A hard Klotski puzzle is a sliding-block challenge whose difficulty comes from the rules rather than from cultural naming: every block keeps its fixed footprint, every accepted action shifts the chosen rectangle by exactly one orthogonal cell, and the only way to escape is to thread a 2 by 2 block through a two-column opening on the bottom edge. On a compact 4 by 5 board with six disclosed starting pieces, the original fixture presented by the Klotski tool has a provable minimum of six atomic moves. That minimum is not a community claim or a leaderboard estimate; an independent breadth-first search discovered 3,234 unique board states and expanded 1,470 queue entries before the first winning state dequeues at depth six, which rules out every route of five or fewer atomic moves.
The reason such puzzles feel hard is that the rule set strips away every shortcut a casual player might try. There is no diagonal slide, no rotation, no jump, no teleport, and no two-cell stride. A 1 by 1 block can only nudge itself into a single empty neighbour, and a 1 by 2 pillar must move as a rigid two-cell rectangle. The 2 by 2 goal is the only block large enough to fill the exit footprint, so the puzzle reduces to a planning problem: how do you clear the central lane without ever wedging another piece into a corner that you will need to back out of later?
What Makes the Original Fixture Hard
The original Lizely fixture deliberately compresses the difficulty into 11 occupied cells and 9 empty cells. Every empty cell counts, because every atomic move needs at least one vacant neighbour in the direction of travel. The board is bounded to four columns and five rows, columns indexed zero through three, rows zero through four, so a wrong move can dead-end a rectangle against the outer wall with no escape route.
The piece mix itself is the second source of difficulty. Two 1 by 2 pillars sit on the outer columns at row zero, framing the 2 by 2 goal. Three 1 by 1 blocks (Slider, Guard, and Gate) occupy the middle row at x=1 y=2, x=0 y=2, and x=2 y=2 respectively. That row is the gatekeeper: until Slider at x=1 y=2 clears the goal's lower-left lane and Gate at x=2 y=2 clears the lower-right lane, the 2 by 2 goal cannot descend from y=0 to its target footprint at y=3. Hard Klotski puzzles succeed or fail on whether the solver recognises that the middle row, not the goal, is the real bottleneck.
Piece Identity and Starting Coordinates
| Piece | Size | Top-left corner | Role |
|---|---|---|---|
| Goal | 2 by 2 | x=1, y=0 | Must reach x=1, y=3 (the two central columns of the bottom row) |
| Left pillar | 1 by 2 | x=0, y=0 | Blocks left lane from row zero to row one |
| Right pillar | 1 by 2 | x=3, y=0 | Blocks right lane from row zero to row one |
| Slider | 1 by 1 | x=1, y=2 | First piece to move in the optimal route |
| Guard | 1 by 1 | x=0, y=2 | Sits on the left side of row two |
| Gate | 1 by 1 | x=2, y=2 | Clears the lower-right lane in the optimal route |
Every piece identity is fixed, every starting coordinate is fixed, and every dimension is fixed. There is no random shuffle, no configurable difficulty, and no alternative board state. The hardness lives in the geometry, not in surprise.
How to Solve the Hard Klotski in Six Atomic Moves
Use the following route on the Klotski puzzle tool. The moves are written in the same disclosed order the production code executes them, and each line is exactly one atomic step.
- Select Slider (number key 4, or its selector button) and move Down into row four, freeing the cell at x=1 y=2.
- With Slider still selected, move Left into column zero, parking Slider at x=0 y=3 and clearing the goal's lower-left lane.
- Select Gate (number key 6) and move Right into column three, parking Gate at x=3 y=2 and clearing the goal's lower-right lane.
- Select Goal (number key 1) and move Down into row one, placing the goal at x=1 y=1.
- Move Goal Down again into row two, placing the goal at x=1 y=2.
- Move Goal Down a final time into row three, placing the goal at x=1 y=3 and completing the disclosed exit footprint.
The board awards exactly 1,000 points the moment the final atomic move resolves. Piece selection, movement, and Undo all freeze at that point. The only legitimate way to play again is to press Restart, which restores all six original positions, selects the goal, empties history, resets the move count to zero, and returns the score to zero.
