The situation

You woke up somewhere with no idea where. The ship has instruments and fuel. It does not have a map, and nothing is going to give it one.

So you make your own. Declare where you are zero, pick some axes, and start measuring. Everything you ever see is in that private frame — it bears no relation to anyone else's coordinates, including Earth's.

The way home is the almanac: seven pulsars with exact positions relative to Earth. Useless on their own. Match three of them to stars you charted yourself and you can work out where your frame sits in the real universe, and therefore where Earth is.

Nothing here is faked. Every uncertainty the game shows you is one it actually computed, and the difficulty comes from the geometry rather than from the game deciding to be hard.

Charting a star

Ranging a star tells you how far away it is — a sphere around you, somewhere on which it sits. One sphere is useless. Two intersect in a circle. Three intersect in exactly two points.

So a star needs 3 sightings from 3 different places before it can be solved at all, and the only way to get them is to fly somewhere and measure again. Wandering is not flavour around the mechanic — wandering is the mechanic.

• Range all in reach measures everything within 150 pc, at 1.5 hours each.
• Ranging the same star twice without moving adds nothing — the second sphere is the first sphere. It replaces rather than accumulates.
• Star type matters enormously. A pulsar has a sharp timing signal and charts far better than a variable. The star card shows the noise for each.

On the chart, each charted star is a pin with a stalk down to the reference plane so depth is readable, and a wireframe sphere that is its genuine 1-sigma uncertainty — the star is inside it about two thirds of the time.

The flat-track trap

Three vantage points define a plane, and three spheres meet at two points mirrored through that plane. Both fit your measurements perfectly. The game cannot tell them apart and neither can you.

That is what a red pair on the chart means: two candidate positions, dash-linked, often hundreds of parsecs apart. Picking the wrong one is a real mistake, not a rounding concern.

The fix is the climb slider. Left at zero your flight path stays flat, every vantage point lands in the same plane, and nothing ever resolves. Move off that plane and the ambiguity collapses.

The star card's off-plane baseline is the number that governs it. If a solve cannot separate the two candidates, the game says so rather than guessing — you will never be shown a confident position that is secretly a coin flip.

Staying found

A jump goes roughly where you aimed it, not exactly. Dead reckoning does not know about the error, so it accumulates: every jump you are a little more lost than you think, and Your position ± tracks how much.

This compounds into everything. A sighting taken while 8 pc adrift is an 8 pc wrong reference, permanently — you cannot chart a star better than you knew where you were standing.

Take a fix is the cure: range stars you have already charted and solve backwards for yourself. It is only as good as the chart it is taken against, so fixing against a poor chart tells you little, and the game will say so rather than pretend.

Errors add in quadrature, so ten jumps leave you about three times worse, not ten. The second half of a long run costs proportionally less certainty than the first — which is what makes "when do I stop and fix" a real decision.

Your two instruments

They fail in different directions, so neither replaces the other.

• The scanner times a star's own signal. Reaches 150 pc and stops dead. Precision depends on the star's type — superb on pulsars, poor on red dwarfs. 1.5 hours a star.
• Parallax watches how far a star shifted against the sky since your last stop. Works at any distance, degrades with the square of it, and depends only on geometry — so it beats the scanner on red dwarfs and variables. 4 hours a star, and it needs you to have moved.

Turn on show parallax shift on sky to see the traces. The slider hides stars that moved more than the threshold, which reads backwards until you try it: near stars swing right across the sky and bury the small shifts of the distant ones, and the distant ones are the whole point.

Instruments take damage from approaching hazardous stars unshielded. Damage costs reach as well as precision, which is the sharper loss. Repair sensors costs hours in proportion to how bad it is.

Fuel, damage, salvage

A jump costs 2.5 fuel to start plus 0.11 per parsec. The fixed part matters: a hundred small hops drift almost nowhere but cost a hundred start-ups, so precision by shuffling is available and expensive rather than free.

Approach & scoop flies to the selected star and refuels there. Which stars give up fuel depends on your scoop grade; which you can survive depends on your shielding. Both are in the Survey panel. Diving into something hot without shielding blinds your instruments.

Ruins are derelicts scattered through the field, some beside stars and some in open space. They are cold and dark, so nothing passive will ever show you one — only the active scanner finds them, out to about 60 pc, and only along the path you actually fly. Route matters as much as destination.

Salvaging one is the only way the ship's tiers ever improve. Fly within 6 pc and spend 30 hours. A site is logged where your chart thought you were when you found it, so one discovered while badly adrift is written down in the wrong place.

Finding Earth

You need three almanac pulsars, and each needs two separate things done to it:

• Identified — fly within 18 pc and read its timing signature. That tells you which almanac entry it is, and therefore its true coordinates. Getting that close to a pulsar will hurt your instruments.
• Charted — solved in your own frame, like any other star. A landmark you flew to but never properly charted contributes nothing.

Anomalous sources show as orange diamonds on the sky and in the Almanac panel. Most are not landmarks at all — variables and white dwarfs read as anomalous too — and you cannot tell without closing on them. The star card records which ones you have already ruled out.

With one match you know your distance from a known point. With two, the answer is determined up to a spin about the axis joining them. With three, well spread, it locks — and the Survey panel stops saying Earth: unknown.

Controls

Chart / Sky

Your own map, or the view out of the window.

Drag

Orbit the chart, or look around the sky.

Scroll

Zoom the chart.

Click a star

Select it. Details appear top-left.

Shift-click (sky)

Range that star directly.

bearing / climb

Aim. Bearing sweeps the reference plane, climb lifts out of it. The dashed line shows exactly where you will end up.

distance

How far, up to 120 pc a jump.

Move

Go.

Wander

A random jump, for when you have no better idea.

link aim and sky view

Point the sky at something and the aim follows, and the reverse. Aim by looking.

Esc

Close this guide.

The rail along the bottom holds the legend, the almanac, salvage sites, and everything currently in sensor reach. Each collapses on its own.