Every other beam game here snaps its light to a grid. Light Ignite does not. Elements are placed freely, beams refract at genuine angles, prisms split white light into its components, and lenses converge or diverge depending on exactly where you put them.

Branching Factor has built a small optics simulation and wrapped levels around it, and the tension between those two things is the whole story of the game — continuous physics makes for beautiful sandboxes and awkward puzzles.

How long a sitting actually lasts

Five minutes if a level yields, twenty-five if it does not, and the difference is not obviously related to how hard the level looks. Continuous positioning means a solution can be a few degrees away from something you already tried.

The sandbox mode changes the shape of a sitting entirely. With no target to hit, I lost half an hour more than once simply arranging prisms, which is a legitimate use of the application and not really a puzzle session at all.

Board readability and art direction

Easily the most attractive game here. Beams glow, dispersion produces real colour gradients, and lens caustics are rendered properly rather than faked. On an OLED screen in a dark room it is genuinely lovely.

It is also the least legible. Overlapping beams sum towards white, colour does load-bearing mechanical work, and on a busy board it is hard to tell which of four rays is the one that needs to reach the target. Beauty and readability are in direct conflict here and beauty won.

Not for everybody

Players who want a puzzle with a definite solution will find this maddening. Continuous placement means there is rarely one correct configuration — there is a region of configurations that work, and finding its edge is fiddly rather than clever.

It is also a poor choice for anybody with colour vision differences, because the dispersion levels require telling beam colours apart with no secondary cue at all. That is a real exclusion and the settings do not offer a way around it.

The level list and the long tail

Levels introduce elements steadily — mirrors, then lenses, then prisms, then portals and black holes — and each addition genuinely changes what a solution can look like. The teaching is good and the variety is real.

What does not progress is the precision demand, which stays high throughout. There is no point at which the game becomes forgiving about placement, and for some players that will make the later levels a test of patience rather than of understanding.

Small screens and large hands

Landscape and two-handed in practice: one hand holds, one places, and the placement is fine enough that a thumb on a moving train will not do. This is the least commute-friendly game in the catalogue.

Screen size matters more here than anywhere else. On a compact phone your fingertip covers a meaningful share of the board and the fine rotations become guesswork; on a tablet the same levels are comfortable. If you only have a small phone, this is the one game here I would actively steer you away from.

Three frames from the game

What works, and what does not

The verdict

A gorgeous simulation and an uneven puzzle. On a tablet, in the evening, with the sandbox open, it is one of the most pleasant things here. As a phone puzzle on a bus it is close to unusable, and the score reflects both halves.

Scores on this site run from 6.1 to 8.7 and are one person’s judgement. How the scoring works.