"Partial Kaal Sarp" is not a classical category

Kaal sarp is a single geometric question with exactly two possible answers, so there is no state of affairs that “partial” could be describing. Somebody has looked at one of the two charts and said those words. Perhaps it was framed as the lesser version: not the full affliction, only some of it, so the wedding is not refused outright but a rite is required first. That framing is the thing to examine, because the underlying test does not have a lesser version. It has an answer, and the answer is yes or no.

What does the kaal sarp test actually ask?

It asks whether all seven grahas fall inside one of the two halves the Rahu–Ketu axis cuts the zodiac into. That is the whole question. Everything else printed under the heading is commentary on it.

The computation runs in four steps, and none of them is hidden. First, compute the sidereal longitude of each of the seven: Sun, Moon, Mars, Mercury, Jupiter, Venus, Saturn. Second, compute the longitude of Rahu; Ketu needs no separate calculation, because it sits opposite by construction. Third, take the arc that runs forward from Rahu to Ketu and ask, for each graha in turn, whether its longitude falls inside that arc. Fourth, look at the two tallies. Seven and zero in either direction is the condition. Six and one is not. Five and two is not.

Notice what is absent. There is no weighting, no koota, no threshold, no total to compare against a pass mark. The eight kootas of Ashtakoota produce a number out of thirty-six and a reader can reasonably argue about where the line should sit. This produces a boolean. You can run it yourself against a printed chart with a pencil and a list of longitudes, which is more than can be said for most of what arrives on a matching report. The kaal sarp checker prints the sorted list and marks which side each graha is on, because the list is the argument.

Rahu and Ketu are not in the sorted set. The seven are the classical grahas, and the nodes are the ruler being measured against, not two more things to measure. Counting them would make the test circular — they are always exactly on the boundary. Uranus, Neptune and Pluto are outside the classical scheme entirely and no formulation of this test includes them.

What are Rahu and Ketu, if they are not planets?

The Moon’s orbital path is tilted a few degrees against the plane of the Earth’s, so it spends half of each month above that plane and half below. It has to cross twice. Those two crossings are Rahu and Ketu: two points, not two bodies, exactly opposite one another because a plane cuts a circle in exactly two places. And they travel backwards.

This is not obscure or disputed astronomy. It is the same geometry that governs when eclipses happen, which is why the nodes are documented in detail by people who have nothing to do with this subject. NASA’s account of eclipses and the saros describes the node shifting steadily westward as the reason an eclipse series eventually ends. A solar eclipse can only occur when a new Moon happens near a node. The word Rahu attaches to the ascending one, Ketu to the descending one, and the eclipse association is the origin of the imagery about a serpent swallowing the light.

Two consequences follow, and both matter for the test. The first is that the axis is a line, not two independent objects. Anything that moves Rahu moves Ketu by exactly the same amount in the opposite direction, so the two halves stay two halves. The second is that the nodes drift backwards through the zodiac, completing a circuit in a little under nineteen years. The axis is not fixed scenery; it is slowly sweeping past the grahas.

There is a real technical fork here that reports almost never disclose. The node can be computed as a smoothed average or as its instantaneous osculating position — the mean node or the true node. The true node wobbles either side of the mean, and the gap between them is comfortably large enough to move a graha from one side of the axis to the other. Two calculators that disagree about kaal sarp very often disagree about nothing else at all: one took the mean node, the other took the true one, and neither said so. Our planetary positions tool states which it uses.

A binary verdict, reinterpreted as a gradient

When a report says partial kaal sarp, it usually means something like this: six of the seven are hemmed in, and only one has escaped. Stated plainly, that is a description of the condition being absent. One graha outside the arc is precisely the case the test was written to exclude.

A test with one bit of output cannot return a fraction of that bit.

The move being made is a familiar one, and it is not unique to this category. A binary verdict that most charts fail is commercially inert: there is nothing to prescribe for the great majority who simply do not have it. Reinterpret the same test as a gradient and every chart lands somewhere on the gradient. Six of seven becomes “partial”, five of seven becomes “mild”, and a condition that was uncommon becomes a condition that is nearly universal. Nothing about the chart changed. Only the reading did.

