How Hazard Telegraphing Works

Telegraphing is the difference between a hard game and an unfair one, and it is more specific than "give the player a warning". There are three distinct kinds, and a game needs all three because they solve different problems.

Three kinds of telegraph

Spatial — the hazard is somewhere you can see before it is somewhere you can be hurt.

Temporal — the hazard becomes dangerous on a schedule you can learn.

Behavioural — the hazard moves along a path you can predict.

A static spike needs only the first. A blinking beam needs the first and second. A patrolling cutter needs all three.

Spatial: nothing spawns on you

The most important rule and the simplest: everything enters from beyond the play area and travels into view.

No hazard is ever created at the player's position, near it, or inside the visible field. Every object crosses the screen edge before it can matter.

That single rule eliminates the majority of unfair deaths in scrolling games, and it is easy to violate accidentally. A generator that places obstacles "somewhere in the next chunk" will eventually place one just ahead of the player at exactly the wrong moment. Ours generate relative to a spawn point beyond the right edge, which makes the failure structurally impossible rather than merely rare.

The same principle applies in maze and platform games, where the equivalent is that the layout is fully visible before you enter it. Hazards are placed at generation time and do not appear later, so anything that kills you in a maze was on the screen when you turned the corner.

The one deliberate exception is hidden collectibles — which are rewards rather than hazards, and which show a faint outline rather than being genuinely invisible.

Temporal: fixed periods

Timed hazards — beams, gates — are dangerous for part of a repeating cycle. Two properties make them fair.

The period is fixed. A beam with a 2.2-second period keeps that period for as long as it exists. It does not vary, drift, or randomise. That is what makes counting possible, and counting is the correct technique.

There is a warning phase. Before the beam becomes deadly it changes appearance, so the transition is visible rather than instantaneous.

The reason to prefer counting over reacting is arithmetic. At 400 pixels per second, reacting to a beam works — you have time. At 760, it does not; the window between the warning and the beam is shorter than a choice reaction. But a counted rhythm is not a reaction at all: you know when the beat falls, so your move is planned rather than triggered.

Neon Gem Run is built on this. Its beam gates run on periods that shorten as the difficulty curve rises, and the energy-stone columns past each gate are placed inside the safe window — so collecting them cleanly is confirmation that you read the rhythm correctly.

Behavioural: one cycle is everything

Moving hazards follow deterministic paths. A patrolling cutter travels a fixed range on a fixed axis at a fixed speed, driven by a sine function — which means its motion is perfectly periodic.

The practical consequence is the most transferable advice in the genre: watch one cycle before committing. That single cycle contains the hazard's range, speed, axis and turning points. There is nothing left to learn afterwards.

A cycle takes one to two seconds. In a stage-based game where nothing moves until you do, watching one costs nothing at all. In a timed game it costs a second or two of clock, which is far less than a wrong guess costs.

Roaming actors in maze games are the interesting variant. They are not periodic — they bounce along corridors and periodically pick a new heading — but they are also deliberately not pursuing. They never track the player's position, which means they are traffic to be timed rather than hunters to be outrun.

That distinction matters for fairness. A pursuing actor cannot be predicted, so the only response is flight; a non-pursuing one can be read, and reading it is a skill.

The test we apply

For any hazard, the question is: could a player who was paying attention have avoided this?

Not "could a perfect player". Not "is it theoretically possible". Could an attentive human, looking at the right part of the screen, have seen this coming and responded in time?

If the honest answer is no, something is wrong, and it is usually one of four things:

• The hazard appeared inside the play area rather than travelling in.
• A timed hazard has no warning phase, or its period varies.
• A moving hazard's path is not deterministic.
• The reaction window has compressed below what a person can answer — usually because a difficulty curve has no ceiling.

Our testing pass for this is deliberately unsophisticated: play badly on purpose, note every death, and ask whether each cause was visible beforehand. Any death where the answer is "not really" is a bug rather than a difficulty setting.

It is worth saying that this does not constrain difficulty much. Ultimate Gem Challenge runs at the highest hazard density on the site — up to 88 per cent of lanes blocked — and every single hazard in it travels in from the edge, follows a visible path, and leaves exactly one lane clear. It is extremely hard and entirely readable, which is the combination worth aiming for.

Frequently asked questions