Two apps, two Rahus, sometimes two different signs — the second silent setting in Vedic software, measured and resolved.
Put one birth into two respected apps and compare the Rahu row. Often you will find a gap of a degree or more — far larger than any modern ephemeris error, which is measured in fractions of an arc-second. Occasionally the two Rahus sit in different signs, at which point downstream statements start diverging loudly: one report declares a Kala Sarpa yoga, the other does not; one places Rahu in the 7th house, the other in the 8th; one says Rahu entered Meena in March, the other says April.
As with the ayanamsa, nobody is miscalculating. The two engines are answering different questions about the same orbit, and each question has a defensible pedigree. To choose between them intelligently — as a user or as an engine author — you need to know what a node actually is.
The Moon's orbit around Earth is tilted about 5°09′ to the ecliptic (the Sun's apparent path). Two planes that intersect define a line, and that line meets the zodiac circle at two opposite points: where the Moon crosses the ecliptic going north (the ascending node — Rahu) and going south (the descending node — Ketu). They are geometric points, not bodies — which is precisely why the tradition's imagery of an invisible, shadowy graha is astronomically perfect: the nodes are where eclipses happen, the "swallowing" points of Sun and Moon. When the classical texts call Rahu the eclipse-maker, they are describing the ascending node's actual function.
Now the subtlety. The Moon's orbital plane is not fixed: the Sun's gravity drags it, making the node line regress — slide backwards through the zodiac — completing a full circle in ≈18.6 years (≈ −3′11″ per day on average; ≈1½ years per sign). But that regression is not smooth. Superimposed on it is a wobble driven by the Sun's changing pull across the lunar month and the eclipse seasons: the instantaneous node swings around the average position with an amplitude of up to ≈1°45′, on a period of about 173 days (half an eclipse year). During parts of that wobble the instantaneous node briefly moves forward along the zodiac.
Figure 1 — One orbit, two readings: the smooth average versus the instantaneous wobble.
So the software question is exact: when a chart asks "where is Rahu?", do you answer with the smooth average (mean node) or the instantaneous geometry (true node)? Western computing culture — via the Swiss Ephemeris, whose node default is "true" — pushed much modern software toward the true node. Indian ephemeris tradition — the printed panchangams, Lahiri's tables, the Rashtriya Panchang lineage — has always tabulated the mean node. Your two apps inherited different ancestries.
The gap between the two Rahus on any given day is the wobble's instantaneous value — anywhere from 0 to ≈1°45′, averaging ≈1°07′ over time (the mean of |A·sin| for amplitude A is 2A/π). From that, the population-level consequences follow directly:
| Question | Computation | Result |
|---|---|---|
| How often do the two conventions put Rahu in different signs? | mean |gap| ÷ 30° | ≈ 3.7% of days/births (peaks near sign boundaries at up to 5.8%) |
| …in different nakshatras? | mean |gap| ÷ 13°20′ | ≈ 8.4% of days/births |
| …in different whole-sign houses? | same as signs | ≈ 3.7% |
| How far apart can apps date Rahu's sign change? | ±1°45′ ÷ 3′11″/day | up to ≈ ±33 days; a few weeks is routine |
| Does it change Manglik status, Guna Milan, nakshatra porutham? | those depend on Mars/Moon | No — node model touches only Rahu-Ketu themselves |
Table 1 — Computed disagreement rates between mean-node and true-node output. "Different sign ~3.7% of births" means roughly one chart in 27 carries a sign-level Rahu discrepancy across apps.
The sign-ingress row explains a perennial internet argument. Every eighteen months or so, "Rahu enters [sign]" becomes a trending topic — and the dates circulating from different sources disagree by weeks. They are not sloppy; they are mean-node dates and true-node dates. Neither side knows the other is answering a different question.
Most single-chart interpretations shrug off a degree of Rahu. One famous verdict does not: Kala Sarpa yoga/dosha, classically read when all seven classical grahas stand on one side of the Rahu–Ketu axis. That verdict is binary, and the boundary is the nodal axis itself. If the planet closest to the axis sits within the mean–true gap of the moment — up to 1¾ degrees — then one convention closes the circle and the other breaks it: Kala Sarpa under mean node, no Kala Sarpa under true node, or the reverse.
