Why Chinese Calendars Disagree: Dates, Time Zones, Sources

Two Chinese calendars, opened on the same phone on the same afternoon, can disagree. Sometimes the disagreement is a single day in the lunar date; sometimes the dates match and the lists of favoured and discouraged activities do not. Neither situation necessarily means one of the calendars is broken. Chinese calendar data is produced by converting astronomical events into calendar days, and that conversion involves choices: a reference time zone, a set of tables, a compilation tradition for the almanac columns. Each choice can shift the output. This page sets out the four main causes of disagreement, with a worked example, and finishes with a procedure for deciding which entry to use.

The time zone is the usual culprit

A lunar month begins on the day of the new moon. The new moon is an instant: the moment the moon passes between the earth and the sun. That instant is the same everywhere on the planet, but the calendar day it belongs to is not, because a day is defined locally.

Chinese calendars are normally built on Beijing time, UTC+8. A calendar built for Vietnam, Thailand, Malaysia or Indonesia may use the local zone, which is UTC+7 or UTC+8 depending on the country. The difference is one hour, and one hour is enough. If the new moon falls within an hour of local midnight, the two zones place it on different calendar days, and the first day of the lunar month moves by one.

This is not a rounding error or a bug. It is the direct consequence of asking "which day was the new moon?" in two places where the day starts at different moments. The same effect applies to solar terms, which are also instants: a term that falls late in the evening in Beijing can fall after midnight in a zone one hour ahead, and then the printed date differs.

The practical consequence is that a one-day discrepancy between two calendars is most likely to appear in months where the new moon falls near midnight. It will not appear consistently; it will appear in some months and not others, which is exactly the pattern people describe when they say a calendar is "sometimes off".

Different data versions and different rule sets

The second cause is the data behind the calendar rather than the arithmetic in front of it. Astronomical positions are computed from tables, and those tables have been revised repeatedly over the past century. A calendar generated from an older table can place a new moon or a solar term a few minutes away from where a modern table places it, and a few minutes is enough to move a date when the event sits near midnight.

For the recent period the differences are small and the mainstream tables agree on almost every day. The further back or forward you go from the present, the more likely a discrepancy becomes, which is why historical date conversions and far-future projections often disagree by a day.

There is also a rule-set question that affects the leap month. The standard rule places a leap month in a lunar month that contains no major solar term. Applying that rule requires both the new moon times and the solar term times, so any difference in either input can change which month is doubled. For the modern period this is rare, but it is not impossible, and it is one reason two calendars can disagree about the number of months in a year.

Why the almanac columns disagree even when the dates match

The yi and ji lists, the clash, the day deity and the direction of the clash are not computed from the sky. They are looked up in tables compiled from traditional source texts, and different publishers compile from different texts, add local custom, or resolve conflicting entries differently.

The result is that two calendars can agree perfectly on the lunar date and the day's stems and branches while disagreeing on whether the day favours a wedding. This is normal. The date is an astronomical output; the activity list is a compilation.

The clash and the day deity usually agree, because they follow mechanically from the day's branch. The yi and ji lists are where variation concentrates, often by one or two activities out of a dozen. How those columns are structured, and what each one claims, is covered in the guide to the almanac's activity columns.

Different year boundaries in different contexts

A third source of confusion is not a disagreement between calendars at all, but two boundaries in use within the same tradition. New Year's Day, the first day of the first lunar month, is the boundary used for the festival and for ordinary conversation. Lichun, the solar term around 4 February, is the boundary used for the year pillar in a birth chart.

In most years the two are days apart, and in 2027 they are two days apart: Lichun on 4 February, New Year's Day on 6 February. A person born in that gap has one animal year by festival reckoning and another by chart. Two sources that appear to disagree about that person's sign may simply be using different boundaries.

The same issue affects the question of when the year changes for naming purposes. A publication that says the Goat year begins on 4 February is not contradicting one that says 6 February; they are answering different questions. The festival side is set out in Chinese New Year 2027, and the chart side follows the solar terms.

A worked example: neighbouring countries, one day apart

The clearest illustration is historical rather than hypothetical. In 2007 the Vietnamese new year fell on 17 February, while the Chinese new year fell on 18 February. The two calendars used the same rule and the same kind of data; the difference came from the time zone used to place the new moon on a calendar day. Vietnam used its own local time, China used Beijing time, and the new moon fell close enough to midnight that the two zones assigned it to different days.

Nothing about either calendar was wrong. Each was internally consistent and correct for its own reference frame. The episode is useful because it shows the size of the effect: exactly one day, in one year, for one boundary, caused entirely by a one-hour offset.

The same mechanism explains why a lunar birthday looked up on a calendar built for one country can differ by a day from the same birthday looked up on a calendar built for another. The conversion method is described in converting a lunar birthday.

How to decide which entry to use

Start by fixing the frame. Decide which reference the date is for: a festival observed in a particular country, a birth chart, or a personal record. A festival observed in Vietnam should be read on a calendar built for Vietnam; a birth chart should be read on the reference the practitioner uses, normally Beijing time.

Then check the source's stated time zone. A calendar that does not say which zone it uses cannot be reconciled with one that does, and comparing them will produce an apparent error that is really a missing piece of information.

Then compare only like with like. Check the lunar date and the day's stems and branches first; if those agree, any remaining difference is in the compiled activity lists and is a matter of tradition rather than calculation.

Finally, keep one source for the whole cycle. Mixing a conversion from one calendar with an annual check on another is the most common way families end up with two dates for one festival. The date is easy to verify: open the daily Chinese calendar, select the day, and read the lunar date and the term printed with it.

A one-day difference between two calendars usually means the two were built for different places or from different tables. It does not mean the tradition is unreliable, and it does not mean one of the sources is careless.

What it does mean is that "the Chinese date" is not a single fact independent of context. It is a fact about a place and a reference frame, and once the frame is stated the disagreement usually disappears. When it does not disappear, the cause is almost always the compiled almanac content rather than the calendar itself, and that is a difference of tradition, which no amount of checking will resolve into a single answer.

For most practical purposes the procedure is short: pick a source that states its zone, use it consistently, verify the date on the page rather than from memory, and treat differences in the activity lists as what they are — different compilations of the same tradition. When two entries still disagree after that, check the day on a calendar that states its reference and compare the lunar date and the day pillar; if those match, the calendars are describing the same day and only the commentary differs.