How to read any matching report critically

The score on the front page, in the largest type the document owns, is the least informative thing in it. A number out of 36 records a conclusion and discards the method that produced it. It does not say which tradition was applied, how close the result was, which of the eight components gave way, or whether an equally documented reading would have landed somewhere else.

All of that is either already in the report or one question away from whoever produced it. What follows is five things to establish, ordered so that the cheapest and most decisive come first, and two more to put if the answers keep arriving. Every one of them is a question about method and not about judgement, which is why a careful practitioner answers them without any difficulty at all.

Question one: which tradition produced this verdict?

Establish the rule set before you look at a single number, because every other line on the page means something different depending on the answer. Not the brand of the software — the tradition. The Charak-lineage North Indian Ashtakoota scoring out of 36, the southern reading in B. V. Raman’s lineage, the Kerala ten poruthams, or one of the widely used software presets.

This single fact reorganises the document. Ashtakoota and the ten-porutham system are not two dialects of one method. They weigh different things, one of them has a veto and the other does not, and they can reach opposite conclusions about the same couple without either making a mistake. A score of 22 means one thing on a page where a Rajju failure would have stopped the assessment outright and something else on a page where nothing can.

The scale of the disagreement is measurable, not rhetorical. Run every one of the 11,664 possible ordered pairings of birth star and quarter through five documented readings — the whole population of nakshatra-pada combinations rather than a sample drawn from it, so there is no error bar — and 56.4% trigger a tested difference or veto across the combined census checks, not simply opposite school verdicts.

If the report names no tradition anywhere, that is the first thing to ask for, and it is a reasonable request, not a pointed one. A reading you cannot label is a reading you cannot compare with any other, which means it cannot be confirmed either.

Question two: where does this score sit in the distribution?

A score means very little without the distribution behind it, and no report prints one. Every number in a matching report is presented as though its meaning were self-evident. None of them are.

Measured over every possible pairing, the mean Ashtakoota total is 20.1 out of 36 and the median is 20.5. The conventional acceptance line of 18 is cleared by 66.7% of all pairings, which makes clearing it the ordinary case, not a distinction. Only 6.9% reach 28 or more, and 2.8% reach 30. The top of the scale is genuinely thin; the middle is crowded.

Two things follow, and both of them change how a family reads the same document. A total in the low twenties is unremarkable, not lukewarm: it is where most couples land, including most couples who married decades ago and were never scored at all. And a total of 16 or 17 is not a different species from 18 or 19. It is one lookup table away, which is exactly the distance between two traditions that disagree about a single cell. There is a fuller treatment of what a given score actually means if the number itself is the thing keeping you awake.

Ask where your score sits in the distribution, not whether it crossed a line drawn by convention.

Question three: which components lost points, and how common is that loss?

The koota breakdown is the most useful table in any report and the one most families skim past on the way to the total. Read it as a list of what went wrong, and then ask how often each of those things goes wrong for everybody else.

The rates are nothing like each other. Full marks on varna are earned by 62.5% of all pairings, so losing that single point is both common and cheap. Full marks on yoni are reached by only 7.3%, which means a partial yoni score is close to normal and reads as a defect only to someone who did not know the base rate.

The two components that carry real weight are scored all or nothing, and between them they account for most refusals. Bhakoot is zero for 50% of all pairings — exactly half, because six of the twelve possible rashi distances trigger it. Nadi is zero for 33.3%, roughly one pairing in three, and it takes 8 points with it when it goes. Neither is rare enough to be treated as a sign that something unusual has been found in a particular chart.

That base-rate framing is the whole value of the question. A report says “nadi dosha present” in the register of a diagnosis, and a family hears a diagnosis. The same sentence read against the population is a statement that this couple is in the same position as roughly a third of every possible pairing, which is a different sentence entirely, and it is still perfectly true.

Who else in India is holding a report like this one?

A great many households, and the practice is not shrinking in the way people assume. This is a digression from the checklist and it earns its place, because the isolation is often the worst part of receiving a refusal and the isolation is not warranted by the numbers.

Marriage in India is still overwhelmingly arranged through families rather than assembled privately, and the economics literature on it is unusually good. The most-cited study of mate selection in modern India, working from matrimonial advertisements and the responses to them, found families trading off caste against every other attribute in ways that are stable and measurable, which tells you the screening step is a real institution and not a formality. Chart matching sits inside that screening step for a large share of households, which is why a refusal on astrological grounds arrives with the force of a decision rather than an opinion.

The demographic picture from the most recent National Family Health Survey is of a country where marriage is close to universal and happens young, which means the number of matches being screened every year is very large. Put that beside the base rates in question three — a third of all pairings failing nadi, half failing bhakoot — and the scale becomes clear without anyone needing to guess at it.

None of which makes a particular refusal easier to live with. It does mean the questions on this page are not an unusual thing to ask, and that a practitioner will have been asked them before.

Question four: which houses, and from which reference points?

If the report flags a Mars dosha, two settings decided that flag and a report almost never prints either of them. Ask for both, in writing, because they are the two settings with the widest measured consequence in the entire document.

The first is which houses count. The lists differ between traditions, each following the verse it descends from, and the list most software ships is the union of two of them. The second is the reference point: whether Mars was judged from the ascendant alone, or from the Moon and Venus as well. That second question is the one nobody asks and it is the one that moves the number.

Once you have the two answers written down, the manglik line stops being a fact about a person and becomes a fact about a named reading. That is the change to engineer, because a fact about a person cannot be discussed with anybody, and a fact about a reading can be taken to a second astrologer, to a printed text, or to the other family. Why those settings vary at all, and who chose them, is worked through in the post on what the software decides.

Question five: which cancellations were checked, and what did each require?

