How Vespiary works — the whole method, in public

The scores

Every pair of people gets five numbers. For each question two people both answered, an answer earns credit for how close it sits to the other person's — an exact match earns 1, one step away earns partial credit, the opposite pole earns 0 — weighted by how much that question actually splits people. Agreeing on something everyone answers alike is nearly free, so it counts for almost nothing; agreeing where the population genuinely divides is the signal. Items you have marked as core (up to ten, on your results page) count double for you. A pool needs at least 5 shared questions before its score shows at all, and profiles also show where a match ranks: "closer to you than X% of members."

Meaningless questions, and skips. A few of the metaphysical items carry one extra option — "The question itself is meaningless — there is no proposition here to evaluate" — because for a real school of thought (the logical positivists; our own Logical Positivism profile answers this way) "God exists" isn't false, it isn't even a claim. Two people who take that option match fully on the item; against a rating, the item simply doesn't count. And skipping is a first-class act here, not a failure: every explicit skip is recorded, what you refuse to answer is part of your reality model, and when two people have skipped five or more of the same questions their profiles say so — the same reality-boundary. Skip rates also feed the admin's question-health view: a question everyone skips is a broken question.

Practices. Twenty items ask not what you think but what you do — gardening, hunting, training, stocking supplies, where you would live — and, crucially, why: each item's options are motives, not yes/no. What you do leaks what you believe, often more honestly than what you say. Practices sit in their own layer: they don't move your Twin % or World %, but they are a lens of their own — Life % ("lives like you"), shown on profiles, cards, and the quick quiz alike, and they have their own map (the "How you live" map on the bubbles page). Because the options are categorical picks rather than steps on a scale, Life % is scored by simple matching: the share of shared items where you picked the same option, with no partial credit for near picks — there is no halfway between cats and dogs. It needs at least 8 shared practice questions before it shows, and it counts toward Kindred %: two people can share every conclusion and still live different lives — or share none, and live the same one.

A note on "facts." Five items in the bank have answers that can be checked against the world — global poverty trends, the Monty Hall problem, where a tree's mass comes from. The site deliberately takes no position on them. They are scored exactly like every other conclusion: as a place you stand, next to other people standing there or elsewhere. Vespiary's job is to find the people inside your reality bubble — not to grade your bubble against reality. Two people can be wrong about all five, in the same way, and be perfect matches. Whether a bubble's shared picture is *true* is a question for the forum, for argument — not for a scoreboard.

Opposition % is the collision score. It runs over the belief items only, and for each one it multiplies the distance between two answers by how far both people stand from the unsure center. Two people at opposite poles, both certain, score near 100; a person who answers "genuinely unsure" to everything opposes nobody, however different their other answers look — you can only collide where both of you actually stand somewhere. It is a public score, like the others: it describes a relationship, not a failing.

Put Twin and World together and most pairs fall into one of four shapes, shown as a small label — you two: — on profiles and cards. A Kindred mind thinks like you and lives in your world. A Mirror thinks like you and believes the opposite — your reasoning twin. A Tribesmate shares your conclusions but got there by completely different machinery. An Opposite is a genuine head-on collision: different world, held with certainty on both sides.

Two lenses, and reasoning twins

Dozens of questions form a belief battery: flat statements — God exists. Eating meat is morally wrong. Free markets do more to reduce poverty than government programs do. — each rated Certainly false → Certainly true, answered in seconds. No scenarios, no loopholes to hide in: the reasoning questions measure the machinery, these measure where it landed. On your results page the two lenses produce a contrast view: your reasoning twins — people whose Twin % with you is high (75%+) but whose beliefs diverge — with the exact beliefs you split on, side by side. A disagreement from someone who reasons badly is easy to dismiss; the same disagreement from your own machinery, honestly run by someone else, is the one argument that can't be waved off — which is exactly why a tribe's shared conclusions, the ones nobody inside it stress-tests, are the ones worth surfacing.

Consistency % comes from eleven mirrored pairs hidden in the bank: the same question asked twice with exactly one detail swapped — whose reasoning, whose belief, which side funded it, how far away it is. Same answer both times scores 1; one step apart scores ½; two or more steps scores 0. Consistency is not correctness: a person can be consistently wrong, and the site never merges the two numbers. Consistency is also private — it appears only on your own results page, never on your profile, the board, or your share card.

Bubbles

Every member's answers form a position across all questions. Members are placed in the bubble whose defining position they're nearest to. Bubble names are human-approved summaries of what actually distinguishes the cluster — never insults, never self-chosen labels. As membership grows past a few hundred, clustering will be computed from the answer data itself rather than archetypes. The bubbles page opens with the map: every person plotted by similarity in two spaces — a thinking map and a belief map — where distance on the page means difference of mind, and the bubbles are drawn as the regions they occupy.

Why “vespiary”?

A vespiary is a wasps' nest. Wasps don't find one ready-made; they build it themselves, together, and then work out of it. We're borrowing that as a metaphor. Your vespiary is the group of people you put together here, one person at a time, from others who share your model of reality — shared rules for what exists, what counts as evidence, whom to trust. The site doesn't ask what you call yourself; it measures your answers and helps you find them — but the building is yours: save the people you want to think with. Some conversations only work once the basics are shared — what humans evolved to eat is not a discussion you can have with someone who doubts evolution, or for whom grain is doctrine. Inside your vespiary those premises are settled, so you can explore further and go deeper together. Your vespiary page maps exactly that — what's settled among your people, and what's still live.

The scores wear the site's colors everywhere: Twin — thinks like you World — believes like you Life — lives like you Kindred — the blend

Privacy

Location is a city label and a ZIP used only to compute distance — never an address, never a pin. Answer visibility is yours to set. Accounts and all their data delete in one click in Settings. We do not sell data, and there are no fake or AI-generated members: every profile is a real account. (If AI participants are ever added for the debate project, they will be labeled as AI, in the open.)