Chart

Ten in the morning is not the same hour everywhere

A birth chart is built from the moment you were born. A clock does not measure that moment — it measures an administrative agreement about which moment to call what. Two corrections recover the difference, and one of them can move an entire pillar.

When someone casts a BaZi chart, the first input is usually a clock reading. That is the only thing most people have, and it is the wrong unit.

Civil time is organised around time zones, and a time zone is centred on a meridian. China, for instance, runs the whole country on UTC+8, whose meridian is 120° east. If you are born in Beijing — about 116.4° east — you are 3.6° west of the meridian your clock is keeping. The sun does not care. It is simply not at the position your clock claims.

The two corrections

Longitude correction. The earth turns 15° per hour, so one degree of longitude is four minutes of solar time. Being 3.6° west of the meridian puts true solar noon 14.4 minutes later than the clock says.

Equation of time. The sun is not a perfect timekeeper. Earth's orbit is elliptical and its axis is tilted, so apparent solar time drifts against mean solar time by up to about ±16 minutes over the year. On 3 November it runs roughly 16.4 minutes ahead.

Add the two and you get the true solar time at the place and moment of birth. Everything downstream — which two-hour branch the hour falls into, and therefore the whole hour pillar — is derived from that number, not from the clock.

An example, from the live engine

Same civil clock reading, same date, same minute — three different birthplaces. These are real outputs from the chart endpoint, not illustrations:

3 November 1990, 10:00 civil time
BirthplaceLongitude offsetEquation of timeNetTrue solar timeHour pillar
Beijing−14.4 min+16.4 min+2.0 min10:02乙巳
Berlin−6.4 min+16.4 min+10.0 min10:10乙巳
Ürümqi−129.5 min+16.4 min−113.1 min08:07甲辰

Source: the studio's /api/paipan endpoint, queried 23 October 2026.

The first two rows show the correction being small and unremarkable: a couple of minutes, no change to the pillar. The third row is the one worth looking at. Ürümqi is in the same time zone as Beijing — the same clock on the wall says the same thing — but it sits 29° further west. The correction is nearly two hours in the other direction, and the hour pillar moves from 巳 to 辰.

Two people born at the same clock minute in the same country, with same-day birth dates, do not get the same chart. This is not a subtlety for specialists. It is a two-hour error in the input.

Near a boundary, two minutes is enough

The Beijing row looks harmless because its true solar time is 10:02 — nowhere near a pillar boundary. Move the clock reading two minutes and that stops being true.

Beijing, 3 November 1990 — same day, four minutes apart
ClockTrue solar timeHour pillar
10:5710:59乙巳
10:5811:00丙午

Two minutes of correction is nothing until your birth minute sits on the edge of a two-hour block. Then it is the difference between one pillar and another — and the hour pillar is a quarter of the chart.

This cuts both ways, and it is worth saying plainly: if your recorded birth time is itself uncertain by ten minutes, no amount of solar-time precision will save the chart. A correction applied to a wrong input is still a wrong chart. The correction removes one class of error — the systematic one — and leaves the other one alone.

What the studio actually does

You give a date, a minute and a birthplace. The birthplace supplies the longitude and the time zone; the engine computes the longitude offset and the equation of time, converts to true solar time, and casts the four pillars from that. The page shows the arithmetic — the offset in minutes, the solar time, the method — so you can check it rather than take it on faith.

You can cast a chart yourself, free, on the chart page. It returns the four pillars, the day master, the strength of the chart and the element it is short of, in about a second.

Put in a date, a minute and a birthplace. The engine does the rest and shows its work.

Cast a chart →