47 km a Minute

What is being measured, and what is not

This does not test whether astrocartography works. It measures how far a stated output moves when an input is uncertain — an internal-consistency measurement. Anyone quoting it should be given that sentence: it says nothing about whether the technique predicts anything, and reading it as though it does will get the reader corrected.

What it does establish is a limitation of the technique itself, ours included. At a birth time rounded to the hour, a line cannot support a claim about a named city almost anywhere on Earth. That is not a fact about Cosmiclogy. It is a fact about astrocartography.

Method

The study runs the product’s own path — cosmiclogy_relocation.assess_confidence, not a reimplementation — over a 192-city gazetteer and six sample charts, at two precisions that differ only in how the time was stated: 09:07, as read off a document, against 09:00, rounded.

Displacement is the span between the two edges of the uncertainty window a stated time implies, recomputed at each city’s own latitude. A time given to the minute implies ±1 minute; a time rounded to the hour implies ±30.

The ratio between those two line figures is 29.988 against an uncertainty ratio of exactly 30. That is arithmetic restating its own input, not a finding, and it is recorded here so that nobody quotes it as one.

The error is not uniform — it scales with cos(latitude)

A reading for Nairobi or Singapore is roughly twice as sensitive to a rounded birth time as the same reading at 45–60°. No tool in this category tells anyone that.

The gazetteer holds almost nobody above 60°, so that band is excluded here rather than printed: its median would rest on three cities. The gradient is visible across the four bands that carry 34–62 cities each, and the comparison to draw is 0–15° against 45–60°.

Why bands survive and lines do not

A line runs where one body stood on one of the four angles at birth, and the angles are pinned to the Earth’s rotation — so a line inherits the whole of a birth time’s uncertainty. A band is a latitude where two bodies meet a shared condition, and it barely moves: 0.3–0.4 km at every latitude measured, against four figures for a line.

That gap is why discarding lines and reading bands at a rounded birth time is a different reading rather than a hedged one. It is stated here as the order-of-magnitude difference it is; dividing one by the other produces a ratio with more digits than either number earns.

The data

All 2,304 rows are published — one per chart, per city, per stated precision, carrying the per-city displacement figures the medians above are taken from, the reading tier assigned, and whether the line survived its orb.

Download the dataset (JSON)

Published under CC BY 4.0. Generated 2026-09-04 by scripts/study_birth_time_sensitivity.py. If you quote a figure from it, the sentence to carry with it is the one at the top of this page: this measures internal consistency, not accuracy.

Common questions

How much does a rounded birth time change an astrocartography map?

Every minute of uncertainty moves a line about 47 km — a median of 47.4 km per minute across 2,304 assessments, ranging from 24.3 to 55.8 with latitude. Round to the nearest hour and you have inherited ±30 minutes, which puts the line somewhere in a band roughly 1,400 km wide.

Does this study test whether astrocartography works?

No. It measures how far a stated output moves when an input is uncertain — an internal-consistency measurement. It says nothing about whether the technique predicts anything.

Why are some latitudes more affected than others?

The error scales with the cosine of latitude, so a reading near the equator is roughly twice as sensitive to a rounded birth time as one at 45–60°.

Can I download the data?

Yes — all 2,304 rows, as JSON, linked above.

Keep reading