A mystery solved by reading the floor plan
Clue Manor begins with a very visual idea: several guests are hidden somewhere inside a manor and each of them knows something about their position. One may know that they are in the Library. Another might be next to a sofa. A third may only know a row or a column. None of those statements has to solve a position on its own. The answer emerges when they are combined with the rule that governs the entire board: no two guests may share a row or a column.
That turns the floor plan into more than scenery. Every room, outside wall and piece of furniture contributes spatial information. Placing a guest does not merely occupy one cell; it removes an entire row and an entire column from everyone else's possibilities. A clue that looked broad a moment ago can suddenly become exact. Once all visible guests are placed, the only unused row and column cross at a cell that needs no extra statement at all: that is where the Hider was waiting.
The Blupoli version is built around that gradual deduction loop. It lets you keep pencil marks, place guests, drag a guest onto the plan, undo an experiment and inspect individual clue regions without turning the interface into a second puzzle. The challenge should come from understanding the constraints, not from fighting the controls.
The manor is really a constraint system
At first glance the board looks like a decorated grid. A more useful way to read it is as several constraint layers occupying the same space. The grid provides rows and columns. Walls divide the plan into named rooms. Furniture creates special cells: sofas and armchairs remain playable and can appear in clues, while tables, plants and statues physically block the cells they occupy.
Each clue cuts down that space. If Ada is in the Drawing Room, every cell outside that room disappears from her candidate set. If she is also known to be on row 3, only cells where the room intersects that row survive. You do not have to choose between those cells immediately. You can record several candidates and let another placement remove a column or a row later.
This matters because deduction puzzles become much easier to reason about when uncertainty is allowed to remain visible. Knowing that a guest has three possible cells is useful information. You do not need to convert every partial fact into a final answer the moment you discover it.
One guest per row and column changes the value of every clue
The global exclusion rule is easy to remember, which can make it easy to underestimate. Suppose Bruno has three possible cells inside one room. Two lie in columns already occupied by other guests. Without learning anything new about Bruno, the third cell becomes forced. Placing him then removes his row and column for everybody else, so the consequence propagates.
On a 4×4 board there are three visible guests plus the Hider, and that propagation is easy to follow almost at a glance. A 7×7 game has six visible guests and enough space for a broad clue to stay unresolved through several moves. The principle does not change. Every confirmed placement makes the effective board smaller for the remaining people.
A useful habit follows from this: after placing someone, do not immediately hunt for a new clue. First scan what the occupied row and column have done to every other candidate set. Often that propagation contains more useful information than rereading the whole puzzle from the beginning.
Not every clue carries the same amount of information
An exact row-and-column clue can identify one cell. A row or column clue usually leaves several possibilities. A room may be broader still. An outside-wall clue, a wing of the manor or a relationship with a sofa describes an area whose size depends on the particular floor plan. Those differences are what make difficulty meaningful.
Easy games contain more clues that reduce the board quickly. As you move through Medium, Hard and Expert, clues tend to leave larger candidate sets and therefore require more interaction between guests. The important measure is not how long a clue sounds. It is how many actual positions on this board still satisfy it.
The interface makes that visible. Selecting a guest highlights every cell compatible with all of that guest's clues. You can also tap an individual clue fragment—such as “Library” or “north wing”—to highlight only that region. Highlighting is not a solver button. It simply turns language into geometry so you can reason directly on the plan.
Read the building before making the first move
It is worth spending a few seconds reading the manor itself. Which rooms stretch across many rows? Which are concentrated in only two columns? Where are the outside walls relevant to clues? Where are sofas and armchairs? Those details determine how informative each statement really is.
“In the Library” may be weak when the Library has eight open cells, but it can become strong after occupied columns remove half of them. Likewise, “by an outside wall” sounds broad on an empty board yet may become decisive once only one compatible row remains. A clue's practical strength changes as the game state changes.
There is therefore no permanent ranking in which a room clue is always weaker than a row clue. The best clue is the one that currently leaves the fewest viable configurations. Learning to notice when a previously broad clue has become sharp is a large part of the game.
Furniture is part of the language, not decoration
The furniture exists for logical reasons. Sofas and armchairs are seats, and their cells remain open. A guest may sit on one or occupy a neighbouring cell when a clue says so. Tables, plants and statues are different: they block their cells and reduce the usable floor area.