How often? Take the planet nearest the axis across many charts: the chance it falls inside the gap zone on a random day is a few percent of the charts that are near-Kala-Sarpa to begin with — a small slice of all horoscopes, but a meaningful slice of precisely the people googling "do I have Kala Sarpa dosha", because near-boundary cases are the ones that get conflicting answers from different sites. If you have received a yes from one app and a no from another, the odds are high you are exactly such a boundary case — and the disagreement is the setting, not your fate. (Our house view on doshas applies here as everywhere: a dosha reading is a starting point for understanding and remedies, never a sentence. A verdict that flips with a software toggle should be held especially lightly — and that is an argument for engines to disclose their toggle, which is the real point of this article.)
It is the instantaneous truth of the geometry: at any moment, the plane crossing is where it is, wobble included. If an eclipse happens, it happens at the true node. Astronomy software naturally produces it; the Swiss Ephemeris made it a default; and some modern astrologers argue that if we can compute the real thing, averaging is an anachronism.
Astronomical: the node is not a body but an orbital element — and for elements, the secular (mean) motion is the physically meaningful trajectory; the 173-day oscillation is a periodic perturbation around it. Mean motion is what the node's long-term dynamics actually are. Classical: the jyotisha tradition is unambiguous that Rahu moves permanently retrograde — vakragati — and the dasha, gochara and shastra logic is built on that steady backward march of about 1½ years per sign. The true node periodically moves direct, which contradicts the texts' description outright; the mean node retrogrades always, exactly as described. When a computational choice must be made, the model that matches the tradition's own stated behaviour of the graha has the stronger claim inside that tradition. Practical: the Indian ephemeris lineage — Lahiri's tables, printed panchangams, the documents in your family's cupboard — tabulates mean nodes. Choosing mean keeps a modern chart comparable with seventy years of printed jātakams.
juno.date computes Rahu and Ketu as mean nodes. We did not begin there by ideology: during our validation campaign (documented in our AstroSage review), our engine's node positions were cross-checked against established references to arc-second agreement — and the exercise surfaced exactly the ancestry split described above, settling our default on the Indian-ephemeris convention. The switch is recorded in our engine changelog, and every report prints the node longitudes to the arc-second in the Nirayana sphuta section, so a true-node practitioner can see precisely what we computed and translate.
Suppose two sites gave you opposite Kala Sarpa answers and you now know why. Three practical steps. First, prefer the reading whose convention matches the rest of your documents — for almost every Indian family, that is the mean node, the same lineage as the printed horoscope made at your birth. Second, look past the binary label at the actual geometry, which any good report shows: a chart with every planet hugging one side of the axis reads the same regardless of a boundary technicality, and a chart that barely closes on a toggle was never a strong yoga in either convention. Third — in our standing house style — remember that even the tradition itself treats Kala Sarpa with nuance and remedies, not fatalism; a verdict that a software setting can flip deserves curiosity, not fear. The sky did not change between your two browser tabs.
It is worth tracing how the split reached consumer software, because the story explains why neither camp thinks of itself as making a choice at all. The Indian lineage is print-first: for most of the twentieth century, an astrologer's node positions came from a published ephemeris — Lahiri's tables above all — and those tables, following the siddhantic treatment of Rahu as a smoothly regressing point, print the mean node. Software written inside that culture (and any engine validated against Indian printed references, ours included) reproduces mean nodes as a matter of fidelity. The Western lineage is library-first: when astrological computing standardised on the Swiss Ephemeris in the 1990s, that library's convenient default for the node was the true (osculating) position — a perfectly reasonable astronomer's choice — and a generation of apps worldwide inherited it silently, including many India-focused apps built on Western libraries. Neither camp's programmers typically documented the setting, because each believed it was simply computing "Rahu". The result is today's quiet incompatibility: your two apps disagree not because one is sloppier, but because one descends from Lahiri's printing press and the other from a Zurich code library. Once you can name the ancestries, the mystery evaporates — and the diagnostic in Section 6 tells you in a minute which family tree any given app belongs to.
