Methodology

計The two layers

A reading here is built in two separate steps, and they are kept strictly apart. The first is a deterministic calculation: the eight characters, the hidden stems, the Ten Gods, the interactions and the ten-year cycles are computed from astronomy and calendar rules. The second is interpretation, which reads that structured result and writes about it.

The interpretation layer is not allowed to recalculate anything. Before any paid report is stored, it is checked against the chart it claims to describe — the Day Master, the day pillar, the strength result, the absent elements and the luck-cycle years must all match, or the report is not issued.

節Solar terms decide the month, not the calendar

Saju months do not begin on the first of a calendar month. They begin at the twelve major solar terms — the instants when the sun reaches 315°, 345°, 15° and so on. The Saju year turns at Ipchun (立春) in early February, not on 1 January.

We compute those instants from the sun’s apparent geocentric longitude rather than reading them from a table, so any year from 1900 onward is handled the same way. If your birth falls within a day of a term change, the report says so, because at that distance a birth time recorded to the nearest hour can move the month pillar.

時Historical time zones and daylight saving

The clock on the wall when you were born was not necessarily the clock your birthplace uses today. Korea ran on UTC+08:30 between 1954 and 1961 and observed daylight saving in 1987 and 1988. The United States changed its daylight-saving rules in 1974, 1975, 1987 and 2007.

Rather than maintain a private table that would go stale, we read the IANA time-zone database through the runtime, which carries the full historical record. If your birth time falls in an hour that was skipped or repeated by a clock change, the report tells you which interpretation was used.

陽Three different clocks, and why the difference matters

“Solar time” is not one quantity. Three are in circulation, they disagree by up to half an hour, and services routinely blur them. We separate them explicitly.

Worked example, and the reason this page exists. A birth recorded at 11:30 PM on 9 November 1984 in Seoul (126.98°E, KST = UTC+9, no daylight saving):

• The clock in force represents the 135°E meridian. (126.98 − 135) × 4 minutes per degree = a longitude correction of −32.1 minutes, giving 10:57 PM in local mean solar time.
• The equation of time on that date is about +16.2 minutes — it peaks near +16m24s on 3 November. Adding it gives 11:14 PM in local apparent solar time.
• 10:57 PM falls in the 亥 hour and leaves the day pillar alone. 11:14 PM falls in the 子 hour and, under Yajasi, advances the day. The Day Master comes out as 丁 under one method and 戊 under the other.

The two corrections can point in opposite directions, so applying only the first is not a milder version of applying both — it can land on the far side of a boundary. That is why we apply the full apparent-solar conversion by default: the service says apparent solar time, so it computes apparent solar time. Every method that would give you a different chart is published alongside it with its exact pillars.

Korean practice is genuinely not uniform here. The roughly 30-minute longitude correction is near-universal in Korean perpetual calendars; whether the equation of time is added varies by calculator and by school, and there is Korean academic work devoted specifically to how apparent-solar correction should be handled. We take a position, label it precisely, and show you the alternatives rather than presenting ours as the only answer.

Order of operations is fixed and matters: the instant is resolved first, then the time conversion is applied, and only then are the hour branch and the day boundary read from the converted time. Reading a branch before converting would silently reintroduce the error this whole section exists to remove.

夜The Yajasi midnight convention

The 子 hour runs from 23:00 to 01:00, which means it straddles midnight, and Korean schools do not agree on which day it belongs to. This service uses the Yajasi convention: a birth at or after 11:00 PM in the applied local time takes the following day’s pillar. A second living tradition changes the day at midnight, so an 11:30 PM birth keeps the earlier day. The day boundary is a separate setting from the time method — we compute all six combinations of the two and publish every one that gives a different chart.

This is not a small difference. It changes the Day Master — the character that represents you — and therefore changes the whole reading. Whenever a competing convention would move your day or hour pillar, your report lists it in full rather than settling the question silently.

曆The lunar calendar

If you know your birth date on the Korean lunar calendar, we convert it using the classical rules rather than a lookup table: months run from new moon to new moon, the month containing the winter solstice is month 11, and in a thirteen-month year the leap month is the first one that contains no mid-term. Instants are evaluated on Korea’s civil clock, which is the meridian the Korean lunar calendar is published against.

論Where the traditions disagree

Some things in Saju are calculation and some are judgement. We try never to present the second as the first.

• Whether the equation of time belongs in the correction is unsettled in Korean practice. The longitude correction is near-universal; the equation of time is applied by some calculators and not others. We apply it, say so, and show the chart you would get without it.
• Whether a Day Master is strong or weak is an interpretive call. Schools weigh the month branch, the roots in the branches and the hidden stems differently. We show our support and drain numbers so you can see how the conclusion was reached.
• The useful god (용신) is the most contested topic in the whole tradition. We do not name one. Instead we describe which elements support and which stretch your Day Master under the strength reading above, and we say plainly that this is our stance rather than a settled fact.
• Hidden-stem weightings, the treatment of partial harmony frames, and how much a clash matters all vary by school. Where a structure is partial we label it partial.

源Sources and knowledge policy

Our interpretation framework draws from classical Myeongri principles, Korean Saju terminology, traditional calculation methods and major Korean interpretive approaches. We do not copy sentences from other services or from copyrighted modern books; the rules are structured independently and the English is written for this product.

We do not claim to be a complete or final system, and we do not claim any accuracy rate. Saju is an interpretive tradition, not a measurement.

信How We Build Trust

A transparent Korean Saju system built for a global audience.

• Korean cultural authenticity, with the Korean terms kept and explained
• Deterministic chart calculation, separated from interpretation
• Global birth cities with historical time zones and daylight saving
• Calculation conventions published, including the ones that disagree
• A structured interpretation framework rather than one generic prompt
• Clear modern English
• Identity and timing, not personality alone
• Individual and relationship analysis
• No fear-based selling
• No subscription at launch

What the last validation actually checked:

• Sexagenary day anchors cross-checked at 1949-10-01, 2000-01-01 and 2024-01-01.
• Solar-term instants compared against published KASI values for 立春 2023–2025 and 立冬/大雪 1984.
• Equation of time checked against its annual extremes and its four zero crossings.
• Historical time zones exercised for Korea 1959 and 1987 and for United States daylight saving in 1975 and 1990.
• Lunar conversion checked against Korean New Year 1984 and 2024 and Chuseok 2024.

據Sources and how they may be used

Every source is listed with its copyright position, because a claim to transparency that skips the legal question is not one. Only public-domain classics, lawfully licensed material, public-institution data and rules written for this product are used. Where a modern work disagrees with us we say so; we do not store or reproduce its text.

異Recorded disagreements

These are the points where competent practitioners differ. The ones marked as changing the chart are the reason your report lists alternative charts.

版Versioning

Every paid report records the versions that produced it, so a reading can always be explained rather than merely re-run.

Have a chart you believe we calculated wrongly? The FAQ explains what to send, and a calculation error is always a refund.