Every tradition carries conditions under which a dosha is set aside, and a report that raises a dosha without listing them has done half the work. There are dozens of these conditions across the classical registry — some cited everywhere, some popular in one region and absent from the texts entirely, and several that contradict each other.

The variation in provenance is the thing to hold on to. Some conditions are stated plainly in a printed source you can open: Raman’s Muhurtha sets out nadi cancellation conditions in his own words and not by allusion, which is why they can be implemented without guessing. Others are recited constantly and turn out to have no locatable source at all — the most famous nadi cancellation everyone quotes has not been found in a named printed edition of Muhurta Chintamani or in the Parashara literature. A rule like that is not necessarily wrong. It is unverified, and there is a difference, and a report should say which it is relying on.

The question to ask is not “is there a cancellation”, which invites a yes or a no and tells you nothing. It is “which ones did you check, and what did each one require”. A rule that did not apply is almost as informative as one that did, because it tells you what would have had to be different — and sometimes what would have to be true of a different prospective match, which is a piece of information families routinely want and rarely receive.

The precedence between rules matters too, and it is the part that trips up even careful practitioners. Several cancellations limit each other, so the order they are applied in changes the outcome. A registry that fires rules in one order and a practitioner working from memory in another can reach different results from the same set of conditions, with nobody at fault. The cancellation reference lists the conditions with the text each one comes from, which is the only form in which this is checkable instead of assertable.

Traditions prescribe remedies as well as cancellations, and the two get run together on a lot of reports. They are different categories. A cancellation is a condition under which a tradition holds that the dosha does not apply. A remedy is something a tradition prescribes. A report should mark the second as what a tradition prescribes and not as something with a demonstrated effect, and it is fair to ask which of the two you are being handed.

Two more, once the first five are answered

The ayanamsa, and the birth time to the minute. These come after the first five because they are less often decisive, and they come at all because when they are decisive they are decisive about everything.

The ayanamsa is the offset that converts a tropical longitude into the sidereal one your chart is drawn from, and published values differ by enough to move a Moon across a nakshatra boundary. If your Moon sits well inside its star nothing happens. If it sits near an edge, the birth star itself is a function of a setting, and every rule downstream is reading a different input. That is set out at length in the ayanamsa page.

The birth time question is the cheapest one on this entire list and the one people skip. Ask what time was entered, to the minute, and whether the finding would survive being ten minutes out. For many charts it would, and you can stop worrying about it. For some it would not: in the published sample case the bride’s ascendant moves from Tula to Vrishchika on a shift of 3 minutes, which rotates the entire house numbering and with it any verdict that depends on which house a planet occupies.

Specificity, not confidence

The difference between a practitioner to trust and one to question is almost never the firmness of the verdict, and almost always the granularity underneath it.

A good answer sounds like this. “North Indian Charak lineage, Lahiri ayanamsa, houses 1, 4, 7, 8 and 12 counted from the ascendant only, nadi zero because both Moons fall in the same group, I checked the same-rashi condition and the different-pada condition and neither was met.” That is a sentence you can take to another astrologer, to a text, or to the other family, and every clause in it can be independently confirmed.

A weaker answer is not necessarily a dishonest one. Practitioners who trained by apprenticeship often hold their rules as practice and not as citations, and a rule learned at a teacher’s elbow has no page number attached to it. That is a real limitation and it is not a character flaw. It does mean the reading cannot be checked against anything, which is worth knowing before a family treats it as settled.

What should make you slow down is a report that presents its output as the only arithmetic available. The astronomy underneath every one of these readings is common ground, resolved against the JPL ephemeris to a fraction of an arcsecond; nobody disputes where Mars was. The disagreement lives entirely in the rules applied on top, and a document that hides that is hiding the only part that was ever open to question.

The published sample, run through the five questions

Priya was born in Pune with her Moon in Ashwini; Arjun in Chennai with his in Purva Bhadrapada. Five readings, totals from 13.5 to 17 out of 36, two accepting and three refusing, and the readings that scored the couple lowest are among the ones that accepted them.

That detail is the one people stare at longest. It is not an error.

Question one separates the five documents immediately, because two of them are not scoring out of 36 in the same sense at all. Question two places every one of those totals below the population mean but well inside the crowded middle of the distribution. Question three finds a component paying zero and taking a large block of points with it, which is what produces totals in that range.

Question four is where the documents diverge most visibly: the groom is manglik under three of the five readings and clean under the rest, from identical planetary positions, on the strength of the house list and the reference points alone. Question five explains why two readings accept a couple that three refuse. 19 rules fired across the five documents for this one couple, each citing the text it came from.

Run the five questions on your own report the same way. Take a pen, write the headings down the margin, and fill in whatever the document itself can answer. Whatever is left blank is your list for the person who produced it, and it is usually shorter than you expect.

A refusal you can take apart is not thereby a refusal that was wrong. Establishing that a verdict came from a particular rule set does not establish that the rule set is mistaken, and nothing on this page is an argument that your family’s tradition should be overruled. What the five questions establish is narrower and more useful: whether the refusal is unanimous across the documented readings or particular to one of them, which component produced it, how common that component’s failure is across the whole population, and which settings were chosen when others were available. None of that decides anything for a family, and it is not meant to. A comparison across traditions cannot tell you which tradition your two households should follow, because that is a question about lineage and practice, and it cannot tell you anything about a marriage. It will occasionally produce the least welcome result of all, which is that every documented reading agrees with the one you were given — that happens, it is a real finding, and a comparison incapable of returning it would be worthless.

What this site can supply is the comparison, not the ruling: the same two sets of birth details under all five readings, with the rule that caused each difference named and cited to its source. That is what the matching tool produces. If the more urgent problem is not the report but the conversation waiting at home, the practical version of these questions is in how to ask for the reasoning.