Amaze!

Two moves tell you almost everything

The first board we reached was smaller than the screenshots on most game listings: an L-shaped corridor wrapped around a white central void. The ball began in the lower-left corner. There was no score chase, decorative map, or long explanation competing for attention. The game placed a finger prompt next to the board and waited for a direction.

We clicked the game canvas, pressed Right, and then pressed Up. Right sent the ball across the entire bottom corridor and painted every square red. Up then carried it to the upper-right stop and painted the right edge. The ball did not move one cell at a time. It continued until the corridor ended.

That small detail changes the puzzle. A route that looks like ordinary grid navigation is really a sequence-planning problem. Every command has a fixed destination, and a useful turn must both add paint and leave the ball somewhere that opens the next useful line.

Reading the board before moving

Start by looking for dead ends and long uninterrupted lanes. A dead end is not automatically bad; it can be the only place from which a perpendicular corridor becomes reachable. What matters is whether the next direction can reach unpainted ground.

On the tested L-shaped board, Right followed by Up was productive because both moves covered new cells. Reversing immediately would only send the ball over an already painted lane. As boards become denser, that same test is useful: before committing, ask where the ball will stop and which unpainted corridor will be available from there.

The visual language is unusually readable. Dark squares are unpainted. Red squares are complete. The ball's resting point shows the origin of the next command. Our screenshot is taken after the two verified inputs, so the red bottom and right edges are the result of play rather than promotional art.

How it plays

Amaze! suits short, deliberate sessions. The input is quick, but the useful thinking happens before the key press. There was no timer visible in our first-level test, so we could inspect the board without pressure.

The publisher says mouse gestures and WASD are also supported. We did not need those routes for this test; the recorded sequence used the arrow keys. We also did not test later mechanics, alternate modes, or mobile gestures, so this note makes no claims about their behavior.

A useful first-session rule

Do not judge a move by how many squares it paints now. Judge it by the stop it creates. A shorter move that ends at the mouth of an untouched corridor can be more valuable than a long sweep that strands the ball against painted walls.

Our test stopped after establishing the movement rule and capturing the changed board. We did not extrapolate a difficulty curve or claim a level count from that short run. What we can say confidently is narrower and more useful: directions resolve as full slides, the paint state is immediately legible, and a two-move opening already rewards planning one stop ahead.