Manglik dosha: the same chart, five different answers

Manglik is not one finding. It is at least five, they are held by different lineages, and on the same charts they disagree more often than they agree. Somebody read a birth chart, said one word, and a conversation about a marriage stopped. The word arrived with no house number attached and no lineage named. That omission is the whole difficulty, because the five readings in common use answer the question differently, and every one of them can be traced to a text.

None of what follows says a verdict was wrong. It says a verdict was under-specified — which is a different complaint, and a fixable one.

What does the word manglik actually claim?

It claims that Mars occupies one of a short list of houses that a particular tradition counts as difficult where marriage is concerned. That is the whole of it. There is no scale, no severity reading built into the word itself, and no statement about a person’s temperament. A chart has twelve houses, Mars sits in one, and a list is consulted.

Because the test is a membership check against a list, it inherits every property of the list. If the list is long, more charts fall inside it. If the list is short, fewer do. And if two lineages hold two different lists — which they do — the same planet produces two different answers with no arithmetic error anywhere.

The name travels by region. In the north it is usually mangal dosha or manglik; in Karnataka and Andhra, kuja dosha; in Tamil Nadu, chevvai dosham, which we treat separately in the Tamil reading of the same test. Four names, one planet, and settings that differ along with the vocabulary.

How far apart do the five readings land on the same charts?

The narrowest documented reading calls 11.7% of charts manglik and the broadest calls 73%, with the charts held completely still between the two runs. The narrow end is Astro-Vision (serious Kuja dosha only). The broad end is OP / SKU panchangas. Nothing moved between those two figures except the rule table.

One chart in nine at one end, three charts in four at the other, and the only variable was which tradition was asked.

The corollary matters more than either endpoint. Taken across the same corpus, 61.4% of charts are manglik under at least one of these traditions and clean under at least one other. That is not a statement about any particular chart, including the one being argued over in a particular family. It is a statement about how often the traditions part company at all: for most charts there exists a citable reading that says manglik and a citable reading that says clean.

How the figure was produced is part of the claim. Each tradition’s own house set and its own reference points were applied to one fixed corpus of 2000 generated charts, and the runs differ from each other in nothing else.

Where did five different house lists come from?

They came from different verses, transmitted faithfully by different lineages, and no surviving text reconciles them. It would be tidier if one list were a transcription slip that propagated into software and could now be corrected. That is not what the sources show.

One verse, attributed to the Brihat Parasara Hora Sastra, enumerates the 1st, 4th, 7th, 8th and 12th houses. It does not mention the 2nd. A second verse, attributed to the Manasagari, enumerates the 2nd, 4th, 7th, 8th and 12th. It does not mention the 1st. B. V. Raman reproduces the second enumeration in Muhurtha, where he attributes it to the Kerala Sastra.

So the northern reading drops the 2nd because its verse drops the 2nd, and the southern reading drops the 1st because its verse drops the 1st. Two texts. Two enumerations. Two lineages that each transmitted their own correctly.

An astrologer working in one lineage and an astrologer working in the other can look at the same chart, both be entirely right within their own tradition, and hand back opposite verdicts. Neither has been careless. They are answering slightly different questions and using one word for both answers.

At the ends of the range sit two further readings, and they are what stretch the spread. One confines the finding to the 7th and the 8th and treats every other placement of Mars as not raising it, reserving the word for the two houses nobody disputes. The other counts seven houses and is the only documented set that includes the 5th, which no other list on the page counts at all. Both are in daily commercial use. Both are cited.

Neither end is an outlier in the sense of being unattested, and neither is a single verse either. They are working lists: a lineage or a publisher settled on the houses it counts, printed them, and passed them on, so the authority comes from long use rather than from one enumeration handed down whole. That is an ordinary way for a rule to reach a practitioner, and it is not a criticism of either reading. It only becomes a difficulty when a working list is presented as though a text had given those houses together, because a family then has no way to tell which of the two kinds of list it has been handed.

What do all five readings agree about?

Four houses appear on every documented list: the 4th, the 7th, the 8th and the 12th. That agreement is the least discussed part of the whole subject, and for a large share of charts it is the only part that matters.

