One couple, two reports, fifteen points apart

A family checks a match on one website and gets 29.5 out of 36 — comfortably above the 18 that most people treat as the dividing line, and well into the range a report will describe as very good. Somebody checks the same two birth certificates on a second site and gets 14.5. Not 28 instead of 29. Half the score.

The natural conclusion is that one of them is broken. It is almost never true. Both tools can be running the same eight kootas, the same classical tables, the same arithmetic, and arrive at those two numbers honestly. The disagreement comes from a decision made before any of the matching begins — and in most cases neither tool tells you which decision it made.

This guide is about those decisions. We are going to name each one, measure how much it is actually worth in guna points, and rank them. The measurements are ours: we ran 4,000 simulated couples, with births spread across 1990 to 1998 in Indian Standard Time, through the same Ashtakoot scoring engine that produces our own match reports, changing one convention at a time and counting how often the total moved.

The short answer, before the detail: the choice of ayanamsa is worth more than every other cause put together, and the input error everybody worries about — the birth time being a few minutes off — is worth the least.

What the 36 gunas actually read — and what they never look at

Ashtakoot, the eight-fold matching used across North India and most of the software on the internet, reads exactly one quantity from each chart: the sidereal longitude of the Moon at the moment of birth. That single number is a position on a 360-degree circle, and everything else in the score is derived from it.

From that longitude two labels are produced. The rashi is which of the twelve 30-degree signs the Moon fell in. The nakshatra is which of the twenty-seven 13-degree-20-minute lunar mansions it fell in. Four kootas are read from the rashi and four from the nakshatra, and they are not worth the same.

Varna (1 point), Vashya (2), Graha Maitri (5) and Bhakoot (7) come from the rashi — 15 points in all. Tara (3), Yoni (4), Gana (6) and Nadi (8) come from the nakshatra — 21 points, including the two heaviest kootas in the system. So a shift small enough to move the Moon into the neighbouring nakshatra while leaving it in the same sign can still put 21 of the 36 points in play, and a shift that crosses a sign boundary puts the other 15 in play. Our 36-guna guide walks through what each koota is testing.

Two things follow that surprise most people. First, the birthplace does not enter the 36 gunas at all. The Moon’s geocentric longitude is the same number whether the birth was in Chennai or Chicago; only the time zone matters, because it is what converts the local clock time into the universal time the ephemeris needs. If two calculators disagree and you are wondering whether you typed the city wrong, that is not the cause. (The birthplace does matter for Mangal dosha, which is read from houses, and houses need a place.)

Second, nothing else in either chart is consulted. Not the ascendant, not the Sun, not the dashas, not the seventh house, not Rahu or Ketu. Whatever else a report shows you alongside the number, the number itself came from two Moon longitudes and nothing more. That is why our guide on matching without a known birth time can say the missing ascendant costs the Ashtakoot score nothing.

Cause one: the ayanamsa, and why it dwarfs everything else

Astronomical software computes the Moon’s tropical longitude, measured from the equinox. Vedic astrology works sidereally, measured against the fixed stars. The offset between the two is the ayanamsa, and it has to be subtracted before anything else happens.

There is no single agreed value. Several ayanamsas are in serious professional use and they do not agree with each other. Our own engine ships twenty of them and defaults to Lahiri, the official Indian value and the one most Indian panchangs use. Computed for a 1994 date, Lahiri stands at 23.776 degrees, Krishnamurti at 23.679, Yukteshwar at 22.398 and Raman at 22.330. The Lahiri-to-Raman gap is 1.4463 degrees. Our guide to calculating your own dasha table works the same offset through the Vimshottari periods, where it moves the dates rather than the score.

A nakshatra is 13 degrees 20 minutes wide, so 1.4463 degrees is 10.85 per cent of one. Which means that for roughly one birth in nine, Lahiri and Raman disagree about which nakshatra the Moon was in — and in our sample of 8,000 individual births, 11.10 per cent flipped, which is exactly what the geometry predicts.

Now put two of those people together. Across 4,000 couples, switching from Lahiri to Raman changed the 36-guna total for 28.1 per cent of them. When it changed, it changed by 5.68 points on average. And 7.5 per cent of couples — one in thirteen — crossed the 18-point line in one direction or the other, which is to say the ayanamsa alone decided whether the report read as a pass or a failure.

Yukteshwar, 1.3783 degrees from Lahiri, behaves almost identically: 26.7 per cent of totals move. Krishnamurti, only 0.0969 degrees away, moves 1.8 per cent. The size of the gap is the whole story, and it is worth noticing that a tool declaring itself a KP-system calculator will usually agree with a Lahiri one, while a tool built in the Raman lineage often will not.

None of these is wrong. They are different answers to a genuinely open astronomical question — where the sidereal zodiac begins — and a serious practitioner in each lineage can defend theirs. What is not defensible is a calculator that does not say which one it used, because then its number cannot be reconciled with any other number.

