How we calculate
Every step below changes what your chart says. Where practitioners disagree we name what we chose, what we did not choose, and why — and where we have not settled a question, we say that instead of picking quietly.
Of the four calculators we surveyed, none documented where they place the boundary of a day. That is why this page exists.
The solar terms
Each solar term is the instant the sun's apparent ecliptic longitude crosses another multiple of 15°. We compute all of them with an astronomical library and ship the result as a fixed table rather than computing at read time, so the same birth always resolves the same way.
We ship 3,864 of them, covering 1900–2060. Births outside that range are refused at the form rather than estimated.
We compared 696 of them against the national observatory's published figures. The median difference on the twelve terms that open a month is 0.29 minutes.
Three deviations beyond that are tracked individually rather than absorbed into a tolerance. Only one of them can move a Month Pillar, and only for births inside a six-hour window on a single day in 2011. We would rather the list of known deviations be visible and grow than have it quietly averaged away.
The observatory's own published data is authoritative but not infallible — we found an error in it. A check that automatically treats the published figure as correct would draw the wrong conclusion, so ours does not.
Conventions we apply
- Clock correction
For births in Korea we move the recorded clock back by 32 minutes before reading any pillar from it. Korea keeps standard time on the 135° meridian while Seoul sits near 127°, so the clock runs ahead of the sun by about half an hour.
We doApplied to Korean births only, as a fixed −32 minutes from the 135° meridian.
Others doSome calculators apply a longitude correction everywhere, and some apply none at all. Applying it everywhere would move every non-Korean chart we produce; applying none would put Korean hour pillars half an hour away from the reference almanacs.
WhyWe matched a reference Korean almanac, which applies it. A birth at 12:50 in Seoul is read at 12:18.
- Equation of time
The sun runs up to about 16 minutes ahead of or behind clock time across the year, independently of longitude.
We doNot applied.
Others doSome practitioners add it on top of the longitude correction, which shifts the hour pillar for births within about a quarter hour of an hour boundary.
WhyThe reference we matched does not apply it. Adding it would move us away from that reference.
- Where the day begins
A birth between 23:00 and midnight belongs either to the day that is ending or to the one about to begin, and the two answers give different Day Pillars.
We doMidnight (正子時). The day turns at 00:00.
Others do23:00 (子時) — the day turns when the Rat hour opens. We compute this on request; it is recorded in the chart's methodology either way.
WhyMidnight matches the reference. We also report which half of the Rat hour a birth falls in — 夜子時 before midnight, 朝子時 after — because that distinction is read even under the midnight convention.
- We do not resolve time zonesa property of the standard, not a bug
We read the clock as it was recorded where it was recorded. We do not convert it to any other zone before reading pillars from it.
We doNo timezone conversion. Daylight saving, where it was in force, is removed first.
Others doA calculator that converted everything to one reference zone would give different Year and Month Pillars for some births.
WhyThis is a defined property of our standard, not a bug: two people born at the same absolute instant in different countries can receive different Year and Month Pillars from us. If you are comparing our output against a calculator that converts, this is where the two will part.
- How the first decade age is rounded
The age your first decade period starts at comes from counting days to the nearest solar term and dividing by three. The remainder has to go somewhere.
We doRemainder 0 or 1 rounds down; remainder 2 rounds up. Ages are Korean counting age, 1 at birth.
Others doSources record schools that always round up, and schools that always round down.
WhyTwo independent sources agree on the rule we use, and it reproduces the reference almanac's start ages across all four direction cases.
- Twelve Life Stages for Yin stems
The twelve-stage cycle runs forward for Yang stems. Whether it also runs forward for Yin stems is the single most-argued point in this table.
We doYin stems run in reverse (陽順陰逆).
Others doSome modern schools run Yin stems forward as well, which changes strength judgements downstream.
WhyThe reference almanac's output matched reverse across all ten stems — four natal pillars plus ten decade pillars, fourteen positions, no exceptions.
Where the schools disagree
These are the Symbolic Stars whose derivation is genuinely contested in our sources. Some of them you can switch; the switch is recorded in your chart's method, so two charts made under different settings stay distinguishable rather than overwriting one another.
- Kui Gang 魁罡switchable
We useThe classic four: 庚辰 · 庚戌 · 壬辰 · 戊戌.
Others useA modern six-pillar list adds 戊辰 and 壬戌.
Not settledWe cannot tell the two apart from the reference we have — it produced 庚辰, which is on both lists. Settling it needs a 戊辰 or 壬戌 day pillar.
- Gwimun Gwansal 鬼門關殺switchableKorean practice
We useAny pair among all eight characters.
Others useAdjacent pillars only — year–month, month–day, day–hour.
Not settledThe reference charts we checked contained no qualifying pair, so it did not decide this. The pair list itself is not in doubt; only how far to look.
- Red Charm 紅艷switchable
We useDay Pillar only, on the base table.
Others useA variant table adds 申 for the 甲 · 乙 · 庚 · 壬 Day Masters, and some sources look at all four branches.
Not settledOne source, which itself says the basis differs between sources. We do not produce this star beyond the Day Pillar.
- Yang Blade 羊刃
We useYang Day Masters only.
Others useSome schools include Yin stems as well, under the name 陰刃 / 飛刃.
Not settledOpen. We took the narrower reading rather than produce a star we cannot source.
- White Tiger 白虎
We useDay Pillar only.
Others useSome sources look at any pillar.
Not settledOpen. Note that 戊辰 and 壬戌 appear both here and on the modern Kui Gang list, so the two questions are entangled.
- Heavenly Nobleman 天乙貴人
We useDerived from the Day Master.
Others useSome traditions derive it from the Year Stem as well.
Not settledOpen. We produce the Day Master derivation only.
Questions we closed
These were live disagreements. We settled them against a reference almanac rather than by argument, and the checks are repeatable.
- Peach Blossom, Travelling Horse and Canopy Star are read from the Year Branch
These three are part of the twelve-star cycle, and the cycle's basis is the Year Branch. Testing the alternative — the Day Branch — against a reference chart failed on eight of eight positions; the Year Branch matched on eight of eight.
- The Day Pillar comes from the national observatory, not from our own arithmetic
The sixty-day cycle runs unbroken and is independent of any school, so it can be looked up rather than derived. We take it from the Korea Astronomy and Space Science Institute's published calendar data. Its year and month fields are not usable for this purpose — they follow the civil calendar rather than the solar terms — and we do not use them.