The New Religion Census

The headline numbers

Why trust this denominator over the published ones? Because the published ones count a different animal. Melton's ~840-group American census, the "40–45 new religions founded per year in the US," Korea's "500+ foundings since 1945," Barrett's 6,000 African independent churches — those populations are overwhelmingly Bible-inheriting schisms, imported franchise branches, and charisma-only prophets who never wrote anything. Filter Korea's 500+ down to people who actually wrote out a new system and you are left with about six names. Writing a complete religion is rare enough that the worldwide two-century list fits on one page — which is what makes the census possible at all.

Outcomes: the census vs. the engineered controls

Grades: ★ major = ≥1M real adherents sustained across two or more generations. ◐ established = ~50k–1M, or smaller but deeply institutionalized and multi-generational. ○ persistent-small = alive and self-reproducing under 50k. ✝ dead = extinct, caretaker-relict, or absorbed. The second bar is the control group nobody expected: religions deliberately engineered by intellectuals who removed the supernatural parts — Comte's Religion of Humanity, the French Revolutionary cults, Ethical Culture, the Monist League, Sunday Assembly. Their record across 250 years is uniform.

The designed-class law, uniform across 250 years: no engineered religion that stripped out supernatural metaphysics has ever exceeded ~10⁴ committed members. Comte built liturgy, sacraments, a calendar, a priesthood — and peaked at 93 members in London. Their best outcome is secular downstream yield: the Saint-Simonians' alumni built the Suez Canal; Positivism got its motto onto Brazil's flag. But note the two designed systems that kept a supernatural payload and built real machinery — the Arya Samaj (infallible Vedas + a conversion rite + 900 schools) and Won Buddhism (karma-dharma metaphysics + an elected clerical council, zero schisms in five generations) — are both alive at scale. Design is not the fatal variable. Naturalism is.

The designed-supernatural option — engineering a religion that lives

Which raises the question the Comte class never answered: if you are honestly designing a religion — no angel visited you, and you say so — where does the sacred authority come from? The census's designed-supernatural survivors each solved that problem with a different mechanism, and the menu is worth laying out, because it is the actual engineering playbook:

Two properties are shared across the designed-supernatural menu. First, every mechanism keeps something genuinely at stake — karma, spirits, salvation — so the binding force the naturalists deleted is intact; the founder's honesty about construction changes where authority sits, not whether the stakes are real to the believer. Second, admitted construction confers immunity to exposure: the Wicca–Eckankar law (exposed origins don't kill) applies trivially to a religion that never hid its origins — there is nothing to expose. The designed-supernatural class trades the hot ignition of a prophet's charisma for slower, steadier compounding on institutions — which is exactly the trade the succession data says wins centuries.

The census

Every qualifying founding we could locate, worldwide, with what each one actually claims about reality (names link to Wikipedia). Corpus = what the founding generation wrote or dictated. Engine: F fertility/lineage, C conversion, B both. Motive is classified two sections down (V vision · S probable scam · D designed · C$ commercial · M mixed). The two ratings are separate axes and deliberately so, because they don't correlate: Soph = intellectual sophistication of the system (internal coherence, engagement with the philosophy and science of its day, argumentative depth) and Silly = the laugh test — would a person hearing the founding stories for the first time start laughing? Mormonism scores M·H: a genuinely worked-out materialist theology of eternal progression… dictated from golden plates read through a seer stone in a hat. H/M/L on both; the cross-tab gets its own section below. Borderline corpus calls are marked ⚑. Numbers are census/scholarly floors.

The adjacent class — major-corpus reform movements and schisms (excluded from the core rates; counting them raises every success number)

And the inverse control: the biggest mass successes with no qualifying corpus at all — Zion Christian Church (zero published theology, 5–12M), Kimbanguism (hymns only, 2–6M), the Marange Apostles (slim visionary text + polygynous fertility, 1–3M), Rastafari (borrowed pamphlets, ~1M — and it survived the literal death of its living God in 1975 by doctrinal denial). Where a Bible is already supplied, schism substitutes for scripture-writing. A corpus is neither necessary nor sufficient. What it is, is a marker — see the verdict below.

