Master advanced Pyramid Solitaire strategy. Card counting, probability-based decisions and score maximisation explained for experienced players.
Pyramid Solitaire is the patience game where analytical thinking produces the most disproportionate returns. The average casual player wins 10–20% of games; a player who applies card counting, probability-weighted decision making, and score maximisation principles consistently wins 30–45% — two to three times more often on identical deals. The reason the gap is so large is that Pyramid contains more information than most players use: the pyramid's 28 cards are all visible from the start, every s
Pyramid Solitaire deals from a single standard 52-card deck divided across three zones at any point in the game: the pyramid (initially 28 face-up cards), the stock (initially 24 face-down cards), and the waste pile (initially empty, growing as stock cards are drawn). Card counting in Pyramid means maintaining an accurate model of which ranks remain in each zone at each point in the game — and using that model to make decisions that would be arbitrary without it.
Track Kings independently from all other ranks. Kings are the only self-pairing rank in Pyramid — they remove themselves without needing a partner. A King accessible in the pyramid is always the highest-priority removal because it frees a position at zero pairing cost. More importantly, a King buried in the pyramid is a permanent blocker until it becomes accessible: it can never be paired away from below, and anything above it must be cleared before it can be removed. At the start of every game, locate all four Kings immediately. Kings in rows 1 through 4 (counting from the apex) are serious threats — identify their row depth and mark which pyramid cards must be cleared to reach them.
Track the 13-rank completion count for each suit. Each suit contributes four pairing relationships to Pyramid (A+Q=13, 2+J=13, 3+10=13, 4+9=13, 5+8=13, 6+7=13, K=13 alone). With four suits and six pair-types plus one King per suit, the full pyramid clear requires 24 pairs and 4 Kings. Tracking which specific rank pairs have already been completed tells you exactly which pairs remain to be found — and therefore which pyramid cards are currently unpairable because their partner has already been removed from the deck entirely.
The unpairable card concept is the most important product of card counting. An unpairable card is a pyramid card whose partner rank has been completely exhausted — all four cards of the complementary rank (adding to 13) have already been played. A 6 is unpairable when all four 7s have been removed; a 9 is unpairable when all four 4s are gone. An unpairable pyramid card cannot be cleared by any subsequent play; the only path to clearing it is if another pyramid card blocking it from below can be removed, eventually making it irrelevant to the clear path. Identifying unpairable cards early is the single most valuable product of card counting in Pyramid: an unpairable card in a critical pyramid position (rows 1–4) is a strong signal that the complete clear is no longer achievable, allowing you to switch to score maximisation mode without wasting further passes.
Maintain a running count of each complementary pair's availability. At any point in the game, for each of the six pair types (A+Q, 2+J, 3+10, 4+9, 5+8, 6+7), track how many of each rank remain in the combined pyramid+stock+waste pool. The critical thresholds: if both ranks of a pair type have zero remaining in pyramid positions and at least one card of each rank remains in stock or waste, the pair can still be completed from the stock — but only if those cards appear in positions where the waste top and stock draw align correctly. If one rank of a pair type has zero remaining anywhere, all pyramid cards of the complementary rank are unpairable.
Use the waste pile as a known card register, not just a discard location. Every card on the waste pile is a known, enumerated card — it has been drawn from the stock, its rank is visible (waste pile top) or was visible when drawn and should be mentally logged. Expert Pyramid players maintain a rough mental model of the waste pile's composition, particularly for ranks that appeared early in the stock draw phase. This model enables three specific decisions: knowing whether the waste top's complement exists elsewhere in the stock; predicting how many draws separate a buried waste card from the top; and calculating whether cycling the waste pile (in implementations that allow multiple passes) will produce better or worse pairing opportunities than the current pass.
Card counting provides the raw data; probability assessment converts that data into better decisions at the specific moments when multiple valid moves exist and one is objectively better than the others. In Pyramid, these decision points arise primarily in three contexts: choosing between multiple simultaneously accessible pyramid pairs; deciding whether to pair a pyramid card now or draw from the stock in hope of a better sequence; and deciding when to use the waste top versus drawing a fresh stock card to extend a pairing run.
