Method, Accuracy and Sources

This page explains where each figure comes from, how large the error is, and which points are genuinely disputed between schools rather than settled. You should finish it able to judge whether these results are worth using.

How solar terms are computed

A solar term is the instant the Sun's apparent ecliptic longitude crosses a multiple of 15°. Those instants come from lunar-javascript, the same open-source library the long-running production calculator on our other site is built on. That is a deliberate choice: this site is only useful if a chart here matches the chart someone was already given, and sharing the library makes that structural instead of something a test has to keep re-proving.

Accuracy, stated honestly

Lichun 2024 comes out at 16:27:07 UTC+8 against a published 16:27 — agreement to the minute the published figure is rounded to. Before this site standardised on one library it ran a second, ephemeris-based implementation alongside; across 144 terms the two agreed to within 41 seconds, and across 24,000 pillar comparisons from 1930–2029 they produced identical charts. So: the seconds shown are real output, not padding, but treat the last few of them as indicative rather than authoritative. A 41-second uncertainty only ever changes a chart for a birth inside that window.

Mixed clocks — which pillar reads which time

The year and month pillars, and the luck cycles, are read in Beijing time, because that is the basis the solar-term table is expressed in; comparing a local clock against Beijing boundaries misplaces anyone born near a Jie. The day and hour pillars are read in true solar time, which is what the classical hour actually means. The Ten Gods are always taken against the true-solar day master, so a chart whose Beijing date and true-solar date differ still reads from the right reference.

The two corrections in true solar time

Clock time to true solar time = longitude correction + equation of time. The longitude correction is the gap between the birthplace and the timezone's central meridian, about 4 minutes per degree, and the meridian is derived from the zone's real historical offset — looked up per birth date, so DST and past timezone changes are handled rather than assumed away. The equation of time uses the same closed-form approximation the production calculator uses (good to roughly ±30 s); a full ephemeris would be tighter, but matching the existing charts matters more than the last half minute.

Where schools disagree, and where this site refuses to choose

The Zi hour: three conventions for 23:00–24:00 differ by a whole day pillar; all three are computed and the active one is labelled. The default matches lunar-javascript so results line up with most Chinese calculators. Element weighting: each visible stem counts 1, and a branch's hidden stems count 1 / 0.5 / 0.3 for the principal, secondary and residual qi. These weights match the existing production calculator so the two agree; they are still one convention among several and should be read as a rough comparison, not a verdict. Luck pillar start: three days to a year, without finer sub-day apportioning.

What is deliberately not offered

This site calculates. It does not judge auspiciousness, forecast fortune, or generate personality readings. The reason is simple: a calculation can be checked and a reading cannot. Almanac do-and-avoid tables and the fortunate/unfortunate ratings of the various star systems disagree between sources and have no verifiable standard, so rather than copy a table we cannot check, we leave it out.

Method and Accuracy