Not every solitaire deal is guaranteed to be winnable. However, losing a deal does not prove it was impossible from the start. An earlier move may have trapped a required card, changed stock access, or consumed the workspace needed for the solution.
Three states that are often confused
A solvable deal
At least one legal sequence of moves reaches the win condition from the initial layout. A player can still lose a solvable deal by choosing a different branch.
A dead current position
The initial deal may have had a solution, but previous moves have made victory unreachable from the current state. Undoing or restarting and changing an earlier decision can recover the solution.
An unsolvable deal
No legal sequence solves the initial layout under the exact rules. Proving this generally requires exhaustive analysis or a solver that accurately models those rules. A visual guess is not always enough.
Why the variant and settings matter
Klondike includes hidden tableau cards and stock access. Draw one, draw three, pass limits, and whether foundation cards can return all affect reachable moves. The same card order can behave differently under different settings.
FreeCell exposes most information, which makes planning more direct, but that does not justify claiming that every possible deal or every numbered-deal collection is solvable. The deal set and movement rules must be specified.
Spider changes sharply with suit count. Evidence about one-suit play cannot be treated as a four-suit solvability result.
Some digital games select winnable deals
A digital implementation may test generated deals with a solver or draw from a curated set. A “Winning Deal” mode can therefore promise a solution, but it does not promise that the solution is easy to find. Daily Challenge, Random Deal, and Winning Deal modes may use different generation rules within the same product.
Solitaire JP does not describe every deal as guaranteed winnable unless the player or provider documentation makes that condition clear.
Why published win rates are difficult to compare
- Draw-one and draw-three Klondike are different rulesets.
- Pass limits, undo, hints, and automatic moves vary.
- Some samples include all random deals; others use selected solvable deals.
- A statistic may describe casual players, experts, or a solver.
- Abandoned games may or may not count as losses.
A percentage without its source, rules, sample, and definition of a completed game is weak evidence. Site likes and play starts are also not measures of mathematical solvability.
How to review a lost deal
- Find the earliest position where two plausible moves were available.
- Compare an information-revealing move with a foundation-only move.
- Check whether an empty column or free cell was occupied without a return route.
- Ask whether a destination existed when the needed stock card appeared.
- If the same deal can restart, change one branch and compare the resulting access.
Use losses as strategy evidence
You do not need to prove every failed deal impossible. It is more useful to ask whether you maximized future choices, recovered temporary spaces, and organized the tableau before drawing. The stuck-position checklist provides a repeatable review order.
For specific rule effects, compare Klondike draw modes and Spider suit settings.