Follow it through. If Mars sits in one of those four houses, every tradition on the page counts it, from the two-house reading at the narrow end to the seven-house reading at the broad end. A second opinion will return the same verdict. A third will too. The finding is stable, and a family arguing about whether some other astrologer might see it differently is arguing about something the lists have already settled.

The entire public dispute is therefore about three houses out of twelve: the 1st, the 2nd and the 5th. The 1st divides the two parent verses in one direction, the 2nd divides them in the other, and the 5th appears in exactly one documented set. Mars has to land in one of those three before the choice of list changes anything at all.

That is a useful thing to know before a conversation starts, because it tells you which kind of conversation you are in. Mars in the 8th and the answer will not move. Mars in the 2nd and the answer sits precisely on the seam between two texts, which is a different situation and calls for a different question.

There is a second layer of agreement further down. The traditions that count the most houses also tend to carry the most cancellation conditions, so the reading that flags the most charts is not always the reading that ends up concluding the most. Counting the flag rate before the cancellation layer, as the census does, measures the widest version of each rule rather than its final one.

Is any of this disagreement about the astronomy?

None of it. Where Mars is at a given moment is not in dispute anywhere on this page. Modern planetary positions come from numerically integrated ephemerides, and the agreement between them is far tighter than anything the rule tables argue about.

The reference series most software ultimately traces back to is JPL’s DE440 and DE441, and positions can be checked against the Horizons service by anyone who wants to verify a longitude rather than take it on trust. For a planet like Mars over the range of dates a birth chart occupies, the residuals are a small fraction of an arcsecond. A house spans thirty degrees.

Put those two scales together and the conclusion is unavoidable. An error of a fraction of an arcsecond cannot move Mars from one house to the next; only a difference in the rule can. So when two reports disagree about whether a chart is manglik, the disagreement has not come from the sky. It has come from the table.

That is worth saying plainly because the opposite belief is common and comforting: that somewhere there is a more accurate calculation which would settle it. There is not. The arithmetic is settled. The doctrine is not, and no improvement in the ephemeris will touch that.

Two things can still move a house number without any disagreement about doctrine, and both are worth naming. The birth time, because the ascendant advances through the zodiac roughly one sign every two hours. And the ayanamsa, the offset between the tropical and sidereal zodiacs, which different traditions set differently. Neither is a dispute about where Mars is; both are disputes about what to measure from.

Is the merged six-house list in any single classical text?

No. The six-house set that behaves as the default in most matching software is the union of the two verses above, and it appears in no single source. Take the northern list, add the 2nd house from the southern list, and you have it: 1, 2, 4, 7, 8, 12. It is a merger, not a transmission.

A union is broader than either of its parents by construction. That is not an opinion about the merger; it is what a union is. Which is why the merged sets sit high in the measured range rather than in the middle of it, and why a chart can be clean under both original verses and still be flagged by the set assembled from them.

The engine here treats the house set as a named parameter with a citation attached rather than a default, and every report it produces cites the text each rule came from. That is not a claim about anyone else’s software. It is a description of what a rule table has to carry before a family can check it.

Worth arguing with: a merged list is a defensible engineering choice if you say you have made it. Being broad is not the failure. Being broad silently is.

Does a longer house list always mean a harsher reading?

No, and the exception is the most interesting thing on the page: one of the short lists carries a condition on Mars that the house numbers alone do not reveal. The verse enumerating 1, 4, 7, 8 and 12 does not stop at the houses. It conditions the finding on Mars being devoid of benefic aspect or conjunction — that is, on nothing helpful sitting with Mars or looking at it.

Read strictly, that verse describes a two-part test. Mars must be in one of the counted houses and must be unaspected by a benefic. A software implementation that reads the houses and drops the clause has widened the rule while appearing to transmit it faithfully.

The consequence runs the opposite way to intuition. A five-house list with a live condition can flag fewer charts than a five-house list with none, and a six-house list applied unconditionally can flag more than a seven-house list applied with care. Counting houses tells you less about strictness than it looks like it does.