Calculate draw probability for each needed rank before deciding to draw. The probability that the next stock draw produces a specific rank is (remaining cards of that rank in the stock) ÷ (total remaining stock cards). If 3 cards of rank 7 remain in the stock and 15 stock cards remain total, the probability of drawing a 7 next is 3/15 = 20%. This simple calculation resolves the most common Pyramid decision ambiguity: whether to pair a pyramid card using the current waste top or draw a fresh stock card hoping for a specific rank. If the probability of drawing the needed rank is above 25% and the waste top alternative would consume a card that has better future use, draw. If below 15% and the waste top alternative is acceptable, take the waste top pair rather than gambling on the draw.
Apply expected value thinking to multi-step pairing decisions. A multi-step pairing decision is one where two valid pairings exist but only one can be made, and the choice determines which subsequent pairings become available. The correct choice is not simply the one that removes the most cards immediately — it is the one that produces the best expected position across the next three to five pairings. To evaluate: for each available choice, identify the subsequent pairing that becomes available if that choice is made, and estimate the probability that the enabling card for the subsequent pairing appears before the stock is exhausted. The choice with the higher probability of enabling a productive follow-up sequence is the correct choice even if the immediate removal count is equal.
Prioritise pairs that uncover pyramid cards whose complement has high stock availability. When two pyramid pairs are simultaneously valid, prefer the one whose removal uncovers a new pyramid card with a high-probability complement remaining in the stock. Example: if removing pair A uncovers a pyramid 6 and three 7s remain in the stock, while removing pair B uncovers a pyramid Jack and only one 2 remains in the stock, removing pair A is superior — the uncovered 6 has a 75% higher probability of eventually finding a partner than the uncovered Jack. This probability-weighted uncovering preference is the most consistently applicable decision rule in intermediate-to-advanced Pyramid play.
Adjust probability calculations after each draw. Each stock draw changes the composition of the remaining stock, which changes the probability of every subsequent needed draw. An expert Pyramid player does not calculate a single probability at the start of a pass and apply it throughout — they recalculate after each draw, particularly when a high-value rank is drawn (raising or lowering the probability for its complement). The recalculation is simple: after each draw, subtract the drawn rank from the stock count and recalculate the relevant pair probabilities. This takes five to ten seconds per draw and produces materially better decisions than fixed-probability assessments.
Pyramid Solitaire is unusual among patience games in that it is scored — most implementations award points for each pyramid card cleared, bonus points for completing the pyramid, and deduct points or count remaining cards for incomplete clears. Because 60–75% of Pyramid deals cannot be fully cleared even with optimal play, the skill of maximising score in losing hands is actually more frequently applied than the skill of completing the pyramid. A player who consistently extracts 22–25 pyramid clears from an uncompletable hand scores substantially better over a session than one who abandons the hand at the first sign of a block.
Switch to score maximisation mode the moment a complete clear is confirmed impossible. The trigger for score maximisation mode is the identification of an unpairable pyramid card in a structurally critical position — a row 1 through 4 card whose complement rank has been exhausted. Once this is confirmed, the strategic objective changes from complete clear to maximum pyramid card removal. Score maximisation mode requires different move prioritisation: instead of optimising for pyramid-clear paths, optimise for removing the maximum number of accessible pyramid cards per remaining stock pass, even if those removals do not contribute to any clear sequence.
In score maximisation mode, prioritise accessible pyramid pairs over pyramid-stock and pyramid-waste pairs. A pyramid-pyramid pair removes two pyramid cards per move — the highest removal rate. A pyramid-stock or pyramid-waste pair removes one pyramid card and one non-pyramid card per move — a lower rate that consumes a stock draw in exchange for only one pyramid removal. In score maximisation mode, pyramid-pyramid pairs are worth approximately twice as much per stock draw as pyramid-stock pairs; exhaust all available pyramid-pyramid pairings before drawing from the stock, even when a pyramid-stock pair is available.
Plan stock draws in score maximisation mode to maximise pyramid-stock pairing density. When no pyramid-pyramid pairs remain and a stock draw is required, the best stock draw is one that produces a rank maximally paired with the current accessible pyramid cards. Before drawing, count how many accessible pyramid cards each remaining stock rank would pair with. The rank with the highest pairing count is the most valuable draw — and while you cannot control which rank appears, you can time draws to occur when the pyramid's accessible card composition is optimised to receive the most likely useful ranks.
