Why two astrocartography maps disagree.

This is not a story about one site being broken. Both of the big free astrocartography tools compute real astronomy, and on the underlying question of where the planets actually were, they agree closely.

They disagree because astrocartography has more than one defensible definition of a line, more than one way to turn your birth certificate into an exact instant, and more than one way to measure how far you live from a line. Each tool picks. Almost none of them tells you which pick it made.

A reader is then left comparing two pictures with no way to account for the difference. That is the actual problem, and it is a documentation problem before it is a mathematics problem.

This is the big one, and it is documented by the tools themselves. Astro-Seek's own FAQ states that it computes astrocartography with what it calls the Zodiacal or Ecliptic projection method by default, and notes that other software defaults instead to the original Mundo or Latitude method developed by Jim Lewis. Astrodienst describes its own maps the other way round, saying that in astrocartography the real positions of the planets in the sky are used, not just their ecliptic points as in a normal chart.

In plain terms. The first convention draws a rising line where the planet was genuinely coming over the horizon. The second draws it where the planet's zodiac degree matches the degree that was rising. Those sound like the same sentence. They are not, because a planet is rarely exactly on the ecliptic, the flat plane the zodiac is measured along. The further off it sits, the further apart the two lines fall.

We measured the gap on Einstein's chart with our own engine. Pluto's Midheaven line lands 3.3 degrees of longitude apart between the two conventions, about 280 km of ground at the latitude of New York or Madrid. Saturn's moves about 0.86 of a degree. The Sun's does not move at all, because the Sun is by definition always exactly on the ecliptic.

That pattern is the fingerprint of this cause, and it matches what people report seeing: the outer planets and the Moon visibly shifted between two maps, the Sun sitting in exactly the same place on both. If your two maps differ most on Pluto and not at all on the Sun, you have found a convention difference and not a bug.

Our choice: the Lewis convention, where the line marks where the planet was really rising, setting, overhead or underfoot. It is the standard the technique was invented with, and it is what the phrase "a planet on your horizon" literally means. The other convention is not wrong, and it has a genuine virtue we do not have: its lines correspond exactly to the angles of a relocated chart.

A birth certificate gives a local clock time. Turning that into a moment in universal time means knowing what the clocks in that place were set to that day, which for older births is genuine history rather than arithmetic. Wartime clock changes, countries that adopted standard time late, cities that ran on their own sun until a railway told them not to.

Get that wrong by a few minutes and every line on the map slides sideways together. We know the size of this one from our own mistake. Einstein was born in Ulm, and our resolver used to apply the local mean time of Berlin, the city that names his modern time zone, rather than Ulm's own. That is 13 minutes and 28 seconds of error. It moved all forty of his lines by about 3.4 degrees of longitude, which is roughly 250 to 290 km on the ground at the latitudes his reading discussed. Every other test in our suite passed while that was true, because natal planet positions barely move in 13 minutes. Only the lines do.

Our choice: the birth instant is resolved from the birthplace itself, not from the label of its modern time zone. Births before standard time use that town's own local mean time, four minutes of clock for every degree of longitude. Ambiguous times, the repeated hour when clocks go back, are detected and recorded rather than silently resolved.

Astrology software carries defaults that change results and are rarely surfaced. The best known in this category is the lunar node, which comes in a true and a mean flavour that differ by a degree or so, enough to move a node line but nothing else. Ephemeris choice contributes a much smaller amount, well under a kilometre for the bodies we draw.

One thing that gets blamed and should not be: the house system. Placidus, Koch and whole sign disagree about where the cusps of houses two, three, five and six fall. They do not disagree about the horizon or the meridian, so they do not move an astrocartography line at all. If two maps differ, the house system is not why.

Our choice: we draw the ten classical bodies, Sun through Pluto, and we do not currently draw node lines. Positions are apparent places for the date: true ecliptic and true obliquity of date rather than a fixed reference frame, apparent sidereal time including the equation of the equinoxes, plus the corrections for light travel time and aberration.

