Why Sukuyo calculators differ: a diagnostic checklist
Why Sukuyo calculators differ: check the original input, lunar conversion, 27/28-mansion table, and displayed field in that order.
Overview
Find the first differing intermediate field. A useful issue report contains more than two final mansion names.
At a glance
- Compare first
- Original input
- Compare second
- Converted lunar date
- Compare third
- Roster and month-start table
Same numbers, different calendar?
Check whether each field expects Gregorian or lunar dates. Also confirm the year and whether a leap-month marker was entered.
An apparent contradiction may be two tools answering different input questions.
Do the lunar dates match?
Compare month, day and leap status. If they already differ, examine the new-moon civil date, timezone and month-assignment policy.
Do not debug the relationship table before establishing the same birth mansion input.
Do the tables match?
Compare the 27-versus-28 roster and the month-start entry. The T1299 first-day row and OpenFate modern row are explicitly different.
Counting day one as the starting position also matters: an off-by-one rule shifts every result.
Are they showing the same field?
A page can show a birth mansion, today’s mansion or an astronomical Moon position. Read the field label and date.
Different field meanings are not a failed calculation.
A useful issue report
Provide the tool URLs, exact inputs, settings, converted lunar dates and displayed mansion names; omit another person’s private details unless authorized.
A comparison with those intermediate fields is reproducible. A claim that one description feels more accurate is not a calculation diagnosis.
Build a first-difference table
Create one row each for input calendar, civil date, timezone, lunar month/day/leap flag, roster size, month-start entry, offset rule and output type. Place calculator A and B in adjacent columns.
Everything after the first unequal row is a consequence until shown otherwise; this avoids blaming the final label for an earlier method change.
Sources and editorial basis
- 宿曜經 · T1299 卷上卷上「大唐月建圖」first-day row 虛室奎胃畢參鬼星翼角氐心 compared with the product row.Historical text for mansion names and calendar tables. Its table is not identical to OpenFate’s modern 27-mansion lookup; it does not establish predictive validity.
- OpenFate Sukuyo calculator and calculation conventionEditorial code check 2026-09-10, policy 1.0.0: calculator-differences; policy.ts, constants.ts, engine.ts and traditional-cycles.ts. Product settings are not an independent historical authority; practical examples are OpenFate editorial applications.Product-specific input and output reference. Editorial checks also inspected policy 1.0.0 and the month-start constants on 2026-09-10; those implementation checks are not an independent traditional authority.