That distinction creates deductions even when no clue mentions a blocked object. A blocked cell can give a room an irregular shape or leave fewer open cells in a row. The floor plan therefore contains information before you place the first guest.
In the Blupoli implementation this is not a visual trick added after generation. The same open and blocked cells are used by the generator, solver and interface. A puzzle is only accepted if the logical model being verified is the same one the player sees.
Pencil, place and clear with one repeated action
A puzzle built on partial information needs a fast way to preserve uncertainty. With a guest selected, the first tap on a cell adds that guest as a pencil mark. A second tap on the same cell turns the mark into a placement. A third tap removes the guest and returns the cell to its empty state. Pencil → place → empty keeps the interaction predictable without requiring a separate mode switch.
Pencil marks belong to individual guests. If both Ada and Clara could occupy one cell, their initials can coexist there while the deduction remains open. When Ada is finally placed, her other notes are no longer needed and incompatible notes in the occupied row and column are cleared automatically. Undo restores the complete earlier state, including those notes.
On desktop you can also drag a guest card directly onto an open cell. Dragging is an accelerator, never the only route. The same game remains fully playable with taps or keyboard controls because a core mechanic should not depend on pointer precision.
The Hider does not need a clue card
The most satisfying part of Clue Manor happens at the end. The Hider is not another ordinary guest with another statement. There is no special clue telling you which room contains them. Their position is a consequence of the finished board.
On an N×N grid there are N−1 visible guests. If all those guests occupy distinct rows and columns, exactly one row and one column remain unused. Their intersection creates one cell. A valid puzzle guarantees that this intersection is open, and that cell is the Hider's location.
This gives the ending a different texture from a normal “last clue” reveal. The final answer appears because all previous deductions have exhausted the remaining alternatives. It also provides a useful mental check: if your supposedly finished placements leave an intersection on a blocked object, something earlier cannot be consistent.
Four sizes create four different information-management problems
Clue Manor supports 4×4, 5×5, 6×6 and 7×7 boards. Increasing size is not meant simply to shrink the cells and call the result harder. Every extra row adds another visible guest, another column and more positions that can survive a broad clue.
The 4×4 board is ideal for learning the grammar. With three visible guests, the effect of each placement is easy to trace. At 5×5, pencil marks start to earn their keep because several plausible positions can remain open at once. On 6×6 and 7×7 boards, managing candidate information becomes as important as understanding the initial statements; one correct placement may affect several guests that were waiting for another restriction.
Size and difficulty are independent controls. You can choose a 7×7 Easy game when you want a large spatial puzzle with generous information, or a 4×4 Expert game when you prefer a compact board whose clues remain open longer. “More cells” and “more demanding deduction” are not treated as the same thing.
What Easy, Medium, Hard and Expert actually change
A useful difficulty setting must alter the puzzle rather than merely change a label. Clue Manor's generator begins with highly precise information and then weakens clues while an independent solver keeps checking that exactly one solution remains. Easy aims for small candidate sets. Medium preserves more alternatives. Hard and Expert try to keep broader regions unresolved long enough that row-and-column propagation becomes essential.
The progression is measured on the generated floor plan itself. The phrase “in the Drawing Room” might correspond to three cells in one manor and six in another. The engine therefore evaluates the real candidate cells behind every clue and adjusts the final clue set toward the information profile of the selected difficulty.
One rule never changes: uniqueness. If weakening a clue creates a second valid completed manor, that change is rejected. A harder puzzle may require a longer chain of deductions, but the final answer never becomes an arbitrary choice between two legal configurations.
Generating a manor takes more than shuffling coordinates
Every game is derived from a seed. The engine first creates a hidden solution with one selected position in each row and column and reserves one of those positions for the Hider. It then divides the floor plan into connected rooms and places furniture without blocking any cell required by the solution.
From that geometry the generator can create truthful clue components for each guest: row, column, room, wing, outside wall, relationship with seating, or combinations of those ideas. A clue is not accepted merely because its wording sounds plausible. The engine can calculate the exact set of open cells that satisfy it.
Those candidate sets are what the difficulty system manipulates. Finally the solver tests the whole puzzle under the one-per-row-and-column rule. The seed keeps the process deterministic, so the same seed, size and difficulty reproduce the same board—important for automated tests and for any shared or daily puzzle context.
The solver is an auditor, not an invisible player
A generator can create a visually convincing floor plan and still produce a poor puzzle. Two clues can be redundant. A large room can leave two equivalent guest arrangements. A blocked cell can make the final unused intersection invalid. Those are structural mistakes, and they need a second system to detect them.