Even with identical lines, two tools can tell you different things about how close you are to one, because measuring the gap between a point and a curve admits more than one answer. This one is worth its own section.

Astrodienst's FAQ phrases its orb as a distance east or west of each line: 150 miles, about 240 km. Read literally, that measures along your parallel of latitude. You travel due east or due west until you hit the line, and that trip is your distance.

We measure the shortest distance to the line instead. Not east or west specifically, but the closest the line ever comes to you, in any direction, over the curve of the Earth.

Picture a line on a map and a town beside it. If the line runs straight up and down, sliding sideways to reach it and walking straight at it are the same trip, and the two measurements agree. Now tilt the line. The sideways slide gets longer while the straight walk stays short, because you are travelling along the tilt instead of across it. Measure sideways and you will report a town as further from a line than it really is, always further, never nearer, and the effect grows the more the line tilts.

Meridian lines, the ones for Midheaven and Imum Coeli, run true north to south, so for those two the distinction does not exist. It is the rising and setting lines that tilt, and they tilt hardest at high latitude.

We found this in our own code, which is how we know the size of it. Uranus's setting line at Stockholm, for a January 1985 chart, measured sideways came out at 690 km. Measured straight at the line it is 399 km. Under our bands the first figure is a faint influence at the outer edge of orb and the second is a real background one, two bands apart, which is the difference between a clause and a paragraph in a paid reading. When we found it, our outer band stopped at 500 km and the first figure was out of range entirely, meaning the line would never have been mentioned at all. Across 48 cities and a spread of birth times, about one in five true in-range rising and setting influences were being dropped or under-weighted the same way. Below 30 degrees of latitude the error was nil. Above 45 degrees it affected roughly one reported line in six.

To be fair to the other reading of it. Astrodienst's east and west phrasing appears in an answer about birth time sensitivity, where the displacement genuinely is east and west, because a clock error rotates the whole map in longitude. They are describing the resolution limit of a printed map rather than defining the shape of an orb, and for the meridian lines most people picture, the two measures are identical anyway.

We chose the shortest distance because a literal east and west orb breaks down where a rising line runs nearly east and west itself, which happens at high latitude, and would imply a band of unlimited width. And because the tradition supports it: Jim Lewis stated his orbs as a distance in miles either side of the line, a uniform band, and demonstrated it by laying his thumb across a line. A thumb measures the short way.

One honest wrinkle. Lewis's orb was a fixed distance measured on paper, which a flat map stretches at high latitude, so his own 700 miles was never quite a fixed distance on the ground. We measure true distance across the Earth's surface, on a single sphere of mean radius, which is more consistent than the practice we inherited it from.

Practitioners read a line's influence as a gradient rather than a boundary. Ours, in kilometres, are the numbers every reading is actually computed with:

Every one of those thresholds is a mile figure someone published, converted and rounded to the nearest five kilometres. The scheme is the one Moses Siregar sets out in print: two hundred miles for the strong band, three hundred and fifty for the moderate one, five hundred as the outer range. Only the innermost band is ours, and it is deliberately stricter than anything in the literature, because it is the only band allowed to say the words under the line.

These bands used to be too tight, and this page said so. Our outer limit stopped at 500 km, roughly 311 miles, which was narrower than every source we could find. Lewis himself read a line out to 700 or 800 miles, and Astrodienst's own encyclopedia says about 1,000 km. So we were printing no line in orb for places the whole profession would read. That is now corrected: the outer edge sits at 805 km, and the band between 565 and 805 km has words of its own, because a line at seven hundred kilometres is real but it is faint, and it should not borrow the language of one at four hundred.

We took Siregar's five hundred miles rather than Lewis's seven or eight hundred, and it is worth being clear that this is still the conservative choice. His is the tightest outer limit published anywhere in the field. We would rather be caught being too quiet than caught inflating.

Distance in miles is Lewis's ruler. It is not the only way to judge a line, and two practitioners writing independently say it is not the rigorous one. The rigorous measure is the orb in degrees from the planet to the angle in a chart re-cast for the place you are asking about. It never touches ground distance, so it does not care about latitude and it is immune to the sideways problem above.

