap is wider: players win around three times as often at turn 1, since turn 3 hides two of every three stock cards behind timing decisions. See turn 1 vs turn 3 for the full comparison.