How the calculations work

Where the positions come from

Planetary positions are computed from VSOP87 for the planets and ELP2000 for the Moon — the standard analytical theories, accurate to about an arcsecond over the range this platform accepts. That is roughly a thousand times finer than any interpretive rule can distinguish, so the ephemeris is not where error enters. The transformations on top of it are, and those are what we test.

The ayanamsha is a direction, not a formula

Rather than fitting a polynomial, this platform pins one fixed direction in space: the point that had a tropical longitude of 23°51'11.5" at the start of the year 2000. The ayanamsha at any other moment is simply where that same direction has precessed to.

This has a property a fitted curve does not: the small wobble in the Earth's axis called nutation appears in both the planet's position and the ayanamsha, so it cancels exactly when one is subtracted from the other. Sidereal longitudes come out clean without anyone having to remember to remove it.

Checked against the Indian Calendar Reform Committee's defining value: the committee fixed the Lahiri ayanamsha at exactly 23°15'00" for 21 March 1956, and this model returns 23°15'01".

The ascendant is solved, not looked up

The ascendant is the point of the ecliptic sitting on the eastern horizon at your birth moment. That is a geometry problem with an exact answer, and the platform solves it and then checks its own answer by putting it back into the horizon equation — the altitude must come out zero, and the point must be on the rising side rather than the setting one. That check runs over hundreds of combinations of latitude and sidereal time in the test suite, so a sign error cannot ship.

Time zones, including the awkward ones

An hour of time-zone error produces a chart that looks entirely plausible and is entirely wrong, which makes it the most dangerous bug in this kind of software. Offsets come from the IANA database, so historical rules apply as they actually were — India's UTC+06:30 from 1942 to 1945 included.

Daylight-saving edge cases are detected rather than papered over. If your birth time falls in the hour that a clock change skipped, or in the hour it repeated, the chart says so and states which reading it used. Which one a birth certificate meant is genuinely unknowable, so the choice is stated rather than hidden.

What we check the numbers against

The strongest tests are the ones that compare against something outside the software. Mesha Sankranti — the Sun's entry into sidereal Aries — has to land between 13 and 15 April; if the ayanamsha were applied with the wrong sign, or not at all, it would land on 21 March instead. The Moon's illumination has to agree with the tithi, which is the same geometry computed two independent ways. Every mahadasha's sub-periods have to sum to the mahadasha itself, at every level.

Where we stop

Only the Lahiri ayanamsha has been checked against an independent published table; the other three are offered and flagged as unverified. The Yoni compatibility matrix is a documented simplification of the full classical table, and its default is the neutral middle rather than a flattering one. The planetary strength score is a transparent composite and is deliberately not called Shadbala, which is a different and more elaborate classical computation.

And the largest limit of all: none of this makes astrology predictive. The astronomy is exact; what anyone reads into it is a tradition, and this platform describes that tradition rather than vouching for it.