AskSoma vs AstroNidan

Two strong claims worth reading closely

AstroNidan describes itself as "the world's first AI-ML powered Vedic astrology platform" and says it uses NASA-based planetary data. It offers a free birth chart, daily horoscopes, career and marriage predictions, Dasha analysis, Yoga identification and Prashna Kundali, plus premium reports and consultations with verified astrologers. It is a substantial product and the feature list is real.

The two headline claims deserve unpacking, because we make a version of one of them ourselves. "NASA-based planetary data" almost always means JPL's DE ephemerides — the same planetary tables that sit under Swiss Ephemeris, and under us. That is a genuine accuracy credential and it is also table stakes: any serious astrology product is working from the same underlying positions. It tells you the arithmetic is right. It tells you nothing about whether the interpretation is.

What "ML-powered" would have to mean to be meaningful

The "AI-ML" claim is the more interesting one. Machine learning implies a model trained on labelled outcomes — charts paired with what actually happened — and in astrology that training set essentially does not exist at public scale. Nobody has a corpus of a million birth charts annotated with verified life events.

So we would rather describe our own architecture plainly than compete on adjective. AskSoma is not a trained predictive model. A deterministic engine computes Dashas, transits, dignities and yogas from classical rules, and a language model puts the result into English. The rules are Parashari, the ayanamsa is Lahiri, and both are published. Nothing is learned from outcomes — which is why we keep a public record of predictions and how they resolved, since that record is the only honest evidence either of us could offer.

Prashna Kundali is a genuinely different instrument

AstroNidan offers Prashna Kundali, and it is worth explaining because it is the feature on their list that we do not have. Prashna casts a chart for the moment a question is asked rather than the moment you were born, on the classical reasoning that a question arising at a particular instant carries that instant's sky. It is the traditional answer for people who do not know their birth time at all.

AskSoma takes the other road for the same problem: birth-time rectification. Rather than casting a new chart for the question, we narrow the unknown birth time by testing candidate times against events that already happened in your life — a marriage, a move, a job change — and keeping the window that fits. Both are legitimate; they answer different questions. Prashna is faster and needs nothing from you. Rectification is slower and gives you a natal chart you can keep using for the rest of your life.

Two different meanings of "validated"

AstroNidan's site carries a "Our Predictions. Validated by Users." section. User validation is real feedback and we run our own equivalent — every Soma reply can be rated, and those ratings shape the product.

But it is worth naming what user validation can and cannot establish, because the distinction is easy to lose. A user rating a prediction highly usually means it resonated — it described something they recognised. Resonance is not accuracy. A description vague enough to fit most people will resonate with most people; that is the Barnum effect, and astrology products are unusually exposed to it.

The only rating that tests a prediction is one collected after the predicted window closes, on a specific dated claim, including from the users for whom it failed. That is why AskSoma's Prediction Ledger records the window's dates when the forecast is made and asks afterwards whether it happened — and why our published accuracy figure will stay unpublished until enough windows have actually closed to make it mean something.