At a glance
- Start here
- Record both sets of inputs first
Record both sets of inputs first
Record the calendar, full date, local clock reading, birthplace, time zone and leap-month status. A month number without its calendar is insufficient. Preserve a foreign birth’s local clock time and zone rather than manually converting it and allowing the software to convert it again. Mark uncertain minutes as a range. A demo value is not a birth record.
Use the changed pillar to narrow the check
If the year pillar differs, check the year boundary. For the month pillar, check the exact solar-term transition rather than assuming every month starts on lunar day one. For a day-pillar difference, inspect the date, zone conversion and midnight-versus-23:00 convention. For the hour pillar, check hour boundaries and solar-time options. This narrows investigation; it does not prove a single cause.
| Difference | First check |
|---|---|
| Year pillar | Year boundary and transition instant |
| Month pillar | Solar terms, not only lunar month number |
| Day pillar | Date, zone and day rollover |
| Hour pillar | Hour boundary and solar time |
Solar time is not a fixed minute correction
Relative to the UTC+8 reference meridian at 120°E, the longitude term at 105°E is minus 60 minutes. Apparent solar time also includes the date-dependent equation of time. This example isolates longitude; it is not a complete apparent-time result for a particular day. Keep the original clock reading and corrections separate, and do not treat a corrected solar clock as a new UTC instant.
Compare day and hour conventions together
At 23:30, midnight and 23:00 rollover conventions can use different day stems. Because the hour stem depends on the chosen day stem, do not splice a day pillar from one result into an hour pillar from another. Save one complete result, change only the boundary setting and recompute. If the uncertain birth interval crosses that boundary, preserve both candidates.
Change one setting at a time and keep the comparison
Create a baseline in the BaZi tool and save its inputs and rules. On each comparison change one option and record the changed pillars. If matching conventions still disagree, compare the calendar data, historical time-zone database and calculation version. A minimal reproducible input is more useful than personal life stories or an identifiable chart screenshot.
Common questions
Does a difference prove one chart is wrong?
First match inputs and conventions. Different conventions can produce different results; disagreement under the same declared convention needs further investigation.
Can a few minutes change a chart?
It can if the interval crosses an hour, day or solar-term boundary. Do not adjust minutes simply to obtain a preferred result.
Try the related tools
Keep the inputs, results and notes together so you can compare them later.
Browse more practical questionsReferences & further reading
Historical sources document ideas and rules of their time, not empirical proof of predictive accuracy. Source notes identify the nature and scope of each reference.
- Hong Kong Observatory: CalendarInstitutional / reference source
Calendar and lunisolar definitions; used for date concepts only.
- ZenTell: BaZi calculation conventionsInstitutional / reference source
This site’s declared conventions; the diagnostic workflow is editorial guidance, not a universal school rule.
- Hong Kong Observatory: Apparent Solar TimeInstitutional / reference source
Official astronomical definition of apparent solar time.
- US Naval Observatory: The Equation of TimeInstitutional / reference source
Official astronomical reference; explains apparent minus mean solar time.