The worked couple: 24.4 arcminutes, and 15 points

Take a groom born 12 March 1994 at 09:20 IST and a bride born 19 January 1996 at 06:15 IST. Real arithmetic, computed with the Swiss Ephemeris and scored with our own engine.

On Lahiri, the groom’s Moon is at 326.0525 degrees, in Purva Bhadrapada; the bride’s is at 252.9265 degrees, in Mula. On Raman, the groom’s Moon is at 327.4988 degrees — still Purva Bhadrapada, so nothing about him changes — and the bride’s is at 254.3728 degrees, which is Purva Ashadha, the next nakshatra along.

The boundary between Mula and Purva Ashadha sits at 253 degrees 20 minutes. Under Lahiri the bride’s Moon is 0.4069 degrees short of it. That is 24.4 arcminutes, and since the Moon travels about 13.18 degrees a day, it is roughly 44 minutes of the Moon’s journey. Everything that follows rests on those 44 minutes.

Mula is a rakshasa-gana nakshatra with adi nadi and the dog yoni. Purva Ashadha is manushya-gana with madhya nadi and the monkey yoni. The groom’s Purva Bhadrapada is manushya-gana with adi nadi and the lion yoni. So: on Lahiri, bride and groom share adi nadi, which scores Nadi at 0 out of 8 and is the flag families fear most; a rakshasa bride against a manushya groom scores Gana at 0 out of 6; and the dog-against-lion yoni pairing scores 1 out of 4. On Raman, the nadis differ, so Nadi scores the full 8; both are manushya, so Gana scores the full 6; and monkey-against-lion scores 2.

The other five kootas do not move at all, because they either read the rashi — and the bride stays in Sagittarius on both settings — or, in Tara’s case, happen to land on the same value. Varna 0, Vashya 2, Tara 1.5, Graha Maitri 3, Bhakoot 7 on both.

Add them up: 14.5 out of 36 on Lahiri, 29.5 on Raman. One report says this match falls well short of the conventional threshold and carries a Nadi dosha. The other says it is a strong match with no Nadi problem at all. Both are arithmetically correct.

The same eight kootas, side by side

It is worth seeing the whole scorecard rather than the total, because the shape of the disagreement is as informative as its size. Three rows move and five do not, and the three that move are precisely the three heaviest nakshatra-derived kootas the bride’s shift could reach.

This also explains a complaint we hear often: that one site shows a Nadi dosha and another does not, on the same couple. Nadi is read purely from the nakshatra, it is worth 8 points, and it is scored 0 whenever both partners share the same nadi. A shift of half a degree that moves one partner across a boundary can therefore create or erase a Nadi dosha outright. Our guide to Nadi and Bhakoot covers what the flag means and the classical conditions under which it is held to cancel.

Cause two: whose Moon — Earth’s centre, or the delivery room

This one is almost never mentioned, and it is larger than most people would guess.

The Moon is close. Close enough that its apparent position depends on where on Earth you are standing when you look at it — the same effect that makes a nearby object appear to shift when you close one eye and open the other. An ephemeris can report a geocentric position, computed as if the observer were at the centre of the Earth, or a topocentric one, corrected to the actual latitude, longitude and altitude of the birthplace.

The largest separation between the two in our sample of 8,000 births was 1.02 degrees — about 61 arcminutes, which is wider than the entire gap between the Lahiri and Krishnamurti ayanamsas, and around 70 per cent of the Lahiri-to-Raman gap. It moved the Moon into a different nakshatra for about one birth in twenty-five, and it changed the 36-guna total for 10.8 per cent of couples, pushing 2.9 per cent across the 18-point line.

Classical Jyotish is geocentric, and so is our engine: the chart is cast from the Earth’s centre, which is what the tradition’s own tables assume and what Indian panchangs publish. But a tool built on a general-purpose astronomy library can enable the topocentric correction with a single flag, sometimes without its author thinking of it as an astrological decision at all. If two calculators agree on the ayanamsa and still disagree, this is the next thing to suspect.

Note what makes this different from the ayanamsa: the topocentric correction actually does depend on the birthplace. It is the one route by which a place can reach an Ashtakoot score, and only on a tool that has switched it on.

Cause three: the birth time you were given

The Moon moves about 13.18 degrees a day, which is 0.55 degrees an hour, or a little over half an arcminute a minute. That sets a hard ceiling on how much a wrong birth time can be worth, and the ceiling is lower than people expect.

Measured on the same 4,000 couples, moving one partner’s birth time by five minutes changed the total for 0.5 per cent of couples. Half an hour changed 2.8 per cent. A full hour changed 5.6 per cent, crossing the 18-point line for 1.3 per cent. Compare that with the ayanamsa’s 28.1 per cent and the ranking is clear: unless your time is out by hours, the setting the software chose matters more than the accuracy of the certificate.