This one deserves an argument, though, because there is a version of the objection that earns an answer. A defender might say: surely six grahas crowded into one half is astrologically different from three and four, whatever the label. That is a reasonable thing to believe. But if it is what someone believes, then the honest move is to define a new measure — largest empty arc, say, in degrees — and report that number, with its own scale and its own sources. What is not honest is to borrow the authority of a named classical category and then apply it to a chart that fails its stated test. The word dosha made the same journey, from something weighed against other factors to a flag that is either raised or not, and the history of the word dosha traces how.

Twelve serpent names for twelve houses

Anant, Kulika, Vasuki, Shankhapal, Padma, Mahapadma, Takshak, Karkotak, Shankhachur, Ghatak, Vishdhar and Sheshnag. Twelve serpents for twelve houses. Which variety a chart is said to carry is settled by which house from the ascendant Rahu occupies, with Ketu in the house directly opposite, so the naming scheme is a lookup on a single number between one and twelve. It is not a severity scale, it is not ordered from mild to grave, and a chart cannot be a little bit Takshak. Reading the list as a ranking is the same category error as reading “partial” as a degree, arriving by a different door.

A second distinction does exist in practice and is more defensible: whether the grahas lie in the arc running from Rahu towards Ketu or the arc running from Ketu towards Rahu. Some practitioners treat those as materially different arrangements and give them separate readings. Whatever one makes of the doctrine, that distinction at least tracks something real about the geometry, which the mild-to-severe gradient does not. The angarak and grahan yoga checker handles the neighbouring combinations that get reported in the same breath as this one and are frequently confused with it.

How common is the arrangement, and why will we not publish a rate for it?

We publish base rates when we can compute them over a whole population, and the kaal sarp condition is not one of those. The koota rates on this site come from scoring every one of the possible nakshatra-pada pairings, an exhaustive count over a closed set of positions. Kaal sarp needs seven real planetary longitudes at a real moment, so any figure for it would be a figure about whichever cohort of birth moments was fed in. Quoting a percentage from a sample of dates and presenting it as the rate would be exactly the sort of undisclosed choice this site exists to complain about.

What can be said without measuring anything is structural, and it is more useful than a headline number. The seven grahas are not seven independent tokens scattered at random. Mercury never appears far from the Sun, and Venus not much further; both are inner planets, so from Earth they stay tethered to it. Three of the seven therefore travel as a loose bundle. Jupiter and Saturn are slow enough that they sit wherever they sat last year. The effective number of independently placed bodies is smaller than seven, and an all-on-one-side arrangement is correspondingly less rare than a naive count would suggest.

The Moon is the opposite case, and it is the one a family should understand. It travels through the whole zodiac in under a month, which means it crosses the Rahu–Ketu axis roughly every two weeks. In many charts the Moon is the only graha on the far side, and it is the graha whose side is most sensitive to the recorded birth time. A birth recorded to the day, not the minute, is usually fine here, but a chart in which the Moon sits within a degree of the axis is a chart whose kaal sarp verdict is a coin balanced on its edge. Whether that is your situation is knowable, and a report ought to tell you when a finding sits that close to a boundary instead of printing the verdict alone.

The positions are not the loose part

Measured against NASA’s JPL planetary and lunar ephemerides, across two centuries of dates, the worst disagreement in any planetary position this engine computes is 0.44 arcseconds. An arcsecond is a three-thousand-six-hundredth of a degree. That measurement is committed to the repository and the figure on this page is read from it, so the claim and the measurement cannot drift apart. Anyone who wants to check a single position independently can pull it from JPL’s Horizons service and compare.

That precision matters more here than it does for most verdicts, because a boundary test is exactly the kind of test where a small positional error changes the answer. A graha eleven degrees clear of the axis is not in dispute. A graha a tenth of a degree clear of it might be.

Which puts the uncertainty where it belongs. The loose parts are the choices: mean node or true node, whether the ascendant is tested alongside the seven, and what the local time offset was at the place of birth. That last one is a surprisingly common failure. India’s civil time has a documented history of offsets and the IANA time zone database records it; a birth time entered without the right offset shifts every longitude in the chart, and the Moon most of all.

