How Accurate Is a Star Moment Map? Our Method, Data and Limits

Our stars come from the HYG database, an open compilation maintained at astronexus.com that merges three standard catalogues: the European Space Agency's Hipparcos catalogue (precise positions and parallaxes), the Yale Bright Star Catalog, fifth edition (the classic list of naked-eye stars with their Bayer letters and Flamsteed numbers), and the Gliese catalogue of nearby stars. The name is the initials. HYG is published under a Creative Commons Attribution-ShareAlike 4.0 licence, and we credit it here as that licence requires.

From HYG we take every star brighter than apparent magnitude 6.5. That is about 9,000 stars, and it is the conventional limit of unaided human vision under a dark, clear sky. We draw brighter stars larger, on a scale derived from magnitude, so that the familiar patterns (Orion's belt, the Plough, the Southern Cross) read at a glance the way they do outdoors. We do not add stars that the person standing there could not have seen, and we do not remove any that they could.

For each star the catalogue gives us a right ascension, a declination and a magnitude. Positions are referenced to the J2000.0 epoch. Everything below starts from those numbers.

A star's catalogue position is fixed to the sky. What you see from the ground depends on where you are standing and how far the Earth has turned. The conversion runs in four steps, and every one of them is textbook celestial mechanics; none is proprietary.

1. Convert local time to Universal Time. The date and time you enter are local to the place you chose. We resolve the place's time zone, including the daylight-saving rules in force on that date, and convert to UTC. The next section explains why this step matters more than any other.
2. Compute Greenwich sidereal time for that UTC instant, using the standard IAU expression for sidereal time as a function of the Julian date.
3. Compute local sidereal time by adding the observer's longitude (east positive). Local sidereal time is the right ascension currently crossing the meridian; it is the number that ties the fixed sky to your spot on the turning Earth.
4. Convert each star to horizon coordinates. From the star's right ascension and declination, the local sidereal time, and the observer's latitude, we compute the hour angle and then the altitude and azimuth using the standard spherical-trigonometry transform. Stars with altitude above zero are above the horizon and are drawn; the rest are not.

Before step 4 we apply precession to move each star's J2000.0 position to the coordinate frame of the requested date. For dates within our supported range this shifts positions by up to about a degree, which is small at poster scale but not something we skip. We do not currently apply proper motion (see the limits below).

We then project the visible hemisphere onto the page with a stereographic projection centred on the zenith. The horizon becomes the outer circle; the point directly overhead is the centre. Stereographic projection preserves angles, which is why constellation shapes look correct, at the cost of enlarging scale toward the rim. This is the projection used by planispheres and by most planetarium software's overhead view, and it is why a Star Moment map and a Stellarium screenshot of the same sky look alike.

The Earth turns 15 degrees an hour. An hour's error in the time moves every star on the map by 15 degrees, enough to put a constellation on the wrong side of the meridian. So the two inputs we ask for most carefully are the time and the place, because together they fix the local sidereal time.

The place is resolved to a latitude and longitude. You can pick a city or drop a precise pin; for a map, the difference between a city centre and a suburb is a fraction of a degree and invisible in print. Latitude decides how the celestial sphere tilts against your horizon and therefore which stars can ever rise. Longitude, with the time, decides how far the sky has rotated.

The time is where most star maps from any company go wrong, and it is usually the time zone rather than the clock. We use a historical time zone database (the IANA tz database, which records boundary and daylight-saving changes back through the twentieth century) so that "9 pm in Houston in 1969" or "11 pm in Lisbon in 1994" resolves to the correct UTC instant even where the rules have since changed. We show you the resolved time zone (for example, America/Chicago) in the designer so you can confirm it before you order.

If you do not know the exact time, choose the hour the moment happened and put it in the caption. A specific time you are fairly sure of is more honest than a default midnight you did not choose.

We draw all 88 constellations recognised by the International Astronomical Union, and only those. The IAU fixed the list and the boundaries in the 1920s, and it is the same set used in every modern atlas and in Stellarium. We use standard constellation line data to join the principal stars in each figure; the lines are a convention rather than a fact, and different atlases join slightly different stars, so if a figure on our map differs by one segment from the one in your childhood book, both are legitimate.

Constellation lines can be turned on or off in the designer. When off, the stars themselves are unchanged; only the lines are hidden. We do not print constellation names on the map. Every map also carries a horizon ring marked with the cardinal directions for your location, so that the orientation is never ambiguous. Because you are looking up through the dome rather than down onto a ground map, east and west are mirrored relative to a road map; the ring makes that explicit.

We show: stars to magnitude 6.5 from the HYG catalogue, sized by brightness; the 88 IAU constellations as lines, optionally; the Milky Way at its real position, optionally; a horizon ring with cardinal directions; your coordinates, date and time as text. The Galaxy design places a nebula photograph behind the chart, and the Moon and Eclipse designs place a photographic disc under or beside it; all three are design elements. They do not show the real phase of the Moon or a real eclipse on your date.

