Method

How this is scored

There are two different mathematical objects here, and conflating them is the mistake that makes quizzes like this feel arbitrary.

The distinction everything follows from

A posterior probability over traditions is what chooses the next question — it answers “which question would tell me most?”. An agreement score is what you are shown — it answers “how closely do my positions match this tradition’s?”. The first is a probability and sums to 1. The second is a similarity, is independent per tradition, and does not sum to anything. Showing you the first while calling it the second is why “87% Buddhist, 84% Unitarian” feels like it means something and doesn’t.

What you are shown

agreement(t) = 1 − ( Σᵢ wᵢ · |uᵢ − vt,i| / (1 + σt,i) ) ⁄ Σᵢ wᵢ

uᵢ your inferred position on axis i, in [0,1] · wᵢ confidence in that position — zero if the axis was never established, reduced if you answered “I don’t know” · v the tradition’s position · σ the internal variation the tradition genuinely tolerates, which softens the penalty for diverse traditions.

Every tradition is scored on its own 0–100 scale and the set is deliberately not normalised. Two traditions can both score 91%, and that is a true and interesting statement about them. Normalising is precisely what makes the numbers depend on which other traditions happen to be in the list.

What the engine does

q* = argmaxq [ H(T) − Σa∈q P(a) · H(T | a) ]

A posterior P(t) is maintained over all 65 traditions. For each candidate question the expected reduction in entropy is computed across its possible answers, and the question that reduces it most is asked. The likelihood comes from the same v and σ through a logistic kernel: an answer far from a tradition’s position, relative to its dispersion, is unlikely under that tradition.

The constraints that keep it honest

  • Between 8 and 14 questions. Below eight it feels cheap regardless of the mathematics; above fourteen completion drops faster than accuracy rises.
  • At least 10 axes established before stopping. Pure information gain will happily resolve a ranking on four axes. You would not believe it, and you would be right not to — a result built on four axes is genuinely fragile.
  • Coverage across all five groups — at least 2 from the divine, and at least one each from knowledge, self, practice and world. A1 is asked directly in every session, because it separates the non-theistic results from the other fifty-five.
  • No more than 2 questions about the same axis. Stops the engine drilling one dimension it happens to find informative.
  • “I don’t know” is a real answer. It yields a flat likelihood — no update at all — and lowers the confidence reported with your result. No position you declined to take is ever imputed to you.

The prior, and a trap

Weighting the starting posterior by how many people hold a tradition makes question selection meaningfully more efficient. It also biases results toward the largest traditions, and if it touched the results the quiz would be worthless. Because these are two different objects, both are available: an adherence-weighted prior for choosing questions, and a flat one for the agreement scores you are shown. Your result never depends on how popular a belief happens to be.

The seventeen axes

None presupposes theism, none uses insider vocabulary, and each is one a thoughtful person can answer about themselves. Every tradition also carries a dispersion per axis — how much internal variation it genuinely tolerates. Hinduism is wide on plurality and narrow on belonging; Reform Judaism is wide almost everywhere. Without that, internally diverse traditions get unfairly penalised for any answer away from their centre.

A · The divine

  • A1 Divine realityWhether anything exists beyond the physical world described by ordinary observation.nothing beyond the naturala divine reality certainly exists
  • A2 PersonhoodIf something ultimate exists, whether it is the sort of thing that can be spoken to.impersonal principle or ordera personal being one can address
  • A3 PluralityWhether the sacred is singular or populated by many distinct powers.strictly onemany powers and spirits
  • A4 LocusWhether the ultimate stands outside the world or saturates it.wholly beyond the worldfully present within it

B · Knowledge and authority

  • B1 Source of truthWhere reliable knowledge about ultimate things comes from.revealed and fixedreason and direct experience
  • B2 Reading of sacred textHow a foundational text — if there is one — is meant to be read.literal and bindingmetaphorical, or no text at all
  • B3 Institutional authorityWho settles a disputed question about what is true or right.clergy and tradition decideindividual conscience decides

C · Self and destiny

  • C1 Survival of the selfWhether anything of a person persists after the body stops.ends at deathcontinues in some form
  • C2 Shape of timeWhether existence runs once toward an end or turns in cycles.one life, then judgmentcyclical return
  • C3 The human conditionWhat is actually wrong with people, if anything.fallen and in need of rescueignorant but fundamentally sound

D · Practice

  • D1 Mechanism of the ultimate goodWhether the best outcome is received or earned.given as graceachieved by effort or insight
  • D2 Centrality of ritualHow much prescribed practice matters to living well.essential and prescribedoptional or absent
  • D3 The bodyWhether embodiment is an obstacle to be managed or a good to be enjoyed.to be disciplinedto be celebrated

E · World and others

  • E1 ExclusivityWhether one way is uniquely right or several lead to the same place.one true pathmany valid paths
  • E2 Origin of the cosmosWhether the universe was meant, or simply is.an intentional actbeginningless or natural
  • E3 Locus of belongingWhether belonging is something you are born into or something you pick.a people, inheritedan individual choice
  • E4 Standing of natureWhether the natural world is a resource, a gift, or a subject in its own right.creation, given for useanimate and itself sacred

The question bank

85 questions, each tagged with the axes it loads on and a loading weight. Two rules govern the writing and both are harder than they sound. No insider vocabulary: “do you believe in grace?” only works for someone already inside a Christian frame. No presupposed theism: “what does God want from you?” has decided two axes before you have answered. Ask about the situation, not the doctrine.

Sources

Every one of the 1105 cells cites one of 66 sources. The whole matrix is published — see the traditions — because an accuracy claim you cannot check is not an accuracy claim.