Refine your birth time from events

From 5, ideally 10 or more; the maximum is 40. But it isn’t only about quantity. An answer “down to the minute” needs at least three independent events from two different areas of life: three marriages are three dates, but they test the same thing. Events in the same area less than 45 days apart count as one: an exam, a hospital stay and surgery within a month are one life event.

No, but the more clearly you remember the day, the more the event narrows the search. A date “give or take a few weeks” helps less than an exact one, and “give or take half a year” won’t find the minute: such an event isn’t counted among those that unlock a “down to the minute” answer. Don’t invent an exact day — pick an honestly approximate date: we count it as imprecise, and it still helps.

It’s an estimate, not proof, and it has one of three levels. “Down to the minute” — when there are at least three independent events from two areas of life and a sufficient model index. “Interval” — when there is a signal in the dates but not enough for the minute. “Working options” — when the dates don’t decide anything yet: these are search points, not an established time. There won’t be seconds: one degree of the Ascendant is about 4 minutes on average, and the dates of events can’t tell finer than that.

No, it isn’t the probability that the time is right. The index from 0 to 100 shows how much stronger the signal of your dates at this time is than the best match that the same events on random dates give anywhere in your interval. With fewer than 7 events it doesn’t rise above 60, and with fewer than 10, above 80. The separate “How the probability is distributed” block splits 100% between intervals of the day based on the dates, your choice of descriptions and your answers to the follow-up questions; this isn’t minute-level precision either, and at the “down to the minute” level this block doesn’t appear.

Two techniques tell minutes apart. Primary directions to the four angles of the chart (Ascendant, Descendant, MC, IC): the Naibod key, conjunctions only, direct and converse, orb 0.5°. A 4-minute time error shifts a direction by about a year — which is why they can see minutes. Solar arcs (1 day = 1 year): planets to angles, and the directed Ascendant and MC to natal planets, orb 0.5°, including squares. On top of that comes a “planet + angle” filter: an event counts only its own planets and its own angles (marriage — Ascendant and Descendant, a move — IC and MC). The scan runs minute by minute — across the whole day or within your interval, with no grid.

At the bottom of the result page there is a collapsed “For astrologers” block. It holds the numbers and terms professional astrologers use: techniques, orbs, plus transits and the progressed Moon. They don’t tell neighboring minutes apart: the progressed Moon is shown for information only, and transits enter the calculation with a small weight, only on the planets and angles relevant to the event. The main screen speaks plain language.

For every request we build a control. We shift each of your events by 1–6 years in both directions — 128 versions — and shuffle them 1,000 times. From this comes the p-value: how often random dates would give the same score somewhere in your interval, not just at this exact minute. We pick the position by its excess over the control, not by the raw score: daytime hours have more reachable arcs, so on any dates they would score higher. We also check stability: we drop one event at a time and shift the dates by 3, 15 and 120 days — if the peak holds, the index is higher.

No. The time calculation doesn’t use a language model: it’s astronomical computation and statistics, so the same data always gives the same result. The description of your look is assembled from ready-made texts and is the same for the same time.

Yes. The catalog has over 330 events across 27 areas — from moves and marriage to service, business, creative work and animals; events can repeat, so several moves or job changes are perfectly normal. The “About your life” step removes questions about what you haven’t had: marriage, children, a job. After the calculation, the chart asks up to 10 follow-up questions about the periods in which an event could have happened and picks the ones that best tell the options apart. If the minute is already found, there are no follow-ups.

No, it’s a hint. Questions about appearance, manner and style carry little weight and don’t unlock a “down to the minute” answer on their own; we don’t promise to guess a sign from appearance. In the “Which is closer to you — A or B?” step you compare two neighboring options: in each row A, B or equal. This narrows the interval, and follow-up questions are asked within it; if no description fits you, choose “None of them is me”. Your choice of descriptions and your answers to follow-ups together reorder the options by at most a factor of two. The time is determined only by the dates of events.

Sensitive topics are a separate step, and only with your consent; without it, such questions appear neither among events nor in follow-ups. The description of a sensitive event isn’t stored: the server zeroes it out, and the interface erases it in the browser first. You can skip the step — the calculation will work on the remaining events, as long as there are at least five.

Yes. Rectification is free — no payment needed. You only need a free Almeo account: without it, the steps with descriptions and follow-up questions don’t work, and the result can’t be saved to your profile. You can run the calculation up to 5 times an hour.