We do not show: the planets, the Moon, the Sun, comets, satellites, or deep-sky objects. Planets and the Moon move against the stars and require their own ephemeris calculation; we have chosen to keep our maps to the fixed stars, and we say so rather than leave it to be assumed. If your memory of the night includes a very bright "star" near the Moon, it was almost certainly Venus or Jupiter, and it will not be on the map.

We also do not model weather, twilight, or light pollution. The map for a summer afternoon shows the stars that were above the horizon behind the daylight, correctly, but nobody saw them. A map for a city centre shows magnitude-6 stars that the streetlights would have hidden. The map records the sky, not the view.

We generate maps for any date from 1 January 1900 to 31 December 2100. Outside that range we decline the order rather than deliver a map we cannot stand behind. The reasons:

Precession. The Earth's axis wobbles in a 26,000-year cycle, drifting the celestial coordinate grid by about 50 arcseconds a year. Our correction handles this well over a century or two either side of J2000.0; the further out you go, the more the standard series approximations diverge from the true motion.

Proper motion. Nearby stars move measurably through space. We do not currently apply proper motion; we plot catalogue positions corrected for precession only. Within 1900 to 2100 the effect is below the width of a printed star for all but a handful of fast, nearby stars (Barnard's Star, 61 Cygni, and to a lesser extent Arcturus), none of which will be noticed on a map. Over longer spans it would be, which is one more reason for the date limit.

Time zones. Before the twentieth century, standard time zones did not exist in much of the world; local mean time varied town by town. Our time zone database is reliable back to roughly the 1880s in most places and less so before, and a wrong offset is a wrong map.

Within 1900 to 2100 none of these effects is visible at poster scale. That range covers every date a living person has a memory of and every date their grandchildren will. We would rather draw the line clearly than sell a map of the sky over the Battle of Hastings and hope nobody checks.

Two smaller limits, for completeness. Our catalogue cut at magnitude 6.5 means we do not draw the fainter stars that binoculars would reveal; a company plotting to magnitude 7 will show more dots in the same field, which is a design choice rather than an accuracy difference. And a stereographic map enlarges the sky near the horizon relative to the zenith, so constellations at the rim look bigger than the same constellations at the centre; this is the projection, not a mistake.

Stellarium is a free, open-source planetarium used by amateur astronomers and in public planetariums. It runs on Windows, macOS and Linux, and a browser version at stellarium-web.org needs no installation. Checking one of our maps against it takes about ten minutes, and we encourage it before you order.

1. Open stellarium-web.org.
2. Set the location. Use the location control at the bottom left: search for the city named on your map, or enter the latitude and longitude printed on it.
3. Set the date and time. Use the time control at the bottom right and enter the date and local time from the map. Stellarium applies the time zone for the location you set; it should be the same zone our designer shows.
4. Turn on constellation lines and labels from the bottom toolbar. Turn off the atmosphere if the moment was in daylight so the stars are visible.
5. Look straight up. Drag the view to the zenith, or use the overhead view. Align using the N, E, S, W marks on our horizon ring; north is at the top of every map.
6. Compare: which constellation is highest; which bright star is lowest in the west; whether a recognisable figure (the Plough, Orion, Cassiopeia, Scorpius) sits in the same part of the sky on both.

If those agree, the map is correct to any standard that matters for a print. If a constellation is on the wrong side of the sky, the time or time zone differs between the two; check the time zone first. If you find a genuine discrepancy, write to us through the help page with the map link and the Stellarium settings, and we will investigate and, if we are wrong, fix it and tell you.

You can run this check on the free preview in the designer before paying anything. That is the point of the preview.

• HYG database, David Nash, astronexus.com (Hipparcos, Yale Bright Star Catalog 5th ed., Gliese 3rd ed.; version 4.2, 2025; CC BY-SA 4.0): https://www.astronexus.com/hyg
• ESA Hipparcos mission (catalogue of 118,218 stars): https://www.cosmos.esa.int/web/hipparcos
• Yale Bright Star Catalog, 5th revised edition (Hoffleit and Warren, 1991), via NASA HEASARC: https://heasarc.gsfc.nasa.gov/W3Browse/star-catalog/bsc5p.html
• IAU, "The Constellations" (the 88 constellations; boundaries adopted 1928, published 1930): https://iauarchive.eso.org/public/themes/constellations/
• IANA time zone database: https://www.iana.org/time-zones
• Stellarium (desktop) and Stellarium Web: https://stellarium.org/ and https://stellarium-web.org/
• Jean Meeus, Astronomical Algorithms, 2nd ed. (Willmann-Bell, 1998), for sidereal time, precession and the equatorial-to-horizontal transform.
• P. Kenneth Seidelmann (ed.), Explanatory Supplement to the Astronomical Almanac (University Science Books), for reference definitions of the quantities above.

Related: [how it works](/how-it-works), [star map designs](/designs) and [pricing](/pricing).