PlaySolitaire.io Research The Win-Rate Observatory October 2026
Solitaire Win Rates: Winnable vs Actually Won
Nine in ten Klondike deals can be won. Players win fewer than four in ten.
- 37 won by players
- 53 winnable but lost
- 10 no winning line
What percentage of Solitaire games do people actually win? In our corrected cohort, players won 37.0% of 138,759 Klondike Draw 1 starts (95% confidence interval: 36.7–37.2%). About 90.5% (90.48% ±0.12) of Draw 1 deals may be winnable in principle. The gap is the difference between a possible solution and real play against hidden cards, ordinary mistakes and incomplete games.
- Actually won
- 37.0%
- Winnable in principle
- ≈ 90.5%
- Games analysed
- 138,759
01 Key findings
Six numbers worth knowing
- 37.0% of Klondike Draw 1 games end in a win 138,759 starts · 95% CI 36.7–37.2%
- ≈ 90.5% of Draw 1 deals have a winning line if every card is known Published estimate
- 35.5% of FreeCell games are won, though 31,999 of 32,000 deals can be solved 4,426 starts · 95% CI 34.1–37.0%
- 60.1% Spider — 1 suit: the most-won mode we measured 3,347 starts · 95% CI 58.4–61.8%
- 17.6% Spider — 2 suits (Random): the least-won mode we can publish 1,566 starts · 95% CI 15.8–19.5%
- ≈ 7% of Russian deals can be won at all, the harshest game we tested exact result · 2,000 deals
02 Results
Winnable vs actually won, game by game
6 game modes have at least 500 eligible starts, enough to publish a player win rate with a 95% confidence interval. The green bar is what players won; the gold mark is how many deals can be won at all, where a published share exists.
| Game | Won by players | Winnable in principle |
|---|---|---|
| Klondike — Draw 1 | 37.0% 95% CI 36.7–37.2% · 138,759 starts | ≈ 90.5% literature |
| Klondike — Draw 3 | 38.0% 95% CI 37.2–38.9% · 12,190 starts | No published figure |
| Spider — 1 suit | 60.1% 95% CI 58.4–61.8% · 3,347 starts | Practically every deal literature |
| Spider — 2 suits (Random) | 17.6% 95% CI 15.8–19.5% · 1,566 starts | Published verified pool: 90 of 90 exact |
| Spider — 4 suits | 20.6% 95% CI 18.0–23.5% · 843 starts | Most deals — if the line can be found literature |
| FreeCell | 35.5% 95% CI 34.1–37.0% · 4,426 starts | 31,999 of 32,000 (99.997%) catalog Which deals cannot be won? |
Player results on PlaySolitaire.io, starts July 15–August 3, 2026; outcomes observed through August 5. Unlimited undo included; intervals are two-sided 95% Wilson intervals. Still collecting: TriPeaks, Pyramid, Golf (strict rules), Yukon, Alaska, Scorpion, Wasp, Forty Thieves, Crescent, Canfield, Eight Off, and Russian have fewer than 500 eligible starts, so they appear only in the ladder below.
03 Solvability
The difficulty ladder: how many deals can be won at all?
The share of deals with at least one winning line, from the harshest game to the most forgiving. It describes the deals, not the players.
Brutal deals under 25% winnable
- Russian ≈ 7% exact
- Forty Thieves ≥ 7.7% lower bound
- Scorpion ≥ 12.8% lower bound
Hard 25–49%
- Golf (strict rules) ≈ 27% exact
- Canfield ≥ 33.1% lower bound
- Crescent ≥ 45% lower bound
Solid: exact solver result or complete deal catalog. Striped: a lower bound (the true share may be higher) or a literature estimate. Spider is not ranked because published sources describe it only in words. For an editorial ranking that weighs player results too, see the hardest solitaire games ranked.
