Skip to content
Lizely

Ice Sliding Maze

Plan complete straight-line slides through five original ice grids without stopping early.

Privacy: your files never leave your device. All processing happens locally in your browser.

How to use

  1. 1.Use an arrow key or the touch direction pad; the puck will slide in that direction until a wall, block, or board boundary forces it to stop.
  2. 2.Plan from one possible stopping square to the next, avoiding cracked stops; crossing the STOP square does not count unless the puck finishes there.
  3. 3.Rest on STOP to earn 200 points and advance, then clear all five fixed rinks for an exact total of 1,000 points.

About Ice Sliding Maze

Ice Sliding Maze is a deterministic route-planning game built around one strict movement rule: after you choose a direction, the puck keeps sliding in a straight line until the next step would leave the board or enter a blocked square. You cannot release the key early, tap twice to brake, or stop on an arbitrary open tile. Every decision therefore changes which walls and boundaries can become your next stopping points. The goal marked STOP counts only when the puck actually rests on that square. Passing across the goal during a longer slide does not clear the rink.

The game contains five original seven-by-seven layouts named Split Shelf, Cold Stair, Blue Switchback, Frost Relay, and Long Thaw. Their walls, starts, goals, cracked stops, and intended routes were authored specifically for Lizely. The shortest solutions increase from three moves on the first rink to seven moves on the fifth. Every cleared rink awards exactly 200 points, so finishing all five creates an exact score of 1,000. The levels are fixed rather than randomized, which makes every route reproducible and lets the automated test suite verify the same puzzles users see.

Movement is evaluated as a stop-state graph. Each reachable stopping square is a node, and each legal arrow input creates an edge to the final square reached in that direction. A separate breadth-first search oracle reconstructs this graph from literal wall coordinates. It confirms that every level is solvable, that the displayed minimum move count is correct, and that each literal solution route reaches the goal in that minimum number of moves. The oracle implements its own slide loop instead of calling the production movement function, so agreement is meaningful rather than circular.

Split Shelf also demonstrates why planning beyond the next slide matters. Its shortest route travels up, left, and down even though the goal begins on the same bottom row as the puck. Sliding directly right stops on one cracked square. Sliding up and then right stops on a different crack. These two error routes are independently reachable and do not intersect the normal winning route. A crack resets the puck to that rink's start. Repeating the same crack signature does not add another mistake, while reaching a second different crack closes the run until Restart.

The board communicates the rule without hiding essential state. Solid dark blocks are impassable. The puck is labeled ICE, the destination is outlined and labeled STOP, the most recent traversed cells receive a restrained accent, and cracked stopping cells use a visible cross. The telemetry panel reports cleared rinks, total committed slides, the current row and column, level moves, and crack count. A boundary or adjacent wall that prevents any movement is an atomic no-op: it does not increase the move counter, change the last path, or create a false failure.

Keyboard players can use the four arrow keys anywhere on the game. Touch and pointer players receive a matching four-button direction pad with at least 44 pixels of height per control. Both input paths call the same reducer and therefore obey identical stopping, scoring, trap, and terminal rules. The shared GameShell supplies Restart, optional local best-score storage, and the double-Escape boss screen. Once the fifth rink is complete or the second distinct crack has been reached, gameplay is frozen until Restart so extra events cannot alter the final state.

This is a discrete recreational puzzle, not a simulation of friction, momentum, acceleration, temperature, or real ice. It does not estimate travel time or physical distance. No player input leaves the browser, no account is required, and no network gameplay service is called. The product value comes from transparent movement rules, deliberately authored route geometry, exact score progression, and an auditable solver that checks the levels independently.

Methodology & sources

Each arrow input advances through open cells until the next step would cross the seven-by-seven boundary or enter a wall. A goal clears only when it equals the final stop. Wall and boundary collisions with zero displacement preserve the exact state. Five literal routes are checked against a separately implemented BFS over stop states, proving shortest lengths of 3, 4, 5, 6, and 7 moves. Two first-level cracked stops are reachable through distinct routes and excluded from the winning path. Crack signatures combine level and stop coordinate for dedupe; the second unique signature deadlocks. Clears award 200, terminal actions freeze, and Restart returns the canonical initial state.

Frequently asked questions

Can I stop the puck in the middle of a slide?
No. Each direction input commits to a complete straight slide. The puck stops only before a wall, block, or outer boundary.
Why did crossing STOP not finish the level?
The destination is a stop-state target. The puck must rest on STOP after a complete slide; passing over it is not enough.
How are the minimum move counts verified?
An independent breadth-first search enumerates each grid's reachable stopping states and proves that the literal route length is shortest.

Related tools

Mini Games guides

View all