Why Do Saju Calculators Disagree, and How to Tell Which Reading Is Right

2026-07-20 · computed with the solar-term engine (the Method)

If two saju calculators give you different charts from the same birth details, they are not offering two valid opinions. One of them is wrong. A saju chart is a calculation with a single correct answer, so a disagreement means one site made an error, and the error almost always comes from one of three places: how the site handles the solar-term boundary, how it treats a birth in the late-night hour, and whether it corrected for your birth time zone. Once you know the three, you can tell within a minute which chart to trust.

Where saju calculators split

1. The solar-term boundary (the big one)

The month pillar and the year pillar do not change on the dates most people expect. They change at a solar term — an astronomical instant when the sun reaches a specific point along the ecliptic. The year pillar turns at Ipchun, the start of spring in early February, not at January 1 and not at lunar new year. The month pillars turn at the twelve major terms through the year.

These crossings happen at a precise minute, and that minute is different every year and different in every time zone. So if you were born within a day or two of a term, a calculator that rounds to the calendar date can easily give you the wrong month pillar — and, if you are near Ipchun, the wrong year pillar. This single boundary is the most common reason two sites disagree, and it changes the most important character in the chart.

2. The late Rat hour (23:00 to 00:59)

The traditional day does not begin at midnight. The hours are divided into twelve two-hour branches, and the Rat hour straddles midnight, running from 23:00 to 01:00. Sites split on what to do with a birth between 23:00 and 23:59. The mainstream Korean manse convention keeps the day pillar of the birth date and only advances the hour branch — the "late Rat hour" or yaja-si treatment. Some engines instead roll the whole day pillar forward at 23:00. A birth in that one hour can therefore come out with two different day pillars depending on the site's sect, and the day pillar carries your day master, the anchor of the entire reading.

3. Time zone and local solar time

The solar term crosses at an absolute moment in time, so the site has to know your birth time zone to place you correctly relative to it. A birth logged in the wrong zone — or a tool that silently assumes Korea — can land you on the wrong side of a boundary. Stricter engines go further and correct for the longitude of your birth city, since apparent solar time drifts from clock time the farther you are from the zone's meridian. For a birth near a term change or near an hour boundary, that correction can move a pillar.

Why a chatbot is the least reliable calculator

A general chatbot is the worst tool for this and the most tempting, because it writes the interpretation so fluently. The trouble is that it does not compute the solar-term times. It infers the month pillar from the calendar date, which is exactly the rounding error above, and it does it invisibly. The output reads as a confident, coherent chart even when the underlying pillars are wrong. Beautiful prose about the wrong eight characters is still wrong. We wrote a fuller account of that failure in why ChatGPT gets your saju wrong.

How to tell which chart is right

You do not need to be an expert to audit two conflicting charts.

• Check whether your birthday is near a term. If you were born within about two days of a solar term, or in the first days of February near Ipchun, treat any calendar-rounding tool as suspect and trust the one that resolves the term time astronomically.
• Compare the month pillars first. That is where disagreements concentrate. If the month pillars differ, the boundary handling is the culprit, and the correct one is the site that switches at the term moment, not at midnight or the first of the month.
• Check the 23:00–01:00 births. If your birth hour is in that window and two sites disagree on the day pillar, the difference is the yaja-si convention. The mainstream Korean answer keeps the birth-day pillar for a 23:00–23:59 birth.
• Cross-check two independent engines. If two calculators built on different codebases agree on every pillar, that agreement is strong evidence. If they disagree, the one that can explain the boundary rule is the one to trust.

What this site does

K-Saju is built for exactly this problem. The engine computes the solar-term times to the minute the way a Korean manse calendar does, switches the year and month pillars at the term instant rather than at midnight, and follows the mainstream late Rat hour convention for the day pillar. It is not trusted blindly either — the engine is verified against a second, independent library across a sweep of 461 distinct dates and times, including boundary cases sitting right on a term change, and the two agreed on every pillar. The full method is documented on the Method page.

The reason to care is simple. Every word of a saju reading is a statement about a specific chart. If the chart is wrong, the reading is confidently about someone else, and nothing in the fluent paragraph will tell you. Get the eight characters right first. You can cast yours here, free, computed to the solar term, and compare it against whatever else you have been handed.