04 Interpretation
Winnable versus won: why the gap is so large
"Winnable" means a winning line exists for a player who somehow knows every card — including the face-down ones — and never wastes a move. Real Solitaire is played against hidden information, ordinary mistakes and sessions that end before a win. That is why the two columns tell different stories. The ceiling describes whether a deal has a solution; the corrected measured rate describes what happened after a valid start on this site.
05 Methods
How the numbers were produced
How we measure real win rates
The published figures use methodology v2: wins divided by eligible mature starts from 15 July through 3 August 2026, with outcomes observed through 5 August. BigQuery joins each start to at most one terminal result by namespaced pseudonymous attempt identity and deduplicates event delivery. A mature start with no observed terminal remains in the denominator as not won. Practice boards, Daily games, Klondike winnable or golden pools and analytics-denied sessions are excluded. The frozen June methodology-v1 dataset remains available separately and is never blended into v2.
Read the complete reproducible methodologyHow the solvability ceilings are computed
The Winnable figures use four types of evidence. Exact: the solver classified every sampled deal under the tested rules. Lower bound: a solution was found for at least the reported share; unsuccessful bounded searches do not settle every remaining deal. Pyramid’s original curation found 1,000 wins among 1,520 sequentially scanned deals, about 65.8%, with four searches reaching the budget. That is a lower bound for the stopped sample, not an exact population estimate. Catalog: 31,999 of FreeCell’s original 32,000 numbered deals are winnable. Literature: for Klondike we quote the published estimate of 90.48% (±0.12) for thoughtful Draw 1 play. Our own Klondike solver checks the deals we serve, but we have not used it to measure how many random deals can be won, so that figure comes from published studies. The methodology records the rules, source scripts and limits.
Read the solvability method06 FAQ
Solitaire win-rate FAQ
What percentage of solitaire games are winnable?
For classic Klondike, research puts thoughtful Draw 1 play at about 90.5% winnable (90.48% ±0.12). Other variants span a huge range: nearly every FreeCell deal is winnable, TriPeaks reaches 96.6%, while Russian solitaire bottoms out around 7%. The results table and the solvability ladder above cover every variant we run, including the more surprising number: what players actually win.
What percentage of solitaire games do players actually win?
In the corrected cohort, players won 37.0% of 138,759 Klondike Draw 1 starts, 38.0% of 12,190 Draw 3 starts, 60.1% of 3,347 one-suit Spider starts, 17.6% of 1,566 current Random two-suit Spider starts, 20.6% of 843 four-suit Spider starts and 35.5% of 4,426 FreeCell starts. Each published row includes its 95% confidence interval.
What is the hardest solitaire game?
By solvability ceiling, Russian Solitaire: our solver settled 2,000 random deals exhaustively and only about 7% can be won by any line of play. Among the corrected player cohorts that currently clear the publication floor, current Random two-suit Spider is lowest at 17.6%. Those are different questions, so the table never treats measured completion as solvability.
Is every FreeCell game winnable?
Essentially — 31,999 of the original 32,000 numbered deals are provably winnable. Players in the corrected cohort won 35.5% of 4,426 eligible starts, which demonstrates that deal solvability and observed player completion are not the same measure.
Why are your numbers lower than other sites publish?
Many published Solitaire “win rates” are theoretical solvability figures or unsourced estimates. Ours are corrected attempt-level observations with exact samples, periods and 95% intervals. Real play and theoretical solvability answer different questions; the gap is a finding, not an error.
How often are these numbers updated?
Deliberately, not silently. Every release has an immutable dataset version, period and correction history. New games remain inconclusive until they clear 500 eligible starts, and every published JSON, CSV, chart and table must agree before release.
07 Open data
Use this study
The data, charts and tables are free to republish under CC BY 4.0: credit PlaySolitaire.io with a link to this page (licence details).
Cite as PlaySolitaire.io Win-Rate Observatory, https://playsolitaire.io/solitaire-win-rates (CC BY 4.0), win-rates-v2-2026-10-21, corrected methodology v2.