Why Six Is the Provable Minimum
Many sliding-block puzzles claim a "minimum" that turns out to be folklore. This fixture is different because an oracle-style breadth-first search, which does not call the production move or collision helpers, exhaustively enumerated the reachable state space. Each state was stored as a flat coordinate vector in the published piece order. For every dequeued state, the oracle independently filled an occupied-cell set for the five non-moving pieces, attempted Up, Down, Left, and Right for every rectangle, rejected out-of-bounds or colliding footprints, and enqueued each unseen coordinate vector once.
That search did not stop until the first winning state dequeued at depth six. Before that moment it had discovered 3,234 unique states and expanded 1,470 queue entries. Therefore no route of zero, one, two, three, four, or five atomic moves can reach the bottom opening. The six-move route listed above is not one shortest path among many guesses; it is a route whose minimum is mechanically verified.
The test that ships with the production code replays the exact sequence Slider down, Slider left, Gate right, Goal down three times through the real move and collision helpers, then validates every intermediate board. Undo is then exercised as a regression: one move followed by Undo must return to the published starting state, and the immutable snapshot logic guarantees it.
Safe Experimentation With Undo and Restart
Hard Klotski puzzles often require several false starts before the optimal route clicks. The tool supports that learning loop without ever corrupting the starting state.
- Tap any rectangle on the board, or press its number key (1 through 6), to select a piece. The selected rectangle gains a visible outline and is named in the status text.
- Use the four large direction controls, or the arrow keys, to attempt a one-cell shift. If the destination footprint is out of bounds or collides with another piece, the move is rejected atomically and nothing changes: positions, history, move count, score, and completion all remain exactly as they were.
- Press Undo (a full-width control beneath the direction pad, or the U, Delete, or Backspace key) to roll back the most recent accepted move. Undo removes the newest snapshot and restores its exact coordinates, reducing the move count by one.
- Press Restart at any time to reset to the disclosed starting state, even after a successful 1,000-point run.
- Repeated keydown events are ignored, so holding an arrow cannot race a piece through several unreviewed states.
Native selector buttons remain reachable to keyboard users via Tab and activate on pointer events, so the same controls work for touch, mouse, and keyboard players.
Atomic Rules and What They Rule Out
Every accepted action changes a rectangle's top-left corner by exactly one orthogonal cell. The rectangle keeps its disclosed width and height, never rotates, and never skips a cell. Before any move is installed, the logic constructs the complete destination footprint, confirms it stays inside the 4 by 5 board, and confirms it shares no cell with any other piece. If either check fails, the move simply does not happen.
That atomic definition is what makes the six-move minimum meaningful. A version of the puzzle that allowed two-cell strides or 90-degree rotation would have a much shorter route and a different proof; this fixture does not, and the contract is explicit that the six-move result applies only to the disclosed original fixture and its exact atomic-move definition. It does not claim equivalence with any culturally named layout or external puzzle collection. If a reader encounters a different Klotski variant elsewhere, the move count, the minimum, and the BFS statistics are not transferable.
Scoring, Privacy, and Local Play
Scoring is deliberately flat. Reaching the exit awards exactly 1,000 points; nothing else changes the score. There is no speed bonus, no move penalty, no hint cost, no partial credit, and no random modifier. The shared game shell also tracks the best score locally and supplies a double-Escape boss key: press Escape twice within the shell timing window to cover the puzzle with a spreadsheet-style screen, then twice again to return.
All gameplay runs locally in the browser. No board state, move history, result, or interaction data is uploaded. The game needs no account, dependency, remote API, device permission, or paid service. The puzzle is recreational spatial-planning practice, not a standardised intelligence, aptitude, educational, or psychological assessment, so the score should be read as a completion marker rather than a measure of ability.
For a deeper look, see Hard Knights and Knaves Puzzles: Solve Five to 1,000.
For a deeper look, see Common LITS Puzzle Mistakes and How to Avoid Them.