The same asymmetry runs through the cancellation layer, where every tradition that raises the finding also documents conditions for setting it aside. Those registries differ from each other too, and we keep them as a cited list rather than a summary, because a verdict that names no cancellation rule has not shown that none applied.

What does a rate this high mean when it is read back onto one chart?

A base rate is a fact about a population, and it changes what a verdict tells you without saying anything at all about any particular chart. If the broadest reading flags roughly three charts in four, then under that reading being flagged is the ordinary case rather than the exceptional one.

That framing is easy to overstate, so here is the careful version. The base rate does not make a finding false. It does not make a tradition wrong. What it does is answer a question families ask constantly and rarely get a number for: how unusual is this? Under the broad readings, not unusual. Under the narrow reading, considerably more so. Both facts are true at once, and which one you are told depends entirely on who did the reading.

The manglik axis is measured separately from the exhaustive star-based census. In that enumeration, 56.4% of pairings trigger at least one tested difference or veto. This combines scoring, cancellation, system and veto checks; it is not just five-school verdict disagreement and does not include the Mars axis.

What would a sceptical relative say to all of this?

The strongest objection is that naming five traditions is just shopping for the answer you wanted, and it deserves a straight response rather than a soft one. Here is the response. Shopping means picking the reading that produces the outcome you prefer and presenting it as the answer. Disclosure means naming which reading produced the outcome and letting the family weigh it. The difference is whether the list gets printed.

There is a test for which one is happening, and it is simple. If a second reading agrees with the first, does the report say so? For a chart with Mars in the 4th, 7th, 8th or 12th, every list on this page counts it, and the honest thing is to say plainly that a second opinion will not move that verdict. A method that only ever finds disagreement is not measuring; it is arguing.

The second objection is sharper and harder. If the traditions disagree this much, perhaps none of it means anything. That is a reasonable position to hold, and it is not one a rate table can settle either way. What we can say is narrower: the disagreement is real, documented, and measurable, and a family entitled to one verdict is entitled to know which of five it received.

What does this measurement refuse to settle?

Three things, and they are stated here rather than buried, because a measurement that oversells itself is worth less than one that does not.

• It does not identify a correct tradition. That question is not answerable by counting, and nothing here attempts it. The traditions are documented, their disagreement is documented, and their relative authority belongs to the people who hold them.
• It does not say what any finding means for a marriage. A tradition concludes something about a chart. What follows from that conclusion is not something a chart, or a census of charts, is able to tell anyone.
• It is a rate over generated charts rather than over a birth cohort, so the endpoints describe the shape of the spread and not a demographic statistic.

What survives all three caveats is the part that matters at a kitchen table: the word on its own does not identify which of at least five findings was made.

Which two facts turn the verdict back into something checkable?

Two, both short enough to ask for in a single message, and neither of them a challenge to anyone’s competence.

• Which house Mars is in. One number between 1 and 12.
• Which houses that reading counts. A list, usually of five, six or seven numbers.

With those two facts in hand the verdict stops being an opinion about a person and becomes a membership test anyone in the room can carry out. Mars in the 2nd is counted by the southern list and not by the northern one, and that sentence can be checked by reading the two lists side by side. Mars in the 7th is counted by all of them.

A third fact is worth having and belongs to a different argument: the point Mars was judged from. The house number is a distance, and a distance needs an origin, which turns out to move the rate roughly ten times as far as the house list does. We take that apart in the attribution between the two axes, and set out the calculation step by step in the anatomy of the two settings.

If you would rather see the lists applied than described, the manglik checker runs all of them side by side and names the source each one came from. It prints the house list next to the verdict, which is the line most reports leave out and the only line that makes a verdict examinable.

One closing thought, offered as an opinion rather than a finding. The five lists have coexisted for centuries without anyone feeling the need to reconcile them, and that is not an accident or a failure of scholarship. They were held in different places by different communities who rarely had to compare notes. What changed is not the doctrine. It is that a family in Pune and a family in Chennai now read each other’s reports, generated by software written by a third party in a fourth tradition, and the seams that were never meant to be load-bearing are suddenly carrying a wedding.