A site that never publishes a correction is not a careful site. It is a site that does not check.
This page is the log. It records errors this site published, with dates, and it records corrections to claims found elsewhere. The method page explains the standard these are held to.
How corrections work here
Errors are fixed and logged, not quietly deleted. The page is corrected, and the entry stays here.
Our own errors are logged on the same terms as anyone else’s. Several entries below are claims this site made about the accuracy of its own work.
Corrections are seva. If something here is wrong, tell us. A source helps and is not required.
Errata: errors we published
21 August 2026: we undercounted the problem on this very page
Earlier the same day, this page reported that about 194 classical attributions across the site were dressed as verbatim quotations. That number was wrong, and wrong in the direction that flattered us.
The count came from a pattern that searched for the titles of five works. It never matched the many attributions phrased through an author instead: Parashara says, Kalyana Varma’s text, Mantreshwara states. It also did not know about several works being quoted the same way, including Chamatkar Chintamani, Brihat Jataka and Lal Kitab.
The real figure, counted from the diff once the work was done, is roughly 690 quoted fragments in about 407 passages across 135 files. The zodiac articles alone held more than four times what we had claimed for the whole site.
There is no excuse worth offering. We measured with a tool we had not checked, on a page whose entire purpose is admitting what we get wrong, while confessing to exactly that failure elsewhere in the same sentence. The quotations are all now removed and the method page carries the corrected figure as a range rather than a false precision.
20 August 2026: we overstated our own Ashtakavarga verification
The Ashtakavarga bindu tables on this site were cross-checked cell by cell against an independent production engine, and we published the result as 504 of 504 cells matching across six charts spanning four countries and both hemispheres.
The cell count and the match rate were correct. The description of the spread was not. The six charts were Mumbai, Delhi, Ahmedabad, Chennai, London and New York: three countries, and every one of them in the northern hemisphere.
This mattered more than a miscounted country. A planet’s sign is an ecliptic longitude and is the same wherever you stand, but the ascendant depends on latitude, and the ascendant is one of the eight reference points every bindu is counted from. So the claim to have tested both hemispheres was not merely inflated; it named the one condition the test had never actually exercised.
Fixed by running the missing test rather than only softening the sentence. Five southern-hemisphere charts were added, Sydney, Buenos Aires, Johannesburg, Auckland and Santiago, and the whole set re-run: 924 of 924 cells matching across eleven charts and eight countries, in both hemispheres. The claim is now true. The full record is on the Ashtakavarga page.
20 August 2026: our tools page had a dead body of text
The tools page template never rendered its own page content. Everything written in the body of that page had been invisible since the template was built, which nobody noticed because the page looked finished. Fixed, and the page now carries its documentation links.
20 August 2026: our dosha count was wrong in our own notes
Internal documentation described the tool as testing “about 10 doshas”. Counted from the source, it tests 31 conditions: the figure was stale because nine conjunction doshas and five debt doshas are generated in loops rather than written out, so a naive count of the test functions missed fourteen of them. The dosha analysis page gives the real figure and the real list.
20 August 2026: our health pages did not defer to doctors
The health and addiction pages described serious conditions, opened on stories in which medicine had failed, and stated that Jyotish “diagnoses the root cause”. None of them told a reader to see a doctor, and the addiction page discussed addiction at length without naming a single helpline or route to treatment.
That was a real failure, not a stylistic one: somebody in trouble could have read those pages and found no way out of them. Both now open with a notice that nothing there diagnoses anything and nothing is a reason to delay treatment, the diagnosis claim is withdrawn, and the addiction page leads with the national helplines rather than with astrology. The relationships page carries a comparable notice about safety.
Corrections to the wider record
An ayanamsa is not a settled fact
Sites and software routinely present their ayanamsa as the sidereal zodiac, without naming it. Lahiri, Raman, Krishnamurti and others differ by enough to move a planet across a cusp or a pada in a chart cast near a boundary. This site computes on Lahiri and says so, and treats charts near a boundary as genuinely uncertain rather than as settled by our choice of parameter.
“Ancient” is not a citation
Claims are commonly attributed to “the ancient texts” without naming one. Where this site cites a classical work it names it, and where it names a work without an edition it now says that the citation points at a tradition rather than a verified line. That is a weaker claim than a footnote, and pretending otherwise would be the error.
A dosha is not a catastrophe for sale
The strongest commercial pressure in this field is on the fear attached to Mangal dosha, Kaal Sarp and Sade Sati. The classical literature that describes these conditions also describes their cancellations, and the cancellations are frequently present. Any presentation that names the condition and omits the parihara is incomplete, whatever it is selling. The tool here tests the cancellations against your actual chart and shows which apply.
Still open
We have not tested whether any of this predicts anything. Sourcing a claim correctly and computing it correctly are different from demonstrating it is true of the world. No entry on this page should be read as evidence for astrology. They are evidence that we check our arithmetic and our citations, which is a smaller claim and the only one we are making.
Our classical attributions are still not citations. Counted on 21 August 2026: more than a thousand references to Brihat Parashara Hora Shastra, Saravali, Phaladeepika, Jataka Parijata and Uttara Kalamrita across the articles, none carrying a verse number and two carrying a chapter. The quotation marks that made roughly 690 of these read as verbatim translation have now been removed, but the underlying weakness stands: an attribution without a verse is still not something you can check. Treat every classical reference here as what the tradition holds rather than what a page says. This remains the weakest thing on the site.
Tell us we are wrong
[email protected], or the contact page. Specific beats polite. If you can name the text, better; if not, tell us anyway.