Why different sites produce different four pillars
Most disagreements are not arithmetic errors but different conventions: birthplace and time zone, true solar time, the zi-hour day change, solar-term boundaries, hidden-stem tables and the start of the Major Cycles. Checked in order, most differences can be located.
Direct answer
Different sites often disagree because they use different conventions, not because one of them miscalculated: birthplace and time zone, true solar time, the zi-hour day change, solar-term boundaries, hidden-stem tables and the start of the Major Cycles. Check them in that order and most differences can be located.
Start by accepting that there is no single official four pillars
Enter the same birthday into two sites and get two different sets of four pillars, and the first reaction is usually 'one of them got it wrong'. In most cases neither did. They chose different conventions. The four pillars are fixed by calendar and stem-branch rules, and at several key points those rules allow more than one accepted practice.
The differences fall into roughly five layers. The input layer: solar or lunar date, whether a leap month is marked, how precisely the birthplace is given, and whether the recorded time is standard time or daylight saving. The time layer: whether true solar time is on, how longitude and the equation of time are handled, and which school's rule governs the day change at the zi hour. The solar-term layer: the year turns at the Start of Spring and the month at a sectional term, but which time zone and data precision are used for the instant? The structural layer: which hidden-stem table is used, and how strength and the Ten Gods are calculated. The cycle layer: how the starting age and the boundaries of the Major Cycles are fixed.
Each layer has a different reach. True solar time usually changes only the hour pillar, though near the zi hour it can involve the day pillar. The zi-hour rule changes the day pillar directly. A solar-term boundary changes the year or month pillar. The hidden-stem version does not change the four pillars at all, but it changes the Ten Gods, strength and useful god. The cycle method does not change the four pillars either; it changes when the Major Cycles begin. Confusing 'different four pillars' with 'different readings' leaves the check without a direction.
The correct order runs from input to time, then to solar terms and structure, and only then to cycles. Fix the input first, compare one layer at a time, and at every difference ask which convention that site chose. This is far more useful than arguing about which site is more accurate.
Comparison table: seven sources of four-pillar differences
| Source of difference | Which pillar is affected | Common symptom | How to check |
|---|---|---|---|
| Birthplace and time zone | Mainly the hour pillar | The same clock time falls into different branches at different longitudes | Check the birthplace, coordinates and time zone used |
| True solar time switch | Hour pillar; near the zi hour possibly the day pillar | One site gives a jia-zi hour, another a yi-chou hour | See whether true solar time is stated, then cast once with it on and once off |
| The zi-hour day change | Day pillar; possibly the hour stem as well | For a birth after 23:00, the day pillar differs by one day and the hour stem may differ | Confirm whether the day changes at midnight or at 23:00, and which day the hour stem follows |
| Solar-term instant and boundary | Year pillar, month pillar | For a birth on the day of Start of Spring or a sectional term, the year or month pillar differs | Check the exact term instant and the time zone used |
| Calendar and leap-month input | The whole year, month and day set shifts | A solar date entered as lunar, or a missing leap-month marker | Check the input type and the leap-month flag |
| Hidden-stem and strength versions | Does not change the four pillars | The same four pillars, different Ten Gods and useful-god conclusions | Compare the hidden-stem table and the method for selecting the useful god |
| Cycle-start method | Does not change the four pillars | Different starting age and different boundaries for the Major Cycles | Compare the cycle-start calculation and boundary rule |
The order of the table is the order of checking: input first, then time, then solar terms, and only then structure and cycles. Skipping steps makes a convention difference look like an arithmetic error.
How to verify: seven steps to locate the difference
Step one: fix the input to one set.
Align the date, time, gender, birthplace, solar or lunar calendar type and leap-month marker before comparing. Many 'different four pillars' are actually different inputs: one site interpreted your solar date as lunar, or a leap-month flag was dropped. If the input is not unified, every later comparison is invalid.
Step two: confirm birthplace and time zone.
The birthplace determines longitude and also the time zone. 'A province' and 'a city or county' can produce different precision, and a birth abroad requires confirming whether the recorded time is local or Beijing time. Check whether both sites matched the same coordinates, then whether their time-zone handling agrees.
Step three: cast once with true solar time on and once off.
Keep everything else fixed, change only the true solar time setting, and compare the four pillars. If the hour pillar moves, the birth moment is close to an hour boundary or the longitude difference is large. If nothing changes, this is not the source. Note that some tools correct only for longitude and omit the equation of time, so their offset may differ slightly.
Step four: check the zi-hour day change.
For a birth between 23:00 and 01:00, confirm which day-change convention each site uses. Changing at midnight and changing at 23:00 produce day pillars one day apart, and once the day pillar changes, the Ten Gods and Major Cycles follow. Look only at the day pillar at this step; do not be distracted by the hour pillar.
Step five: check the solar-term boundary.
For a birthday near Start of Spring or another sectional term, confirm the exact instant of the term and which time zone the site used. The year turns at Start of Spring and the month at the sectional term, so a birth a few minutes before or after the instant can produce a different year or month pillar. This difference usually affects only the year or month pillar.