So we compute both, and your reading prints both. Within 3 degrees of an angle is strong, 3 to 6 is moderate, 6 to 8 is weak but still worth naming, and past 8 degrees the planet is not angular in that chart at all. Same source as the distance bands.

The two measures disagree, often, and we show you when they do. They are not two views of one number. Your reading's lines are drawn where the planet's own body sits on the horizon or the meridian; the relocated angle is the degree of the zodiac sitting there. For a planet away from the ecliptic, those are different places. A line can pass close to your city while its planet is nowhere near your relocated angle, and a planet can sit right on your relocated angle while its line runs hundreds of kilometres away. Neither number is wrong. Vela is instructed to name both, to pitch the claim at whichever is weaker, and never to average them or convert one into the other. A reading that quietly quoted the flattering one would be easier to write and worth less to you.

A test that compares software against itself proves nothing. We learned that the expensive way: our engine had a full suite of passing tests while two real defects sat in it, because the tests checked our arithmetic against a second copy of the same arithmetic.

So every check below reaches outside the engine for its answer, to a different ephemeris, a different transform, a brute force method with no cleverness in it to be wrong, or a chart somebody else published decades ago. All four run on every change. None of them can be satisfied by our own formulas agreeing with our own formulas.

Historical time zone data is not part of our code. It ships inside the runtime, and a routine platform upgrade can carry a revised copy of it. If that happened silently, a birth instant we already computed and sold could change, and every line on that map would move, with nothing to notice.

So we pin the behaviour. 23 historical offsets are asserted directly, one per class of thing that can go wrong: pre standard time local mean time in three countries, United States war time in 1942, the 45 minute offset Nepal uses, Russia's two changes of mind, both sides of a repeated hour when clocks go back, and Ireland, which the data models backwards from everyone else. On top of those, the resolved birth instants of the three published readings in our examples gallery are pinned end to end. A platform upgrade that would move any of them fails our build instead of reaching a customer.

Everything above is our job. This part is not, and almost nobody tells you the exchange rate, so here it is.

The Earth turns. Your lines are pinned to where the sky was over Greenwich at your birth instant, so an error in the clock does not distort the map, it slides the whole thing east or west. All forty lines, together, by the same amount. About 1 degree of longitude for every four minutes.

Astrodienst publishes the conversion in ground distance, and it is worth reading twice:

Sit with the second one. A quarter of an hour of doubt is wider than our entire peak orb band. It is the difference between a line running through your city and one running through the next country. A certified consultant in this field puts the requirement plainly: a birth time accurate to within 15 minutes is needed to derive meaning from the map at all.

Which is why we ask for a birth time and will not sell a line reading without one. A reading built on a guessed time is a confident picture of somewhere you were not born. The time on your birth certificate beats the time your family remembers, every time, and it is usually obtainable.

If your time is close but not certain, say so during intake. Vela will lead with your steadiest material and hold the fine calls loosely rather than quote a precision your data cannot carry.

The anchors we check ourselves against are public. Anyone with the patience can reproduce the sharpest one without us being involved at all.

Take Einstein's birth record: 14 March 1879, 11:30 local mean time, Ulm. It is rated AA, meaning it comes from the birth certificate, and it is the case every serious astrology reference has published for decades, with an Ascendant of 11°39' Cancer and a Midheaven of 12°50' Pisces. Ulm sits at 10 degrees east, so its own mean time runs exactly 40 minutes ahead of Greenwich and 11:30 there is 10:50 universal time. Cast the chart from that instant in any software you trust and those two angles are what should come out. They are what comes out of ours, to the arcminute, starting from the raw date and town.

Then compare the free maps yourself. Draw the same chart on two tools and look at the Sun lines first. If they sit together and Pluto's do not, you are looking at the convention difference described above, and now you know which one each tool made.

If you want the detail behind any figure on this page, or the output of the checks themselves, ask us. Write to hello@astrocartographers.com and a person will send it to you.