Why do two readers say different things about the same birth date?
Because two separate things can go differently, and the one people blame is usually not at fault. The first is the chart: the eight characters come from the year, month, day and hour of birth, and they vary between places because Korea has moved its clocks more than once and because readers hold different conventions about when a day begins. The second is the reading. Identical eight characters, two readers, two different reports, because schools weigh a chart by different rules and because part of the work is judgement. If saju (사주), the Korean reading of the four pillars of birth, is new to you, the beginner page comes first.
On the chart side, answers split in only five places: the solar terms, the hours after 11 at night, daylight saving time, Korea's old standard time, and longitude. Your birth date and time alone decide which can apply to you.
Why does the saju year start in February and not on 1 January?
The month in a saju chart changes with the solar terms, not with the calendar. The terms are twenty four divisions marking where the sun has actually reached, so the saju year begins at ipchun, the start of spring, around 4 February. If you were born in January or the first days of February, your chart is built on the previous year, and your zodiac animal is the previous year's animal.
The same thing happens every month. A term arrives once a month, usually between the 4th and the 8th, so a birth in those days can fall either side of a month boundary, and the boundary is fixed to the minute rather than to the date. If ipchun one year falls at 5:12 in the morning on 4 February, a person born at 4 o'clock and a person born at 6 o'clock that same morning have different month pillars, same calendar day, same city. Places that treat a term as a whole day part company from those that work out the moment, which I have computed to the minute.
Why does a birth after 11 at night give two different day pillars?
Because two conventions about when the day turns over are in live use, and they disagree by half an hour. This affects only births after 11 at night.
| Convention | Hours it covers | Date changes at |
|---|---|---|
| Half hour ja (반시), standard in Korean almanacs | 00:00 to 01:29 | 11:30 at night |
| Full two hour ja | 23:00 to 00:59 | 11:00 at night |
The date changing means this: born at 11:30 at night, your day pillar is built on the following day, not the day on your record. A whole pillar moves, two of the eight characters change, and everything read from them changes. Mulsangjae uses the half hour convention for a practical reason, not a doctrinal one: it is what most Korean readers and almanacs use, so your chart here has the best chance of matching one you were shown elsewhere. If you want both, say so and I will draw the other alongside.
Did Korea really have daylight saving time?
It did, in three separate stretches, and this is the point missed most often. In those periods the clock ran an hour ahead of real time, so the hour on a birth record is an hour later than the sun was.
| Years | Period in force |
|---|---|
| 1948 to 1951 | Spring to early autumn, each year |
| 1955 to 1960 | May to September, each year |
| 1987 | 10 May, 2am, to 11 October, 3am |
| 1988 | 8 May, 2am, to 9 October, 3am |
The Seoul Olympics years catch many people, so a summer birth in 1987 or 1988 is worth checking first. These periods also began and ended at 2am and 3am rather than at midnight, so a birth just before a listed start or end date is easy to get backwards from the dates alone. I work from the transition times, not the dates.
Near midnight the effect is bigger than an hour, because the date moves. Someone born at 12:01 in the morning on a July day in 1987 was really born at 11:01 the previous evening, so the day pillar sits on the day before the birthday and two of the eight characters change completely. When a chart has come out differently elsewhere, this is the first thing I check.
Why does Korean clock time need a thirty minute adjustment?
Korea sets its clocks by the meridian at 135 degrees east while the country sits nearer 127.5 degrees, so the clock runs about thirty minutes ahead of the sun. Saju is read from the sun, so those thirty minutes are put back before the hour pillar is drawn, and a place that skips the step gives a different hour pillar near an hour boundary.
There is an exception, a pleasing one. From 1954 to 1961 Korea kept its clocks on 127.5 degrees, thirty minutes behind today's standard, so a birth then needs thirty minutes added to match the present clock. Since 127.5 degrees is very nearly the true longitude of the peninsula, in those years the clock and the sun agreed and the correction above was not needed.
The east to west spread inside Korea is narrow, so birthplace shifts things only slightly. Gangneung on the east coast and Mokpo in the south west are about ten minutes apart in solar time, which is nothing unless a birth time sits on an hour boundary, where those ten minutes decide it. That is why I ask where you were born as well as when. If the birth time itself is uncertain, that has its own page.
So which one is right?
In most cases neither is wrong. Two of the five are conventions, and a convention is a choice rather than an error: whether the day turns at 11:00 or at 11:30, and whether to apply the longitude correction. Careful readers differ on both, and a chart drawn either way is defensible, provided the reader knows which convention they work in and says so.
The other three are not conventions. Daylight saving time was really in force. The old standard time was really in use. The moment a solar term arrives is a fact about the sky, not something anyone decides. A chart that ignores those three is working from the wrong eight characters, and no amount of skill further down the page repairs it.
Mulsangjae applies all three. Where a correction has moved your date I say so in the opening lines of the report. Stated plainly it is a small technical note; left unsaid it reads as though I had your birthday wrong.
Why do readers still disagree when the eight characters match?
The remaining differences are differences of reading, and they are real. The largest is which single element the chart is judged to need. One way weighs the chart and corrects the imbalance, draining what is in excess and supplying what is thin. The other looks at climate first: a chart born in deep winter with no fire is warmed before anything else is weighed, one born in high summer with fire everywhere is cooled first. Both are legitimate and often agree. Where they name different elements, the advice in the two reports points different ways. The five elements page goes through it slowly.
The second difference is emphasis. A chart holds more true statements than fit in one report, so the reader decides which lead: the ten year cycles, with the birth chart as background; or the day master outward; or sinsal (신살), the symbolic stars, which some use heavily and others hardly touch. Side by side those look like contradictions when they answer different questions about one chart.
The third is method. Free apps mostly read the day master alone, which is why the output could belong to anyone: one character out of eight doing all the work. Chinese bazi weighs the same eight characters differently and names them in another vocabulary. My own work builds the eight into a single scene of natural images and reads what the things in it do to each other.
Then the part that is uncomfortable to write down and still true. Some of it is the reader's judgement. Where two rules point opposite ways, a person decides which carries more weight here, and another person with the same training may decide otherwise. I would rather say so than pretend the method removes the human being from the work. What I offer instead is narrow and checkable: what I read, which characters I read it from, and where the chart was a close call.
How do I tell which of these applies to me?
Run down this list with your birth date and time in front of you.
- Born 1 January to roughly 8 February: look at the solar terms.
- Born on the 4th to the 9th of any month: look at the solar terms.
- Born after 11 at night: look at the ja hour convention.
- Born in the summer of 1987 or 1988: daylight saving time.
- Born 1954 to 1961: the old standard time.
If none of the five applies and the chart still differs, the calculation is not the problem. What went into it is: usually a solar date entered as a lunar one or the reverse, or a lunar leap month dropped somewhere. Send the date as you have it and I will work out which.
What this page cannot tell you
This is general explanation. It tells you where charts diverge and why, but it says nothing about your own chart. Whether a correction moves your day pillar depends on your exact date, time and place, and which element your chart needs depends on all eight characters together and the cycles passing over them.
The limits of the practice are worth stating. Saju describes tendencies and timing, not outcomes, and a correctly drawn chart is the floor of a good reading rather than proof of one. To see what I do with a chart once drawn, the sample report shows the shape of the whole thing.