The Clue Manor solver takes the finished clues and builds a candidate domain for every guest. It searches assignments without repeating rows or columns. When all visible guests are assigned, it calculates the unused row and column and requires their intersection to be an open cell. The search stops as soon as it finds a second solution, because one extra solution is enough to reject the generation.
The solver's route is never shown during normal play. It does not feed a sequence of moves to the interface. Its job is simply to support one promise: every published board has one final arrangement determined by the rules you can see.
Why the generator weakens clues instead of adding random ones
A common way to build logic puzzles is to add random clues until only one solution remains. That can work, but it can also produce a supposedly hard puzzle full of irrelevant statements or an easy puzzle that depends on one unexpectedly tangled relation.
Clue Manor works in the opposite direction. It starts with maximum information and looks for broader replacements that preserve uniqueness. If an exact position can be relaxed to “row 3” without introducing another solution, the puzzle gains meaningful uncertainty. If it can become “Library” and the other guests still force a unique completion, the deduction has become more global.
The goal is not to assign a scientific intelligence score to a board. It is to make the four tiers recognisable: more local support in Easy, more cross-guest interaction in Expert. The individual manor then provides its own character through rooms, furniture and the particular way candidate zones overlap.
A strong opening: begin with the smallest domains
At the start of a game, select each guest once and look at roughly how many cells light up. You do not need an exact count. The useful distinction is between somebody with two or three plausible cells and somebody who could still be almost anywhere in a large wing. Start with the smaller domains.
If a guest has an exact or near-exact combination, an early placement may be safe. If two remaining cells are both on the same row, you may not know which column is correct yet, but you have learned that this guest will consume that row. That information can already constrain another guest.
Pencil marks are most valuable when a small candidate set spans several rows and columns. Recording it prevents you from reconstructing the same deduction later. The opening objective is not to fill as many cells as possible; it is to create a reliable map of possibilities so the first placement produces a useful chain reaction.
After a placement, propagate before reading something new
Suppose Clara is placed on row 4, column B. From that moment nobody else can use row 4 or column B. If Diego had three candidates and two lie on those lines, Diego is now solved without any change to his own clue. If Emma had a pencil mark in column B, it disappears. This is the core rhythm of Clue Manor.
A productive solving loop therefore alternates between two modes. First, interpret a clue strongly enough to obtain a certainty. Second, propagate that certainty through the entire grid. Only when propagation stops do you go hunting for another promising statement.
This makes large boards much less intimidating. A 7×7 manor may contain many possibilities at the beginning, but you do not need to settle them all at once. Find one sufficiently constrained guest, place them with a reason, and let the row-and-column system do additional work.
Broad clues often become useful later
“West wing” can feel almost useless on the first move. It may cover several rooms and a large fraction of the floor plan. Do not mentally discard it. Its value rises as other guests occupy western columns or remove particular rows.
This is why rereading is part of the intended solution process. You are not repeating the same analysis; the context has changed. A clue that allowed six cells earlier may now allow only two even though its wording is identical. On Expert boards, many decisive deductions come from returning to old information after two or three placements.
A good routine is to revisit the cards with the largest highlighted areas whenever an important row or column becomes occupied. Something that was “too broad to use” may have quietly become the next forced move.
Use furniture clues literally
“Next to a sofa” describes a specific geometric relationship with a sofa cell. It does not automatically mean “in the same room,” “sitting on it” or “sharing its row.” Likewise, “sitting on an armchair” points to seat cells rather than adjacent cells. The clue highlighting helps separate those meanings.
A manor theme naturally invites interpretation, but the puzzle is strongest when every statement is treated as a formal condition: a cell either satisfies it or does not. The theme makes the rule easy to picture; it should not introduce fuzzy assumptions.
When a furniture clue leaves several positions, combine it with global restrictions. A sofa might have four neighbouring cells, but perhaps only one lies on a row that is still available to that guest. Local spatial clues and the global row-column rule are designed to reinforce each other.
Useful mistakes are why Undo matters
Deduction puzzles invite experiments. A chain may look convincing and only reveal a contradiction two moves later. Undo should help you find the first unsupported decision rather than force you to rebuild the entire board.
