How We Calculate — Method, Sources and Limits

Every number on this site is arithmetic, not opinion. This page publishes the tables behind it so you can reproduce any figure yourself, and states plainly where the limits are. It is generated from the same code the calculators and the paid reports run, so it cannot fall out of step with them.

Last updated 2026-11-04

Which system this site uses

Chaldean, throughout. Every calculator, every compatibility page and every paid report uses the Chaldean letter values. Where a page shows a Pythagorean figure it is labelled as such and shown for comparison only.

This matters because the two systems give different answers for the same name, and most disagreements between numerology sites come down to nothing more than that. Chaldean assigns letters by sound and never gives a letter the value 9. Pythagorean assigns them by alphabetical position and does use 9.

Chaldean letter values

The table every name number on this site is computed from.

Pythagorean letter values

Published so you can check a figure another site gave you. Nothing here is calculated with these values unless the page says so explicitly.

Ank-maitri — which numbers are friendly

The classical friendship table. It is used to explain compatibility, never to compute the score. The two are independent sources and on 11 of the forty-five pairs they genuinely disagree — including 1 and 2, 1 and 7, 2 and 4, 2 and 7 — and where that happens the page says so rather than smoothing it over.

Note that this table is not symmetric everywhere. Classical naisargika maitri treats some friendships as one-directional, so a number can count another as friendly without that being returned. That is a property of the tradition, not an error, and we have left it as the sources give it. Our compatibility scores, by contrast, are symmetric by construction: a pair gets the same figure whichever name is entered first.

Vehicle numbers use a different table, on purpose

Vehicle numerology has its own published friendship convention and it is not the ank-maitri table above. We use the vehicle convention as published by Dr. J C Chaudhry on our vehicle calculator, because that is what the established sources in this field use, and a reader comparing our answer to theirs should not find a silent disagreement.

The two tables genuinely conflict. Counting every mulank against every number, they disagree on 36 of the 81 combinations — including 1 and 2, 1 and 4, 1 and 7, 2 and 2. On mulank 8 and 8, for instance, ank-maitri counts the pair friendly while the vehicle convention advises against it. We would rather publish both tables and name the conflict than pick one quietly and present it as settled.

A vehicle number is also read from the digits alone — the last four of the registration, being the part a buyer can choose. The letters are not converted. That differs from name numbers on this site, which do use the Chaldean letter values.

How a compatibility score is produced

Each pair of root numbers has a base score. A relationship-type adjustment is then applied — marriage +3, because marriage rewards stability more than the other readings do — and the result is held between 20% and 97%.

The clamp is deliberate. No pairing is a certainty and none is hopeless, and a table that produced 0% or 100% would be claiming more than the method supports. The same function generates the figure on every page it appears on, so a score can never differ between two pages of this site.

How a name number is scored against a birth date

The compatibility score above reads two people against each other. Choosing a name is a different question: one name number against one birth date, which supplies two numbers of its own — the Mulank from the birth day, the Bhagyank from the whole date.

The name number is scored against each of them on the friendship table above, and the two results are combined with the Mulank weighted twice as heavily as the Bhagyank. That weighting is a judgement, not a derivation: the Mulank comes from the birth day alone and is what Indian numerologists weigh first when a name is being selected. We state it here rather than bury it, and the same function produces the figure on every baby-name page, so it cannot differ between two pages of this site.

Both component scores are shown to the reader alongside the combined one. A name number can suit the Mulank well and the Bhagyank poorly, and a single combined figure would hide that; where the two differ by 25 points or more the page says so explicitly.

Reading the result: 70 and above is a strong fit, 55 to 69 workable, below 55 weak. Scores cluster roughly between 40 and 80 because the friendship table rarely goes to extremes, so nothing reaches 100 and the pages print the best numbers achievable for your particular date as the real ceiling. A ten-point gap between two names is a difference worth acting on; a two-point gap is not, and the comparison page says that rather than crowning a winner.

Until 7 October 2026 the baby-name pages did not score this at all. They asked only whether a name number was friendly with the Mulank OR the Bhagyank, and printed a tick if either passed. That marked quite different names identically strong — and it hid disagreements, because a number scoring 85 against one birth number and 40 against the other was reported as a clean pass. Anyone who used those pages before this date saw a verdict we would no longer stand behind.

