By Online Solitaire Free Editorial Team · Editorial Standards
Learn about rare patience games. Discover strategies, tips, and guides for playing free solitaire games online.
The mainstream solitaire catalogue — Klondike, Spider, FreeCell, Pyramid, Golf, TriPeaks, Yukon, Scorpion, Forty Thieves — covers the games most players encounter online. But the full historical catalogue of patience card games runs to several hundred distinct variants, many of which were played regularly in nineteenth and early twentieth century card-game collections before the digital era standardised the small set of variants that early computer implementations popularised. These rare variants are not inferior to their mainstream cousins — many offer strategic challenges, rule mechanics, and win probability structures that are genuinely distinct from anything in the mainstream set. They reward players who have developed solid foundations in standard solitaire strategy and want to encounter genuinely new decision problems that the familiar variants no longer provide.
This guide covers fifteen rare patience card game variants, each selected for at least one distinct strategic or mechanical property not found in the mainstream catalogue. For each variant, the description covers the basic rules, what makes it structurally different from mainstream variants, the approximate win rate range, and the primary strategic skill it develops or requires. Players who have mastered the probability, sequencing, and positional assessment frameworks described throughout this strategy cluster will find that these rare variants apply those frameworks in unexpected configurations — which is precisely what makes them valuable for continued strategic development.
Before exploring rare variants, it is worth establishing what players typically bring from standard solitaire experience. The core skills developed in Klondike — uncovering chain management, stock timing, foundation balance — apply across the entire patience family, because almost all foundation-building variants share the Ace-to-King suit structure and face-down card uncertainty. The probability framework from the key card probability guide applies to any variant with unknown card positions. Players who arrive at rare variants with these foundations will recognise familiar structural patterns in unfamiliar rule configurations — which is the fastest path to developing competent play in a new variant.
Spider experience is particularly valuable for the rare variants that feature multi-card sequence movement and suit consolidation goals, because Spider's same-suit discipline principle — never make a mixed-suit build when a same-suit build is available — generalises to any variant that penalises colour or suit mixing in tableau sequences. Players without Spider background will find suit-consolidation variants significantly harder to approach than players who have internalised the same-suit priority as a default habit.
1. Canfield (also called Demon Patience). A single-deck variant with a 13-card reserve pile, four tableau columns of one card each, and a three-card stock draw. The distinctive rule: the starting foundation rank is determined by the first card dealt — if that card is a 7, all four foundations begin on 7s and build up through King then wrap around to Ace. This wraparound foundation rule is unique in the mainstream patience family and requires tracking two build directions simultaneously. Win rate approximately 20–30%. Strategic challenge: managing the wraparound sequence while handling the three-card draw constraint.
2. La Belle Lucie (also called The Fan). Seventeen fans of three cards each, foundations built from Ace in suit, only the top card of each fan playable. The defining mechanic is the single redeal: once per game, all remaining tableau cards are gathered, shuffled, and redealt into new fans. Timing this redeal — not too early, not too late — is the primary skill challenge. Win rate approximately 30–40%.
3. Accordion. Cards dealt one at a time into a single row. A card can move left over one or three positions if the destination matches in rank or suit. The win condition is reducing all 52 cards to a single pile — no foundations. Win rate approximately 1–5%. Strategic value: develops pure rank-and-suit pattern recognition under the most severe positional constraints in the patience family.
4. Sir Tommy. All 52 cards dealt face-up one at a time; the player chooses which of four tableau piles each card goes to. Once placed, a card cannot be moved — only the top card of each pile is accessible for foundation play. Win rate approximately 35–50%. The irreversibility of every placement decision makes this a pure forward-planning problem: every incoming card must be assessed against all future cards whose foundation access depends on its pile assignment.
5. Gaps (also called Montana). All 52 cards dealt face-up into four rows of 13. Aces are removed creating four gaps. A card moves into a gap if it is one rank higher in the same suit as the card immediately left of the gap. Two redeals are allowed: completed sequences stay, remaining cards are reshuffled. Win rate approximately 15–25%. Strategic challenge: maximising correctly-ordered sequence length before each reshuffle.
6. Calculation. Four foundations built in arithmetic sequences regardless of suit: step-1 (A-2-3...K), step-2 (A-3-5...Q), step-3 (A-4-7...J), step-4 (A-5-9...10). Four unrestricted tableau piles serve as staging. Win rate approximately 25–40%. The challenge is purely sequencing: routing each card to the pile that preserves the correct arrival order for each arithmetic sequence.
7. Osmosis (also called Peek). Four foundation starters dealt face-up in a column; four rows of four face-down cards in the tableau. The cross-foundation rank dependency rule — a card can only be placed on a foundation if the same rank already appears on the foundation directly above it — creates a sequencing challenge unique in the patience family. Win rate approximately 40–55%.
8. Beleaguered Castle (also called Laying Siege). FreeCell without free cells and without a stock: all 52 cards dealt into eight columns of six, four Aces placed centrally as foundations, tableau builds in any-suit sequence. Win rate approximately 40–55% versus FreeCell's 80–90%. The absence of free cells eliminates the staging safety net that FreeCell's recovery strategy relies on, requiring deeper planning of every column interaction.
9. Sultan (also called Emperor of Germany). A two-deck variant with eight Kings and the Ace of Hearts arranged in a 3×3 at the start. Foundations build up in suit on the eight Kings; the centre King of Hearts stays fixed as the game's centrepiece. Win condition: a completed 3×3 of crowned foundations surrounding the central King. Win rate approximately 20–35%. The fixed centre card creates blocking patterns similar to Forty Thieves with an added spatial dimension.