In Clue Manor an undo state includes both placements and pencil marks. That detail matters because placing a guest automatically clears incompatible notes in the same row and column. If Undo restored only the person and not the notes, the board would no longer represent the earlier reasoning state.
Redo makes it possible to compare those states without manually recreating them. The tools are not intended to turn guessing into a strategy. They are there so a revised deduction or an accidental tap does not destroy the work around it.
A hint should unlock the puzzle, not replace it
The Hint action reveals the correct position of a guest who is not yet correctly placed. It deliberately does not fill an entire room, clear every wrong note or reveal the Hider in advance. The interesting part still comes after the hint: the new occupied row and column may create several deductions elsewhere.
This makes hints useful when a session stalls without changing the nature of the game. One new certainty can restart the propagation loop, and you can continue solving from there rather than watching the interface complete the board.
The final result also keeps track of how many hints were used. That is context, not a punishment. A solve with assistance and a solve without assistance are different kinds of sessions, and both can contribute to learning the puzzle.
The same board has to work for touch, mouse and keyboard
The game sizes itself from the rectangle supplied by the Blupoli Puzzles shell. On compact portrait screens the board prioritises available width, and the game screen can scroll inside the application shell if controls and board together require more height. Even a 7×7 plan keeps usable touch targets instead of being clipped to satisfy an artificial fixed-height layout.
Guest cards become a horizontally scrollable strip on compact screens so they do not squeeze the grid. On desktop and landscape they occupy a side panel where several clues can remain visible at once. These are not separate mobile and desktop products; they are two arrangements of the same controls.
Keyboard navigation is part of the same contract. Arrow keys move between playable cells, while Enter or Space applies the same pencil-placement cycle as a tap. Accessible labels describe rows, columns, rooms and furniture so the state is not encoded by colour alone.
Colour supports the reasoning instead of carrying it
Clue Manor follows the Blupoli Puzzles visual system: solid surfaces, clear room boundaries and the category accent for selection and focus. It does not assign a separate bright colour to every guest. Initials and names identify people even when colour is unavailable or difficult to distinguish.
Rooms are separated primarily by borders rather than a patchwork of backgrounds. That choice keeps a 7×7 board readable when furniture, pencil marks and placed guests are all present. A clue highlight is temporary and always tied to a selected card or textual clue fragment.
Visual restraint matters in a constraint puzzle because several layers of information may be visible at once. Less decorative noise makes it easier to notice what changed after a move.
A miniature example: from four candidates to one
Imagine that Ada's clue says “Library,” and the Library contains four open cells. Bruno is constrained to column C. Clara must be next to a sofa. You place Bruno first because his column, combined with the current floor plan, leaves only one row. That placement eliminates Bruno's row and column C.
Two of Ada's four Library cells disappear—one on Bruno's row and one in column C—so Ada now has two candidates. Clara originally had two cells next to sofas; one lies on the occupied row, so Clara becomes forced. Placing Clara removes another column, which eliminates one of Ada's final two candidates.
Ada's clue never changed. It began with four valid cells and ended with one because information elsewhere altered the context. After Ada is placed, the remaining guest may still have a broad clue, but nearly every row and column is already consumed. Once the final visible guest is settled, the unused intersection reveals the Hider. That is the chain reaction Clue Manor is designed to create again and again with different layouts.
Clue Manor and Logic Matrix: two ways to preserve uncertainty
If you enjoy keeping possibilities open rather than committing too early, Logic Matrix is a natural companion. Its cells preserve symbolic candidates and its clues relate categories. Clue Manor uses physical positions as candidates and keeps the one-per-row-and-column rule active across the whole floor plan.
That changes how progress feels. Logic Matrix asks you to inspect abstract candidate domains inside a matrix. Clue Manor lets you reason in areas: rooms, walls, wings and furniture. Both reward the same discipline of leaving a hypothesis unresolved until another constraint closes it.
The Logic Matrix guide explores that other style in more detail. Moving between the two games is a useful way to practise the same underlying habit—record uncertainty accurately—through very different visual representations.
How it differs from Logic Grid and Einstein Riddle
Logic Grid focuses on explicit relationships between categories, recorded as confirmed or impossible pairings. Einstein Riddle adds written order, adjacency and equality statements. Clue Manor shares the process of translating language into constraints, but makes physical space the primary model.
Instead of reading a relationship table, you can see “row 3” immediately. The Library has literal walls. “Next to the sofa” maps to a small set of cells. That direct translation between wording and geometry makes the game approachable without removing the need to combine clues.
