迷宮探險

Maze Explorer: gameplay and practical tips

Navigate a freshly generated maze and reach the star at the exit. This guide reflects the controls and win conditions in the current browser build. It is reviewed by the PocketJoy game-testing team and maintained by Long-Term New Media Limited.

Gameplay and rules

Start at the top-left corner, cross a randomly generated maze and reach the starlit exit at the bottom right. Move with the arrow keys, WASD or the on-screen buttons, all three being equivalent. Three maze sizes are available, every regeneration produces a brand-new random layout, and the generation algorithm guarantees a route to the exit, so you never need to wonder whether the puzzle itself is faulty. The hint function marks the next cell on a correct route when you are thoroughly lost. Time and steps are both recorded and kept separate, because they measure different things: moving quickly but wandering inflates your step count, while thinking carefully and taking the shortest route gives few steps and a longer time. A fastest time is stored per size and they never overwrite one another.

Practical tips

In an unfamiliar maze the most reliable general method is wall following: from the entrance, keep your right hand — or your left, consistently — against the wall and always turn to the same side at junctions. In a maze whose walls are all connected, which is the class this generator produces, that always reaches the exit, at the cost of a route that may be far from the shortest. If you are chasing a time, a better approach is to spend two or three seconds scanning the whole layout, tracing a rough route backwards from the exit with your eyes and memorising a few key turns before you move. Working backwards usually finds the main corridor faster, because branches thin out near the exit far more than near the start. While moving, learn to recognise dead ends: a corridor that narrows with no branches beside it is probably a dead end and rarely worth walking to the bottom. On large mazes, memorise in segments — fix an intermediate waypoint, reach it, then plan the next leg, rather than trying to hold the whole route from the outset.

Rookie mistakes

The usual problem is heading straight for the exit on instinct, changing direction at every wall, and circling the same region repeatedly. The shortest route through a maze often requires moving away from the exit first, and straight-line thinking is a trap here — on a large maze, the corridor pointing at the exit is quite often the longest dead end on the board. The second mistake is relying on hints, which only show the next cell; following them step by step means abandoning your own judgement and typically lengthens your time, since you stop to press the button at every square. The third is not tracking where you have been. The most common human error in a maze is re-walking a dead end, and without deliberately noting that a corridor has already been tried, you will make the same wrong choice at the same junction repeatedly. The fourth is choosing the wrong size: large mazes reward planning, so if you simply want a quick round, the small size produces better results and more satisfaction.

Who it's for

Suited to players who enjoy spatial awareness and exploration. Because mazes are generated randomly, every run is new and memorising layouts is impossible — you are training general wayfinding rather than knowledge of a particular level, which is quite different from a fixed-level puzzle game. The maze is among the oldest puzzle forms humanity has, traceable archaeologically to Crete and Egypt thousands of years ago, while computer maze generation is a classic application of graph theory; the common algorithms — recursive backtracking, Prim, Kruskal — all produce connected, loop-free mazes, which is precisely the mathematical precondition that makes wall following work. The small size works for a quick round while the large one demands real planning and patience. There is no fail state — a wrong turn just means walking back — and the timer is only a reference for comparing with yourself.

Game features

Three maze sizes are available and a fastest time is stored for each, so they can be challenged in turn without overwriting one another. The run panel shows time and steps separately, and keeping them apart is meaningful: you can move quickly but wander, or think slowly and take the shortest route, and only both figures together describe a run properly — judging by one alone misreads your own standard. Every regeneration produces a brand-new random layout, and the algorithm guarantees a route exists, so the puzzle itself is never in doubt; that also means being stuck is always a matter of your search method rather than luck. The next-step button marks the following cell on a correct route, useful for reopening a hopeless position, though showing only one cell means it never solves the maze for you. The start is fixed at the top-left corner and the goal is the star at the bottom right, so the direction of travel is clear and the exit never has to be found.

Controls and score records

Three inputs are supported: arrow keys, WASD, or the on-screen direction buttons. Phone players can play in full using the buttons with no external keyboard, while desktop players may use either arrow keys or WASD, which are exactly equivalent and chosen by preference. All three can be mixed within one run. Walking into a wall simply does not move you and carries no penalty, so probing routes costs nothing. A fastest time per size is stored in the browser's local storage on your own device, with no account required and nothing uploaded to any server; records do not travel to another device, and clearing browser data removes them. The language button switches the interface between Traditional Chinese and English instantly, updating every panel label and prompt. Once the page has loaded, the game itself plays offline.

Gameplay Q&A

Reviewed by the PocketJoy game-testing team · Operated by Long-Term New Media Limited · hklongtermad@gmail.com