A short digression that earns its keep. In the Puranic image, Rahu is the severed head that swallows the Sun and Moon; eclipses are his act. Strip the story to its mechanism and it is precisely correct: an eclipse can only occur when a syzygy (new or full moon) happens near the node line — the Moon must cross the ecliptic at the same longitude where it aligns with the Sun. The nodes are literally the eclipse-makers. This is also why the true node's wobble has its 173-day rhythm: that is half the eclipse year, the interval between the Sun's successive passages through the node line, when the solar perturbation on the lunar orbit peaks. Eclipse seasons, the node wobble, and Rahu's mythology are three descriptions of one mechanism. We mention this not as ornament but as method: when a classical description encodes real celestial mechanics this cleanly, taking the tradition's own characterisation of Rahu's motion seriously — always retrograde, steadily devouring the zodiac backwards — is not sentimentality. It is reading the same source that got the eclipses right.
A practical close before the FAQ. Open your jātakam on juno.date and find Rahu in the Nirayana sphuta table — say it reads 23°41′ of Kumbha. Now check the same birth on the other app that worried you. Three outcomes: (a) agreement within a few arc-minutes — same convention, any residual difference is ayanamsa-related (see that article); (b) a gap of up to ~1°45′ with the sign usually agreeing — you are looking at mean-vs-true, and nothing in your life changed; (c) a gap that puts Rahu in a different sign — you are one of the ≈3.7% boundary cases, and every statement the two reports make about Rahu's house, sign, and any Kala Sarpa call should be read with this article in hand. For case (c), the useful next step is not to pick the scarier report: it is to look at the actual longitude. A Rahu at 29°50′ of a sign is a boundary Rahu in both conventions — treat sign-based statements about it as soft, whichever app produced them, and lean on readings anchored to the nakshatra and degree, which both conventions nearly agree on.
The dasa timeline is keyed to the Moon's nakshatra, so the node model does not shift your dasa boundaries. It does affect readings of Rahu dasa and the houses/signs Rahu-Ketu occupy — and, at the transit level, when "Rahu changes sign" for gochara purposes (by up to a month between conventions).
KP practice commonly pairs its own ayanamsa with the true node — so a KP chart can differ from a Lahiri-mean chart on two settings at once. When comparing across that divide, translate both: subtract the ≈6′ ayanamsa gap, then allow the node wobble. Our arc-second sphuta display exists precisely so such translations take a minute instead of an argument.
Yes — in both conventions Ketu is the same node line's other end, exactly 180° away. The mean-vs-true gap applies to the axis as a whole.
Because "where is the node" is not the question a dasa or gochara framework asks. Those frameworks model Rahu as a slow, steady influence unfolding over eighteen-month sign residencies; feeding them a point that jitters degrees back and forth every few months mixes a fast periodic signal into a slow-reading instrument. The mean node is the slow signal, isolated. Astronomy offers both; astrology's own structure selects one. That is a modelling argument, not a loyalty oath — and it is the same argument by which astronomy itself uses mean elements for long-term orbital work.
It is half the eclipse year (≈346.6 days) — the interval between the Sun's alignments with the node line. The Sun's pull on the lunar orbit peaks around those alignments, driving the oscillation. The nodes and eclipses are one mechanism, exactly as the Rahu mythology encodes.
Yes — several traditions read Rahu's and Ketu's nakshatras (and their lords) in dasa interpretation, and some remedial traditions key mantras and gemstone timing to them. With an ≈8% inter-convention disagreement rate at nakshatra level, this is another place to confirm the convention before comparing sources — especially for Ketu, whose nakshatra-based readings around Gandanta zones attract anxious searches.
Strongly — navamsa cells are 3°20′ wide, so a gap averaging ≈1°07′ moves the nodes' D9 position in roughly a third of charts. Any D9-based statement about Rahu-Ketu (marriage significations especially) should be treated as convention-dependent unless the engine is known. This is the nodes' version of the divisional-chart amplifier we describe in the ayanamsa article.
The Indian almanac lineage (Lahiri's ephemeris, Rashtriya Panchang tradition and the printed panchangams that follow them) tabulates mean nodes — one reason your grandfather's horoscope and juno.date will agree about Rahu while a true-node app disagrees with both.