Four Pillars · Calculation method
Know which chart you are reading
A birth record is a clock reading in a place. Before interpreting it, we make the calendar and time choices visible.
Method version: 2026-09. This page describes the calculation, its assumptions, and the limits of the reading. It does not claim that a correctly calculated chart predicts events.
Calculate a chart Read the complete fictional sample
What happens to your birth record
- Establish the recorded moment. We combine the Gregorian date and recorded clock time with the named time zone. Historical daylight-saving rules come from the IANA time-zone database. Do not manually subtract summer time before entering your record.
- Find the year and month. The year changes at the instant of Li Chun. The month changes at each of the twelve month-opening solar terms, not at the first day of a Gregorian or lunar month. We compare the same instant in the relevant time zones.
- Apply your day/hour convention. Use local standard time, or select apparent solar time. Then choose midnight or 23:00 as the beginning of the day. The result states the choices used.
- Derive the relationships. Day Master, Ten Gods and hidden stems follow from the resulting pillars. The reading layer interprets these relationships; it does not change the calendar arithmetic.
Our conventions
| Question | Rule used here |
|---|---|
| When does the year start? | At the exact Li Chun solar-term instant. The popular lunar-new-year zodiac answers a different calendar question. |
| When does a month change? | At its month-opening term: Li Chun, Jing Zhe, Qing Ming, Li Xia, Mang Zhong, Xiao Shu, Li Qiu, Bai Lu, Han Lu, Li Dong, Da Xue and Xiao Han. |
| Which clock is used? | Local standard time by default: the historical daylight-saving adjustment is removed. Apparent solar time is an explicit alternative requiring longitude. |
| When does a day change? | Midnight by default; 23:00 is available. The chosen day stem supplies the hour stem. This is a declared convention, including for the late Zi hour. |
| What if the time is unknown? | No invented noon or hour pillar. A date-only result retains only pillars that can be established under the selected settings. A boundary date may leave more than the hour uncertain. Precise luck-cycle timing and a full reading require a known time. |
| What if only a two-hour interval is known? | It is less precise than a recorded hour and minute. Near a term, time correction or day boundary, the result needs an exact time; the warning is part of the result. |
The calendar engine accepts birth years 1900–2050. Historical time-zone records may be less reliable for older dates and do not establish which clock a family actually used. For a Xinjiang birth, a Beijing-clock record and a Xinjiang-local-clock record are different inputs.
Solar time is more than a longitude adjustment
Apparent solar time combines the recorded moment, longitude, and the equation of time. In minutes, the correction from a recorded clock is:
equation of time + 4 × longitude − 60 × recorded UTC offset
East longitude is positive. The recorded UTC offset includes daylight saving when it applied. We carry a correction across midnight into the date; wrapping the hour alone would lose that change. The equation of time is an astronomical approximation, so a result very close to a boundary should not be treated as exact to the second.
For example, 09:20 on a Beijing clock in Kashgar (75.9°E) is approximately 06:24 in local mean solar time before the equation of time is added. Inferring the civil time zone from longitude would miss about three hours.
A clock time can occur twice when daylight saving ends, or not occur at all when it begins. A repeated hour requires you to identify the first or second occurrence from your record; otherwise the calculator asks for clarification. A nonexistent time is rejected. We do not silently select a moment.
How the stems and luck cycles are derived
The day pillar follows the continuous sixty-day cycle. Month stems follow the Five Tigers rule. Hour stems follow the Five Rats rule in San Ming Tong Hui (三命通會): Jia/Ji days start with Jia-Zi, Yi/Geng with Bing-Zi, Bing/Xin with Wu-Zi, Ding/Ren with Geng-Zi, and Wu/Gui with Ren-Zi.
Luck-cycle direction follows year-stem polarity and the sex input: yang-year males and yin-year females move forward; the other combinations move backward. The start is derived from the interval to the relevant previous or next month-opening solar term, using three days per year of age. The displayed age is a traditional conversion, not a forecast of when an event will occur.
Checks you can inspect and reuse
Our downloadable test inputs and expected checks (JSON) cover ordinary dates and boundaries. They distinguish calendar facts, mathematical checks and convention-dependent outcomes. They are a reproducible test set, not an accuracy percentage or a survey of every calculator.
An independent offline comparison of 60 month-opening terms in 1900, 1950, 1988, 2024 and 2050 compared lunar-python 1.4.8 with Swiss Ephemeris 2.10.03 using its Moshier calculation. The largest difference was 4.727 seconds. This checks selected astronomical instants across the supported range; it does not establish identical results at every possible second. Swiss Ephemeris was a separate research check, not the production BaZi calculation engine.
- Astronomical anchors: the Hong Kong Observatory's February 2024 and August 2024 calendars give solar-term instants independently of our implementation.
- Time conversion: IANA time-zone records and the NOAA solar-time equations describe the historical-clock and astronomical layers separately.
- Classical derivation: the month/hour stem rules are checked against San Ming Tong Hui, including the Wu/Gui → Ren-Zi rule.
- Software comparisons: the lunar-python library supplies solar-term calculations. Same-family libraries are not counted as independent astronomical evidence.
- Test-design reference: Jade Almanac's convention dataset contributed boundary-input ideas. Its recorded outputs describe particular versions and settings; we do not use majority agreement as proof of correctness or endorse its interpretive scoring.
What the reading can and cannot establish
Pillars and time conversion are calculations. Element balance, personality themes and luck-cycle narratives are traditional interpretations. A displayed element count is not a clinical measure, a probability, or proof of which element will help you. Pattern, seasonal, strength-based and other schools can interpret the same chart differently.
The complete fictional sample uses the same report structure as the product. It shows what is delivered without turning a fabricated life story into evidence. Use the questions it raises to reflect on your circumstances, not to outsource a consequential decision.
Correction record
September 2026: replaced date-only solar-term boundaries and approximate year-table reuse; corrected the Gui-day hour-stem mapping; replaced longitude-derived time zones with named historical zones; separated unknown time from noon; and exposed the selected time basis and day boundary. Charts near these boundaries, and hour pillars on Gui days, can differ from earlier results.
To report a discrepancy, include a synthetic example if possible, the time zone and calculation options, and the result you expected. Our editorial and corrections policy explains how we handle evidence and corrections.