The Hour Pillar: Why It Is the Most Common Source of Wrong Charts
Ask a practitioner which pillar they trust least, and most will say the hour. Not because the method is weak, but because four separate things can corrupt the input before the calculation even starts.
1. The birth time is a recollection
Most people do not have an exact birth time. They have "around two in the afternoon" or "after dinner," repeated by family for decades. A one-hour uncertainty can move the hour pillar by a full branch.
Practical approach: if the time is uncertain, calculate two charts — one at each plausible boundary — and compare. If the hour pillar is the same, the uncertainty does not matter. If it differs, note it as an open question rather than picking the flattering one.
2. The time is in clock time, not solar time
Two-hour branches are defined in true solar time, not clock time. The gap between them depends on birthplace:
- Longitude. Four minutes per degree from your time zone's central meridian. Beijing is 14 minutes off; Taiyuan is 30; Urumqi is nearly two hours off because China uses one zone across its width.
- Equation of time. Clock vs sundial, roughly ±16 minutes across the year.
Combined, these regularly move a birth time by 30–90 minutes. If the uncorrected time sits near a branch boundary, the hour pillar flips.
Urumqi is the extreme case worth knowing: born at 08:00 clock time in winter, the true solar time is about 06:00 — a different branch, a different hour stem, a different part of the reading.
3. Daylight saving is invisible in the record
China used daylight saving from 1986 to 1991. A birth certificate from August 1990 records clock time at UTC+9, but the chart needs UTC+8 solar time. Applying the standard offset without accounting for DST introduces a full hour of error — enough to move most births across a branch boundary.
Many countries still observe DST, and the rules change over time. This is why our calculator uses the IANA time zone database rather than a fixed offset — it knows what the offset was at that specific date and place.
4. The late-night boundary dispute
The hours 23:00–24:00 sit in an awkward place. That window is 子 hour, but does the day pillar advance at 23:00 or at midnight?
Schools differ:
- Advance at 23:00 — the day pillar rolls to the next day, and the hour is the next day's 子 hour.
- Advance at midnight — the day pillar stays on the current day, while the hour stem is calculated as the following day's 子 hour.
Both are used. The key is to know which convention your calculator applies and to be consistent. Our calculator follows the second convention (no day advance at 23:00), matching the classical Python reference implementation the engine was verified against.
If you were born between 23:00 and 00:00, this is worth checking explicitly — it changes the day pillar, and the day pillar contains your Day Master, which everything else is read against.
How to check yours
- Write down the clock time as recorded, and the city of birth.
- Check whether the calculator asked for the city. If it did not, the hour pillar is uncorrected.
- Look for a stated correction in minutes on the result page. If there is none, the time was taken at face value.
- Run a second chart one hour earlier and one hour later. If the hour pillar changes, your reading depends on the accuracy of that time — treat it as provisional.
The hour pillar covers the least-documented part of the chart, and it is the part most affected by data quality. Being honest about that uncertainty is more useful than pretending the eighth character is as solid as the first.