Never exhaust the stock on an already-confirmed uncompletable position without extracting maximum pairs first. Once score maximisation mode is active, every remaining stock draw should be evaluated against the question: does this draw remove a pyramid card, or does it just advance the waste pile position? Draws that do not directly remove pyramid cards are only justified if they are positioning the waste pile for a subsequent pyramid-stock pair that would otherwise not occur. Avoid drawing past useful cards on the waste pile — if the waste top is paired with an accessible pyramid card, take the pair before drawing further.
Manage the waste pile cycling order in multi-pass implementations. In implementations that allow two or three passes through the stock, waste pile management becomes a combinatorial optimisation problem. The key insight: the order in which stock cards are drawn in pass one determines the order they appear in the waste pile for pass two. Cards drawn early in pass one appear deep in the waste pile in pass two — accessible only after every card drawn after them in pass one has been played or passed over again. Before completing pass one in a multi-pass game, identify which cards remain in the stock, predict which ranks you will need for pass two pairing sequences, and structure the end of pass one to position those cards near the waste pile top for pass two access.
The row-depth clearing priority framework. When no other decision criterion applies, prioritise clearing pyramid cards in the deepest rows (rows 1–3 from the apex) before the shallower rows (rows 5–7 at the base). The reason: each deep-row card covers two shallower cards, so removing a deep-row card expands the accessible pool by two cards; removing a base-row card expands the accessible pool by zero. The row-depth priority is the tie-breaking principle when probability calculations and unpairable-card detection do not distinguish between two available pairs.
Count the pass number against the accessible pyramid card count. At the transition between stock passes, compare the number of accessible pyramid cards to the pass number. On pass one with fifteen or more accessible pyramid cards, play aggressively — use stock draws freely to generate pyramid-stock pairings. On pass two with fewer than eight accessible pyramid cards, play conservatively — every non-productive draw may bury a needed waste card under additional passes of the same irresolvable cards. The pass-count-to-accessible-card ratio is a simple measure of late-game viability: low accessible cards relative to remaining passes means the game is likely in its final productive phase.
Use the waste pile top deliberately, not reflexively. The waste pile top is a permanently visible known card that can be used at any time. Players who immediately use the waste top whenever it pairs with an accessible pyramid card are playing reactively; players who occasionally hold the waste top through a draw in order to reach a better subsequent pairing sequence are playing analytically. Before every waste top use, ask: if I use this pair now, what is the probability that the next stock draw produces a better sequence? If the answer is low — if the waste top pair is a good one and the next draw is unlikely to improve on it — take the pair. If the answer is moderate or high, hold the waste top through one more draw and re-evaluate.
Card counting in Pyramid produces three specific improvements over pattern-recognition play. First, it identifies unpairable pyramid cards early — cards whose complement rank has been exhausted — enabling the switch to score maximisation mode before the stock is wasted on an uncompletable clear path. Second, it provides the raw data needed for probability-weighted decision making at the numerous points in each game where two valid moves exist and one is objectively superior. Third, it enables waste pile cycling management in multi-pass implementations by tracking which cards are positioned where in the waste pile and which stock draws will produce which waste compositions in subsequent passes. Together, these three improvements are responsible for most of the gap between the 15–20% casual win rate and the 30–45% analytical win rate on identical deals. Play our free Pyramid Solitaire game and begin with just Kings tracking and unpairable card detection — these two habits alone produce the fastest improvement from any single starting point.
Score maximisation in an uncompletable Pyramid hand follows a four-rule priority system. Rule one: switch to score maximisation mode immediately upon identifying the unpairable card — do not continue playing as if a complete clear is still achievable. Rule two: prioritise pyramid-pyramid pairs over all other pair types, because they remove two pyramid cards per move at zero stock cost. Rule three: when drawing from the stock is necessary, draw when the accessible pyramid composition maximises the probability of useful pyramid-stock pairings from the likely upcoming ranks. Rule four: in multi-pass implementations, structure pass one's final draws to position high-value pairing partners near the waste pile top for pass two access. Applied consistently, these four rules reliably extract three to five additional pyramid card removals per losing hand compared to unstructured play — which translates directly to higher session scores. See our solitaire win rates guide for context on Pyramid's unwinnable rate.