Methodology
The rules we calculate with
Saju schools disagree with each other. Rather than hide which side we took, we publish it, so you can tell whether a chart elsewhere differs because of a mistake or because of a different convention.
What we adopted
| When the year changes | the year changes at Ipchun (입춘), the first solar term, not on 1 January |
|---|---|
| When the month changes | the month changes at each of the twelve major solar terms (절), not on a calendar date |
| Which time we read the chart from | local mean solar time, corrected for how far your birthplace is from its time zone meridian |
| When the day changes | the day changes at 23:00, when the Zi hour begins |
| Southern hemisphere | charts in the southern hemisphere use the same solar terms as the north |
These are settings, not code. When one changes, the version number on your result changes with it, so an old reading can always be traced to the rules it was produced under.
The sixty-day cycle
Day pillars are counted from 1984-01-31, taken from 남두성 음양 만세력 / 한국어 위키백과 갑자일 목록 475건 교차검증 (2026-09-14).
How the sun is calculated
Solar terms are found by solving for the moment the sun reaches a given apparent longitude, using a published astronomical method (Meeus, Astronomical Algorithms, 2nd ed., ch. 25 (low accuracy) / ch. 28 (equation of time)).
That method states an accuracy of about 0.01 degrees of apparent longitude. The sun moves roughly one degree a day, so the instant of a solar term is not exact. We say so rather than presenting it as exact, and we intend to replace this method with a higher precision one before launch.
The difference between universal time and terrestrial time is applied before the sun's position is evaluated, worth 94 seconds around 2050.
Measured against published solar term times: largest difference 13 minutes. The accuracy page shows the full comparison.
meeus-low-precision 1.0.0.
Where schools disagree, and what we chose
Two of the rules above are genuinely contested. We state our choice rather than implying there is only one answer.
The day boundary. We turn the whole day at 23:00, when the Zi hour begins. This is the standard across Hong Kong, Taiwan and much of South East Asia, and it is also the Korean orthodox position, which declines to split the Zi hour into an evening and a morning half. The competing position turns the day at midnight.
The time standard. We correct for longitude but not for the equation of time. The half-hour correction Korean practitioners apply is a longitude correction, and Korean calculators apply it by default. Adding the equation of time is astronomically purer but puts us up to sixteen minutes away from the charts people are used to.
If you were born between 23:00 and 01:00, your free result shows what the other conventions would give, side by side with ours. That is the one window where the choice changes your chart.
How we judge the balance of a chart
Almost every reading rests on whether the chart backs the Day Master or draws on it. The classical method asks three questions, and Korean and English-language sources ask the same three: does the month branch back it, does the branch directly under it back it, and does enough of the rest of the chart back it.
We answer all three and publish the count. We do not publish a single score, and that is not caution on our part — the sources say the same thing. Schools weigh the month branch, hidden stems, combinations and seasonal temperature differently, so a number implies a precision the method does not have.
Two limits worth stating. We count roots at the main qi of each branch only, because hidden stems are switched off in this build, which makes our count more conservative than a reading that includes them. And we do not read combinations, clashes or luck pillars into the balance.
Leaning one way is not better than the other. This measures support and load inside a chart. It does not measure confidence, resilience, health, intelligence, or the worth of a life, and we do not use it to tell you what work to do or what to invest in.
Time zones
The offset used is the one that was actually in force at your birthplace on your birth date, taken from the standard time zone database, not today's offset. Where a birth moment falls in an hour that was skipped or repeated by a daylight saving change, we ask you which one you mean instead of guessing.