Step six: if the four pillars match but readings differ, check the structural layer.
When the four pillars are identical, the difference lies in the hidden-stem table, strength scoring, Ten Gods derivation or the method for selecting the useful god. A different hidden-stem version can change the Ten Gods configuration; a different strength algorithm can change the useful-god conclusion. This is an interpretive-layer difference and should no longer be called a charting error.
Step seven: check the cycle start last.
The starting age and the boundaries of the Major Cycles belong to the cycle layer and do not change the four pillars. If the pillars agree but the starting age differs, the issue is the cycle-start calculation, not the charting. Leaving this layer until last prevents a cycle difference from being mistaken for a pillar difference.
How we approach it
Our conventions are not the only possible answer, but we try to put the choices in plain view. When a result differs from another site, you can compare the mechanisms item by item:
Input and birthplace are fixed first. Charting needs the solar or lunar calendar type, birth time, gender and birthplace. Once a city is matched, its coordinates are brought in and true solar time is calculated from them. If the input layer is unclear, a more precise result has no comparative meaning.
True solar time is a visible switch. True solar time can be turned on or off in the birth information, and the default follows the matched city. Whether it was enabled is recorded in the chart result. When it changes the hour branch, the notice shows both branches and the offset in minutes.
The zi hour changes the day at midnight. We use the convention that the day changes at midnight: from 23:00 to 24:00 the day pillar remains that of the current day, while the hour stem follows the next day's day stem, and after midnight the chart moves into the next day. The other convention changes the day at 23:00, producing a day pillar one day earlier; the notice states that a school difference exists here.
Solar terms change the pillars. The year pillar is bounded by Start of Spring and the month pillar by the twelve sectional terms, not by the first day of a lunar month or 1 January. When a birth is close to a term instant, the notice marks the nearest term and the approximate distance, warning that the year or month pillar may be sensitive.
Structure and cycles are explained separately. Hidden stems, strength, Ten Gods and cycle start each have their own convention. When the four pillars match but conclusions differ, the difference belongs to the structural or cycle layer. We do not describe that as another site 'getting the chart wrong', and we do not pretend our convention is the only standard.
Limits and disclaimer
Different sites producing different four pillars does not mean one of them is wrong, and it does not mean the two results can be mixed freely. There is no globally unified official answer for charting conventions; what can be compared is the input, the convention and its reach. BaZi readings do not provide medical diagnoses, disaster claims, investment advice or legal conclusions. For major decisions, rely on real-world conditions and qualified advice.
Frequently asked questions
Two sites give day pillars one day apart. Which one is wrong?
First check whether the birth time falls between 23:00 and 01:00. If it does, the difference most likely comes from choosing the midnight or the 23:00 convention for changing the day. If not, check the solar/lunar input, the leap-month marker and the time zone. A different day pillar alone does not establish that either site is wrong.
For a birth on the day a solar term changes, whose year pillar is right?
The year pillar changes at the exact instant of Start of Spring, not at the start of the day. Before the instant it belongs to the previous year; after it, the next. A birth on that day must be timed precisely and the time zone confirmed. If the two sites give different term instants, the year pillar can differ.
Can solar versus lunar input cause a difference?
Yes, and it is the most common input-layer error. Entering a solar date as lunar, or a lunar date as solar, shifts the entire set of four pillars. A missing leap-month marker can shift the month as well. Before comparing, confirm that both sides interpret the input under the same calendar.
How many hidden-stem versions are there, and does it matter?
Different versions of the hidden-stem table exist, mainly differing in how many hidden stems a branch carries and which ones are retained. They do not change the four pillars, but they can change the Ten Gods configuration, strength scoring and useful-god judgement. When the pillars match and the conclusions differ, the hidden-stem version is worth checking.
Is a different Major Cycle start also a convention issue?
Yes. The starting age, cycle boundaries and forward or backward direction all have more than one accepted method, but none of them changes the four pillars. If the pillars agree and the starting age differs, the issue is at the cycle layer, not a charting error.
Is there a standard answer I can look up?
There is no globally unified official four pillars. Calendars and solar terms have precise astronomical definitions, but true solar time, the zi-hour day change, hidden stems and cycle start all involve school choices. The reliable approach is not to find a standard website but to confirm which convention a source uses and remain consistent.
If the charts do not match, can I still trust the reading?
Settle the chart before discussing the reading. When the four pillars differ, every conclusion lacks a shared starting point. Once you locate the layer of difference, either continue with one convention or compare the two side by side. Arguing about accuracy before the charts are aligned only makes the discussion muddier.
Where can I see this site's conventions?
The chart result states whether true solar time is enabled. When a birth time sits on an hour boundary, crosses a branch under true solar time, or falls near a zi-hour or solar-term boundary, the chart notice explains the difference and our choice. The knowledge base and methodology page also explain these conventions.
Related comparisons
Updated 2026-10-20