That said, the hour-scale errors are real and they cluster in identifiable places. A time recorded to the nearest hour or half hour on an older certificate is exactly the kind of rounding that lands you in the 2.8 to 5.6 per cent band. A birth outside India brings daylight saving into it, and a tool that applies the wrong offset for the date is out by a full hour by definition. India itself has a wrinkle: Indian Standard Time has been UTC plus 5:30 since 1906, but a birth recorded in the older Calcutta or Bombay local time, or during the 1942 to 1945 wartime advance, is not on the offset a modern tool will assume.

The practical version: if your time is confident to within a few minutes, stop worrying about it. If it was rounded, or the birth was abroad, that is worth checking before you accept any single tool’s number as final.

Causes that are not computation at all

Everything above concerns getting the Moon’s longitude. A second family of disagreements happens after both tools already agree on the longitude, and these are editorial rather than astronomical.

Cancellation rules. The classical literature holds that Nadi dosha and Bhakoot dosha stand cancelled under specific conditions — the same rashi with different nakshatras, the same nakshatra with different padas, particular lordship relationships between the two Moon signs. Some tools apply the cancellation and hand you a corrected total; some report the raw score and mention the cancellation in a note; some do not check at all. The same couple can therefore be a 21 with a cancelled Nadi on one report and a 21 with an uncancelled one on another, and the second family will read that as a much worse match.

Half points. Graha Maitri and Tara both produce half-point results in the classical tables — our worked couple scored 1.5 on Tara — and a tool that rounds those to whole numbers will drift from one that does not. If two totals differ by exactly 0.5 or 1, this is usually why, and it is the least consequential disagreement on this page.

Extra tests bolted onto the eight. A report may fold Mangal dosha into its verdict, or add poruthams from the Tamil tradition such as Rajju or Mahendra, or apply a house-based check of its own. None of these changes the 36, but they change the recommendation printed under it — and a family reading the recommendation rather than the number will experience that as the tools disagreeing. Our comparison of Dasa Porutham and the 36 gunas works through what happens when a couple clears one system and is flagged by the other.

What cannot change your guna score, including the thing most often blamed

It is worth being just as precise about the causes that get named and are not causes at all.

The lunar nodes. Rahu and Ketu can be computed as the mean node — the smoothed, always-retrograde point that most Indian panchangs and BPHS-lineage software use, and the one our engine defaults to — or as the true node, the osculating point that briefly turns direct. The two are genuinely far apart: over the ten years from 1990 to 1999 they separated by as much as 1.93 degrees, and they sat in different nakshatras on 7.7 per cent of days. That matters for a Kaal Sarpa reading, for a node’s pada, and for anything that reads where Rahu sits. It cannot touch a guna total, because Ashtakoot reads the Moon and only the Moon. If a report tells you the node setting explains your score difference, it has told you something that is structurally impossible.

The ephemeris behind the software. Whether the positions come from the Swiss Ephemeris, from JPL’s data, or from an analytical approximation, modern sources agree on the Moon to well under an arcsecond over the years anyone is matching horoscopes for. An arcsecond is a three-thousandth of the smallest effect on this page. This is a real distinction for a specialist and it is noise here.

The house system. Whole-sign, Sripati, Placidus and the rest divide the chart differently and change which house a planet is read in. Ashtakoot never asks which house anything is in. The house system reaches Mangal dosha, and stops there.

The gender assignment, mostly. The kootas are scored from the groom to the bride in several places — Varna compares ranks in one direction, and Gana treats a deva groom with a manushya bride differently from the reverse — so a tool that has the roles swapped can differ. It is a small effect and an easy one to check: swap them yourself and see whether the number moves.

Five questions to ask before you trust a score

None of this makes matching unreliable. It makes an undocumented score unreliable, which is a different and much more fixable problem. Five questions settle almost every disagreement.

Which ayanamsa did you use? This is the first question and it resolves most cases. If the tool does not say, its number cannot be reconciled with any other tool’s number, and that is a reason to distrust the tool rather than the tradition.

Is the Moon geocentric or topocentric? Almost no consumer calculator answers this. If one of two tools has the topocentric correction on, you have found your degree.

What exactly is the birth time, and what offset was applied to it? Confident to the minute, or rounded to the hour. Inside India or outside. On the offset that was actually in force on that date.

Which nakshatra did you get for each partner, and how close to a boundary? This is the question that turns an argument into an answer. If either Moon is within about half a degree of a boundary, you now know the score is convention-dependent and you can stop treating the number as a fact about the couple.

Were the cancellation conditions checked, and is this the raw score or the corrected one? The gap between a raw and a cancelled Nadi is 8 points and a great deal of family anxiety.

And then the part no calculator does. A total out of 36 is a compatibility screen, not a verdict on two people, and the tradition itself never treated it as one — the eight kootas were one instrument among several, read alongside the seventh house, the dashas each partner is running, and the judgment of someone who had met the families. Our 16 samskaras guide sets the vivah samskara in that fuller frame. If you want the number for your own two charts with its conventions stated on the face of it, run the match and then bring the koota you are worried about to Jyothshi, who can tell you where each Moon actually sits and how far it is from the boundary that would change everything.