Trying all three reveals how the same deductive skill can be represented as a matrix, a set of written relations or a floor plan. Difficulty does not need ever more complicated rulebooks; changing the representation of information is enough to demand a different style of attention.
What to do when the board seems stuck
When no move is obvious, avoid testing a random cell. Run a structured scan. Select every guest and compare the approximate size of their highlighted region. Remove rows and columns already occupied. Revisit room clues that cross those lines. Then inspect furniture, edge and wing clues again.
It is also useful to read pencil marks by row or column instead of by guest. If only one guest can still occupy a particular row, that may be as strong as an explicit position clue. The global rule can create deductions that belong to no single clue card.
If the grid still refuses to open, a hint can supply one placement and let you continue the reasoning from there. The goal is to recover a chain of consequences, not to accumulate unexplained moves.
Why a unique solution changes how a dead end feels
Knowing there is exactly one solution changes your relationship with uncertainty. If two finished configurations appear equally valid, you do not have to pick the one that “looks right.” Some restriction has not yet been used fully. That trust matters most on high difficulties, where broad clues may remain unresolved for much of the session.
Unique solution does not mean there is only one order of moves. Several deductions may be available simultaneously, and two players can begin with different guests. The guarantee concerns the final arrangement, not a mandatory path.
The solver protects the destination while leaving the journey to the player. That separation is what allows procedural generation to be strict without making play feel scripted.
Practise size and difficulty separately
If you are learning the game, keep the board size fixed for a few sessions and change only difficulty. On 5×5, for example, it becomes easy to notice how direct information decreases and cross-guest reasoning becomes more important from Easy to Expert. Then reverse the experiment: keep a comfortable difficulty and move from 4×4 through 7×7 to practise managing more candidate domains.
Separating those axes helps identify the actual problem. You may understand broad clues well but lose track of notes on large boards; then candidate management is the skill to work on. Or you may comfortably manage a 7×7 Easy board yet stall on 4×4 Expert; then the challenge is chaining restrictions, not visual volume.
The point is not to rush toward Expert. It is to become better at explaining why a cell has stopped being possible and recognising when an old clue has become newly useful.
The floor plan as an external memory
A good deduction game should not demand that every candidate live in your head. Clue Manor's floor plan acts as a notebook. Pencil marks preserve possibilities. Placements record certainties. Walls keep room membership visible. Furniture keeps spatial restrictions on the board. Undo even preserves the history of those annotations.
That frees working memory for the interesting part: combining rules. Instead of remembering that Ada might be at A2 or D4, you can see those notes. Instead of reconstructing which row was just consumed, the placed guest remains on it. Progress is externalised.
This is a broader principle across Blupoli Puzzles: the interface should represent the logical state honestly. It should not hide known information to manufacture difficulty, and it should not automate so much deduction that nothing remains to solve. It should make your own reasoning visible.
A different route through placement puzzles
Clue Manor joins the group of puzzles where placing one thing transforms the space available to everything else. If that feeling appeals to you, the Blupoli Puzzles catalogue lets you explore related placement, pattern and spatial-reasoning games. Categories overlap because a single puzzle can exercise several kinds of thinking at once.
The manor contributes a particular combination: very short written clues, a visual map, explicit candidate notes and one global row-and-column rule linking every guest. The Hider gives the ending a structural payoff instead of adding one more ordinary clue.
You can play Clue Manor on Blupoli Puzzles and start with 4×4 Easy to learn the interaction. Once pencil → place → propagate becomes natural, increasing difficulty changes how much information you receive without changing the rules you already understand.
The mystery ends when there is nothing left to guess
Clue Manor works because each layer has a distinct job. Clues describe areas. Rooms and furniture shape the space. Rows and columns connect every guest. Pencil marks preserve hypotheses. Placements propagate restrictions. The Hider appears only when all of those pieces agree.
The most rewarding solve is not necessarily the fastest one. It is the one where you can follow the chain of reasons from an incomplete statement to an unavoidable position. Sometimes that chain is two steps long; sometimes it crosses half the manor. The underlying discipline remains simple: do not close a possibility too early, and do not leave a new certainty unpropagated.
When the final guest falls into place and the unused intersection appears, the board stops being a collection of rooms and becomes a complete proof. The mystery is solved not because you guessed a hiding place, but because no other position can satisfy everything the manor has told you.