Where does the classical evidence actually sit?

Thinner than the presentation suggests, and saying so costs something, so it should be paid. Kaal sarp is not a prominent category in the oldest matching literature. It does not stand alongside the eight kootas as a standard axis of compatibility, and a reader searching the Brihat Parashara Hora Shastra for a chapter on it will not find one in the place the popular literature implies. It circulates far more widely in twentieth-century practice than in the classical corpus.

There is a reason the pedigree question matters more here than it would elsewhere. Most of what a matching report says can be traced: a koota comes from a named table, a cancellation comes from a verse someone can point at, and where a rule cannot be traced we say so. Kaal sarp arrives with a great deal of ancient-sounding presentation and comparatively little of that apparatus, and the gap between the two is where the graded version was able to grow. A reader who has been handed it as an ancient verdict is entitled to know that its documented history is shorter than the framing suggests, and also that this by itself does not make it wrong.

That is not the same as saying it is invented. Yogas defined by the enclosure of grahas between two points are a recognised form in the classical vocabulary, and the general pattern of a chart being read by which arc is occupied and which is empty is old. What is modern is this particular name, this particular list of twelve serpents, and above all the commercial apparatus that has grown around it.

The strict formulation is at least consistently stated wherever the category appears. The graded version is not stated anywhere in the classical texts we can cite. That asymmetry is the argument. We use the strict reading and say so, not because strictness is a virtue, but because it is the only version with a written definition to point at. The engine behind these pages cites the text each rule comes from, and a rule with no text to cite is reported as traditional and unverified rather than dressed up as ancient.

What does any of this have to do with a marriage refusal?

Kaal sarp is not one of the eight kootas and contributes nothing to the thirty-six points, so when it appears in a refusal it is a second, separate verdict travelling alongside the first. Two independent readings have been bundled into one sentence, and each is entitled to be examined on its own.

The bundling matters because the two behave very differently under scrutiny. Matching verdicts diverge because traditions genuinely disagree about doctrine: we ran every possible pairing under five documented traditions and 56.4% trigger a tested difference or veto across the combined census checks. The distinction between its separate checks is explained in the divergence census documents in full. The kaal sarp disagreement is not like that. For once the schools are not arguing about doctrine. They are arguing about whether a category with two states has been reported as though it had ten.

Where the condition is found, various traditions prescribe remedial rites: commonly a puja at a nodal shrine, sometimes a fast, sometimes a recitation. We can describe that these are prescribed, and by whom. We make no claim about what a rite accomplishes, in either direction; that is outside what a computation can speak to, and it is one of the things we have committed not to claim. What we will say is narrower and checkable: a rite prescribed for a condition the chart does not meet the test for is a rite prescribed for something that was not found.

What should you ask the person who gave you this verdict?

Ask for the list of which grahas sit inside the arc and which sit outside, with a longitude beside each name. A report that reaches a kaal sarp verdict without printing that list has not shown its work, because that list is the work.

If the list shows all seven on one side, the strict condition is present and the classical discussion applies to your chart. If it shows six and one, the next question is what the seventh graha is and how far from the axis it sits. The answer is a number of degrees. It either clears the axis or it does not, and if it clears by more than a degree, no reasonable choice of node convention closes the gap.

There is one answer that should end the conversation instead of continuing it: a refusal to produce the list, on the grounds that the matter is too subtle to reduce to positions.

The test is positions. That is all it has ever been.

Subtlety may well belong somewhere in a reading of a chart, but it cannot belong in the step where seven longitudes are compared against one boundary, because that step has no room for it. A practitioner who declines to show the sorting is declining to show the only part of the verdict that is not opinion.

Two follow-up questions separate a considered reading from a repeated one. Was the mean node or the true node used? And was the ascendant included in the test? Neither question is a trap and neither is rude: both are ordinary disclosures, and a practitioner who works from a stated method will answer them without hesitating. Most of what arrives at a family as “there is partial kaal sarp” has never been asked any of them.