SolitaireSolitaire Win Rates: Winnable Deals vs. Games You Win
Not every solitaire deal is winnable. Published solver results describe whether a solution exists under particular rules; they do not predict how often you will win on your first attempt.
What does win rate mean?
Winnability asks whether at least one legal route to a complete game exists. A solver can search alternatives and, in thoughtful solitaire, knows the identities of hidden cards. Player win rate counts wins divided by attempts under stated rules. It depends on decisions, information and whether replay or undo is allowed.
Knowing every card in FreeCell helps planning, but does not make a human player equivalent to an exhaustive solver. Likewise, a failed Klondike attempt is not proof that its starting deal was impossible.
Published results and their limits
| Game and sample | Reported result | What it means |
|---|---|---|
| Thoughtful Klondike, research rules | 81.945% ± 0.084 percentage points | A winnability estimate with all card identities known; the authors report a 95% confidence interval. Not a first-attempt player success rate. |
| Standard FreeCell, original 32,000 Microsoft deals | 31,999 solvable; one unsolvable | Deal #11982 is the exception in this particular numbered collection. Other collections contain other impossible deals. |
| Games played on this website | No measured rate published here | These external results do not establish a win rate for our players or every rules option offered here. |
The Klondike figure comes from Blake and Gent's winnability research. The FreeCell result is documented by project participant Michael Keller in the Solitaire Laboratory FreeCell FAQ. Read our explanation of FreeCell #11982 for more context.
Why the rules matter
- Klondike: Draw 1 and Draw 3 expose the stock differently. Our version permits unlimited redeals and moving a top foundation card back to the tableau. Check the exact research rules before transferring an odds figure.
- Spider: one, two and four suits create different constraints. A percentage without the suit count is incomplete.
- Pyramid: stock passes and the winning condition matter. Our version allows two redeals; an estimate for a one-pass game is not directly comparable.
- Russian Solitaire: same-suit building restricts destinations compared with Yukon. That rule difference alone does not establish a numerical player win rate.
We do not rank all variants by unsourced percentages. Even the same game name can describe different rules, information and stopping conditions.
How to measure your own progress
- Choose one game and keep its rules constant.
- Decide whether you are counting first attempts or attempts with undo.
- Count abandoned deals consistently; dropping losses inflates the result.
- Record wins and attempts together: 12 wins in 40 attempts is 30%, but still a small sample.
Review a difficult move using the Klondike strategy examples or FreeCell planning guide. Improving decisions is useful even when a short winning streak ends.
Sources and review notes
Charles Blake and Ian P. Gent, The Winnability of Klondike Solitaire and Many Other Patience Games: authors' project and solver resources. Michael Keller, FreeCell FAQ, documents the Microsoft deal project and independent solver checks.
Reviewed September 8, 2026. This revision replaces unsourced difficulty rankings with scoped research results and separates solver winnability from player outcomes. See our editorial standards.
Frequently asked questions
What percentage of Klondike deals are winnable?
Blake and Gent report 81.945% plus or minus 0.084 percentage points for the thoughtful Klondike variant they studied, where all card identities are known. This is a solver estimate, not a human win rate or a measured result for this website.
Is every FreeCell deal winnable?
No. In the original 32,000 Microsoft deals, 31,999 are solvable and deal 11982 is not, under standard four-free-cell rules. That result applies to that numbered collection, not every possible shuffle.
Does losing mean my deal was impossible?
No. A losing attempt can result from an earlier choice or missing information. Undo can help you explore alternatives, but failing to find a solution does not prove there is none.
Do you publish win rates for players on this site?
This guide does not contain a measured player win-rate dataset. Its numerical examples come from the linked external research and the Microsoft FreeCell deal collection.