10. Grandfather's Clock. Twelve cards arranged as a clock face serve as foundations, each building in suit to the rank of its clock position (the 2 of spades at the 2-o'clock position builds to 2; the Queen of hearts at 12 o'clock builds to Queen). Remaining 40 cards dealt into eight tableau piles of five, building in descending alternating colour. Win rate approximately 60–75% — among the highest of the rare catalogue. Strategic challenge: simultaneously managing twelve foundations with different target ranks demands multi-objective coordination unavailable in any mainstream variant.
11. Cruel Patience. All 52 cards dealt face-up into 13 piles of four. After moves are exhausted, cards are picked up pile by pile and redealt four at a time without shuffling — a deterministic redeal that rearranges the same card ordering into different pile groupings. Win rate approximately 10–20%; the low rate reflects how many initial deals have no winning path even with optimal play. Develops systematic pile management and recognition of which orderings survive redeal productively.
12. Nestor. All 52 cards dealt into eight columns of six plus one column of four. Win condition: clear the entire tableau by removing pairs of same-rank cards that are both face-up and accessible. No foundations — purely a matching puzzle. Win rate approximately 25–40%. The pair-accessibility planning required develops the same circular-dependency recognition that the three-pattern structural diagnostic targets, applied to a matching rather than foundation-building structure.
13. Windmill (also called Propeller). A two-deck variant: one King placed centrally builds regardless of suit from King down through Ace then up again, cycling continuously until all 104 cards are placed. Four diagonal foundations build from Ace up in suit. The resource competition between the suit-independent central pile and the four suit-specific diagonal foundations is unique to Windmill. Win rate approximately 35–50%.
14. Bristol. Three fans of three cards plus a stock that deals three cards face-up; only the rightmost of the three dealt cards is immediately playable. Win rate approximately 30–45%. The right-to-left access constraint on each stock deal creates a timing dependency similar to Turn 3 Klondike's grouping constraint, with the additional complexity of three simultaneous fan piles whose build sequences interact.
15. Stonewall (also called Flower Garden). Six tableau columns of six cards plus a 16-card reserve that is fully accessible in any order. The reserve's complete accessibility provides exceptional flexibility — any of the 16 reserve cards can be played at any time — making this the most forgiving rare variant. Win rate approximately 65–80%. Despite the high win rate, reserve management (which of the 16 cards to play immediately versus preserve for later) directly develops the resource preservation habits that transfer to free cell management in FreeCell and empty column discipline in Spider.
For players new to rare variants, Grandfather's Clock and Stonewall offer the highest win rates (60–80%) and the most familiar rule structures — both use standard descending-alternating tableau builds. For players seeking challenge comparable to Spider 4-Suit or Forty Thieves, Accordion, Cruel Patience, and Sultan offer win rates below 25% with distinct structural challenges the mainstream catalogue does not provide. For players interested in specific transferable skill development: Sir Tommy develops irreversible placement discipline (transfers to Turn 3 Klondike stock management); Calculation develops arithmetic sequence tracking (transfers to cross-foundation endgame coordination); Osmosis develops cross-foundation rank dependency awareness (transfers to endgame sequencing in all foundation-building variants). For the probability framework applicable to all these variants, see our key card probability guide.
The mainstream variants at onlinesolitairefree.com — Klondike, Spider, FreeCell, Golf, TriPeaks, Pyramid, Yukon, Scorpion, and Forty Thieves — cover the full difficulty, information, and strategy range that the rare catalogue explores in different configurations. The skill foundations built in the mainstream catalogue transfer directly to rare variant play: the probability framework, solver architecture insights from the AI solvers guide, and sequencing principles from the strategy cluster all apply across the full patience family. Mastering the mainstream catalogue with deliberate strategic practice is the most efficient preparation for approaching rare variants with competent play from the first session.
Three habits from mainstream solitaire transfer most efficiently to rare variants. First, identify the variant's key card structure before the first game — which cards unlock which positions, what is the primary blockage pattern? This maps the new variant's challenge onto the framework already developed in standard play. Second, apply the opening scan to the new variant's opening position — the five quality indicators (key card accessibility, sequence density, resource potential, key card clustering, empty column potential) apply in modified form to almost every patience variant. Third, apply the three-pattern structural diagnostic to stuck positions — circular dependency, key card burial, and resource exhaustion appear in rare variants as reliably as in mainstream ones and provide the same efficient resignation signal regardless of the specific rules.
Stonewall (Flower Garden) has the highest win rate at approximately 65–80%, driven by its 16-card fully-accessible reserve that provides exceptional flexibility. Grandfather's Clock is second at approximately 60–75%, and both are the recommended entry points for players encountering rare patience games for the first time. At the opposite extreme, Accordion has the lowest win rate at approximately 1–5% — its single-row constraint and non-standard win condition create an almost purely constraint-satisfaction problem with very few winning paths per random deal.
No. Like their mainstream counterparts, all rare patience games contain intrinsically unwinnable deals whose proportion varies by variant. Cruel Patience and Accordion have particularly high unwinnable rates; Stonewall and Grandfather's Clock have relatively low ones. The three-pattern structural diagnostic applies universally: a confirmed circular dependency, confirmed key card burial beyond accessible depth, or confirmed resource exhaustion all signal genuine unwinnability regardless of which variant is being played — and the correct response in every case is efficient resignation after diagnostic confirmation rather than continued play on a position that has no winning path.