The science check

We took the numbers the classical astronomy states and measured them against a modern ephemeris. Some hold up to within a second. Some do not. Both results are below, and nothing here is stored — every figure is recomputed each time this page is built.

What measurement can settle, and what it cannot

The hope is obvious: we have instruments the siddhanta authors lacked, so we ought to be able to test their principles and keep whichever voice survives. Half of that works. This is that half, and where it stops.

The astronomical texts state numbers — the length of a month, the length of a year, the rate at which the equinoxes shift. Numbers are checkable, so every figure in the table below is recomputed from the ephemeris each time this page is built. Where a text is refuted, this engine uses the measurement and not the text.

Classical statement	Text	Measured	Off by	Verdict
A lunar month is 29.5305851 days. Tithi, amavasya and every lunar festival depend on this. Being right to a fraction of a second is why the classical calendar stayed usable for a thousand years without correction. The most accurate figure in the text. Lunar periods accumulate over centuries of eclipse records, so a long enough observing tradition can reach this precision with no instruments at all.	29.5305851 days	29.5305892 days	0.35 s	Text confirmed by measurement
The Moon returns to the same star in 27.321674 days. Sets the nakshatra the Moon occupies, so it sets the dasha.	27.3216740 days	27.3216610 days	1.12 s	Text confirmed by measurement
A sidereal year is 365.258756 days. This one is wrong by minutes per year, and the error compounds: over the sixteen centuries since the text it amounts to several days. Makara Sankranti now falls on 14 January rather than at the solstice for exactly this reason. This engine uses the measured year, which is why our Sankranti matches the sky and not the text. Harder to measure than the month: it needs a fixed stellar reference and a full year between observations, so errors do not average out the way lunar ones do.	365.2587560 days	365.2564811 days	196.55 s	Text refuted — measurement used instead
The equinoxes shift 54 arcseconds per year. Roughly seven per cent too fast. Projected back to the text's own epoch it misplaces the ayanamsa by about a degree, which is a nakshatra boundary. No ayanamsa in use today is computed the way the text prescribes.	54.00 arcsec/year	50.28 arcsec/year	3.72 arcsec/year	Text refuted — measurement used instead
No siddhanta has a term for the 18.6-year wobble of the Earth's axis. Not a mistake — it is below what their instruments could resolve. It is also why the true and mean ayanamsa differ, a distinction the tradition could not have made and this engine has to.	no term at all	18.90 arcsec	—	Effect unknown to the texts

The pattern is not random. Lunar periods accumulate over centuries of eclipse records, so a long observing tradition reaches sub-second accuracy with no instruments at all — and it did. The sidereal year needs a fixed stellar reference and a whole year between observations, so its errors do not average out, and that is the figure that went wrong. It is why Makara Sankranti has slid to 14 January instead of sitting at the solstice.

The Chitra criterion, which the tradition itself supplies

The Chitrapaksha rule says Chitra (Spica) sits at exactly 180°. That is a real criterion, so it can be tested, and it ranks the candidate ayanamsas on the tradition’s own terms rather than ours.

Ayanamsa	Misses 180° by	In degrees
True Citra	+0.0″	+0.000°
Lahiri	-62.5″	-0.017°
Krishnamurti	+286.2″	+0.080°
Fagan Bradley	-3242.0″	-0.901°
Yukteshwar	+4899.4″	+1.361°
Raman	+5144.2″	+1.429°
Surya Siddhanta	+10600.9″	+2.945°

Worth noticing what this does to us: Lahiri, which this site uses and which Telugu panchangams use, misses its own criterion by about a minute of arc. Raman and Yukteshwar miss by more than a degree and cannot honestly be called Chitrapaksha. The Surya Siddhanta’s own frame misses by nearly three degrees — the text fails its own tradition’s test.

The one mechanism ever offered, tested

Drishti is carried by planetary rays, so its strength should follow the inverse square of the planet's distance from Earth.

Planet	Nearest (AU)	Farthest (AU)	Received light varies
Mars	0.373	2.676	52×
Venus	0.265	1.736	43×
Mercury	0.549	1.451	7×
Jupiter	3.949	6.456	3×
Saturn	8.032	11.048	2×

Mars varies by a factor of 51.6 in received flux across its cycle. Classical drishti strength is fixed by sign-to-sign angle and contains no distance term, so the rules contradict the mechanism offered to explain them. Either drishti is not radiative, or the rules are wrong; the tradition cannot have both.

Where there is nothing to point an instrument at

“Unfalsifiable” is easy to say and easy to wave away, so each of these names which of the three things a checkable claim needs — a quantity, a mechanism, or an observation it forbids — is missing.

Doctrine	Why measurement cannot reach it
Planets signify things — Mars conflict, Venus pleasure, Saturn delay.	No mechanism and no quantity. The assignment is not derived from anything; it is stipulated, and different texts stipulate differently.
Planets act on each other by aspect (drishti).	No quantity: strength is given as sign-to-sign angle with no units. The one mechanism ever offered is that planets act by their rays, and that mechanism is testable — see drishti_distance, where it fails on the tradition's own rules.
Houses govern domains of life — the seventh marriage, the tenth career.	Houses are directions in the sky. Which domain attaches to which direction is asserted, never argued, and the house systems that divide them disagree with each other.
Vimshottari periods time when a planet's significations arrive.	The 120-year total and the individual period lengths are stipulated numbers. No text derives them, and the competing schemes assign different lengths to the same planets.
Mangal, Kaal Sarp and Pitru doshas damage a marriage or a life.	The harm is never specified precisely enough to look for, and the remedies are held to work whether or not the harm appears. Nothing could count against it.
Thirty-six gunas measure compatibility between two people.	There is no independent scale of compatibility to correlate the score against, and the koota rules themselves differ between regions.
Some hours are auspicious for beginning things and others are not.	The one quantity here is exact — the divisions are arithmetic eighths of daylight — but no mechanism connects an hour to an outcome, and no text states what result would count against the rule.
Gemstones, mantras and rituals alter a chart's indications.	No quantity and no mechanism, and worse: success confirms the remedy while failure is attributed to karma or to incorrect performance, so nothing could count against it. That is not a claim about the world.

The count

5 classical statements are quantitative enough to check. 2 survive measurement, 2 do not, and where a text is refuted this engine follows the measurement. Every one of them is astronomy. The predictive doctrines — all 8 families of them — state no quantity and no mechanism, so no instrument reaches them. Better astronomy has not made them more likely; it has only made them more precisely addressed.

So the honest division is not ancient wisdom against modern science. The tradition’s astronomy is science, always was, and is now partly superseded by better measurement of the same quantities. The tradition’s astrology is a different kind of claim, and no improvement in the astronomy converts one into the other. Computing a chart to the arcsecond does not make the meanings true. It only means they are being assigned to the right degree.