A correction, 24 September 2026 — the name compatibility tool

Until this date the name compatibility calculator did not use the score table above. It carried a second table of its own, and that table was wrong in three ways. We are recording it here because a reader who ran the tool before this date was given a figure we would no longer stand behind, and a methodology page that only documents the parts we got right is not worth much.

First, it disagreed with the published table on 19 of the 81 number combinations. On 5 and 7 it had the sense reversed: the published score for that pair is 35, the lowest of any pairing for 7, while the tool's own table listed 5 as one of only two numbers compatible with it.

Second, it was asymmetric on one cell — 4 counted 6 as compatible, 6 did not count 4 — with the result that entering the same two names in the opposite order changed the answer. On a 4,560-pair sweep, 226 pairs scored differently depending on who was typed first, and some crossed a verdict boundary in the process.

Third, the percentage included a few points derived from a hash of the two names. It was there so results would not cluster on the same round numbers, which made every score slightly bespoke and meant no reader could reproduce one from the tables on this page.

All three are gone. The tool now reads the same table as every other compatibility page on this site, so a name pair returns the figure printed above and returns it whichever order the names are entered. Scores moved: across that same sweep the mean fell from 72.6% to 64.1%, and roughly two thirds of pairs now score lower than they did. We would rather publish a number you can check than a higher one you cannot.

It was noticed because a reader compared us against two other sites on the same pair of names. Both of those sites, and this one, computed the same two name numbers — the arithmetic was never in question. The disagreement was in the scoring layer, which is exactly where a site should expect to be asked to show its working.

Where the name meanings come from

A name's number is arithmetic and you can check it from the letter tables above. A name's meaning is not — it is a claim about language, and it needs a source.

Our name database was audited in October 2026 and most of it did not survive. It held 2,998 names but only 943 distinct meanings between them: 340 names were all glossed "blessed", 61 "mature", 35 "day light", and 1,187 of them sat in 111 groups that shared both a word-stem and a copied meaning. That is the signature of a script appending suffixes to a stem, not of a collected vocabulary. Alongside it, 483 names had no meaning at all, and some that did were wrong — alaris was filed as Sanskrit for "beautiful" and is not a Sanskrit word, while agyeya, which genuinely is Sanskrit (अज्ञेय, "unknowable"), was filed as having no known meaning.

The replacement is smaller and sourced. Each name now records where its meaning came from: a dictionary, a named text, common usage, or nothing at all. Where we hold nothing, the page says so instead of offering a guess — which is why a few of our busiest name pages now state plainly that we cannot verify a traditional meaning.

Part of the current set is imported from the Nurturepedia open baby-names dataset, used under the MIT licence. Those entries are marked as common-usage rather than dictionary-grade, because the upstream carries no citation for individual names. Where we have written an entry by hand with a dictionary root, the hand-written one is what the page shows.

We do not record a religion for a name. Names have languages of origin, which can be checked; they do not have religions. Sanskrit-derived names are used by Muslim and Christian Indian families constantly, and sorting a name list by faith misstates how naming actually works.

What we do not do

We do not predict outcomes. Not whether a marriage will happen, not the month a job arrives, not an exam result, not a rank, not a medical outcome. None of those can be derived from a birth date, and our chat and our reports are built to refuse them rather than guess.

Choosing a date is a different thing, and we do it. Asked which dates in a window suit a couple, or which months suit an application, we apply a published rule to the numbers and show the working — that is what muhurat selection has always been, and it is reproducible rather than predictive. The distinction we hold to is that we will tell you which date the method favours, and we will not tell you what will happen on it. Where a genuine muhurat needs a tithi, a nakshatra or a sunrise-based window, we say so and send you to a panchang for your city rather than inventing the precision.

We do not infer gotra, caste or community from a name or a birth date. They descend through family, not arithmetic.

We do not advise on lotteries, betting or any game of chance, and we do not tell you what another person is thinking or whether they will return.

We do not answer medical questions numerologically. If a page or the chat is asked one, it says so and points to a doctor.

Corrections

When we find an error we fix the data rather than hide the output. In June 2026 the score table and the editorial copy disagreed on 36 of 45 compatibility pairs; they were reconciled to a single source and automated checks now fail the build if they diverge again. Similar checks cover the letter tables, the celebrity arithmetic and the compatibility grid.

If a figure on this site looks wrong to you, it is worth telling us — it is checkable, which is rather the point of publishing the tables.