The four organs

Code every case on four traits and the whole census sorts itself. The survivors at scale have all four; remove any one and the failure mode is predictable — each organ's absence kills in its own signature way.

How they die

Primary causes of death or collapse across the full failure record (the 20 core-census deaths plus the wider corpus-bearing failure list and the designed class; a movement can carry two factors, so counts overlap). Hover a bar for the cases.

What "nothing at stake" actually looks like — because it deserves suspicion

Fair challenge: the top bar sounds like a hypothesis ("they lacked the thing we think matters") rather than an observed cause of death like a police raid. So here is the observed part. Naturalist religions don't die dramatically — they fail to bind, and the failure runs the same clinical course every single time, for 250 years: (1) Launch with press attention and real enthusiasm among educated people — Saint-Simonianism drew thousands to its lectures; Sunday Assembly's first events turned people away at the door. (2) Growth stops in the hundreds-to-low-thousands, because the product is the founder's talk-circuit energy, and attending is the entire practice. (3) The demands never arrive. Nothing to fear, nothing to hope for beyond the meetings themselves — so the group is competing, on pure quality of experience, against every concert, lecture series, and political club in town. It loses to the real versions: Comte's church was a worse opera, a worse university, and a worse charity than the actual ones his members already attended. (4) It never captures life's fixed points. Comte designed sacraments on paper; nobody needed his weddings or funerals, because nothing was at stake in them. (5) Recruits skew to educated, low-fertility skeptics who will not drill their children in a creed they themselves hold as a metaphor — so there is no second generation at all (Brahmo Samaj: from the crown of the Bengal Renaissance to 177 people in the 2001 census). (6) The founder dies, and with no living relationship to anything beyond him, there is nothing left to stay for; the remnant converts into a lecture society or a charity — which is a fine thing to be, but is not a religion. Death by evaporation, never by execution. And the causal claim survives its controls: it is not that designed religions were just small and odd — Comte's had a full machinery blueprint (liturgy, calendar, priesthood, sacraments) and still evaporated, while the two designed religions that kept a supernatural payload (Arya Samaj, Won Buddhism) are alive at scale. The variable isolates about as cleanly as history ever allows.

Succession — who inherits the founder's voice?

Every religion in the census eventually faces the same event: the founder dies. Melton's finding is that the death itself rarely kills — what decides the outcome is the succession architecture, which is almost always fixed (or fatally not fixed) years earlier. Sorting the census by succession type produces the cleanest predictor table in the whole study:

The design rules that fall out: the founder's office must be transferable; the transfer mechanism must be written down before it is needed (Baháʼu'lláh's covenant is the type specimen — succession fixed in a document while the founder lived, which is why Baháʼí passed through four transitions with only trivial splinters); legitimacy should rest on text and procedure rather than blood or private designation, because documents can be consulted but designations can be disputed; and if charisma must wind down, step it down gradually (Tenrikyo's twenty-year mouthpiece-bridge) rather than switching it off at a deathbed. Movements that did all four had, collectively, a near-zero fatality rate at succession. Movements that did none of them are the ✝ column.

Where does an indexed-corpus system sit?

Techno-Puritanism's own governance — doctrine living in a citation-stable, versioned, publicly indexed corpus, revised by stated procedure rather than by any person's office — doesn't map onto a single row above. It is a deliberate hybrid of the two best-behaved classes: text-as-successor (nothing to decapitate; the Risale-i Nur property that let leaderless reading cells shrug off ninety years of state pressure) crossed with elected/collegial's core insight (authority located in text-plus-procedure, so disputes are settled by consulting a document rather than fighting over a designation). The engineering intent is to keep the text class's unkillability while fixing its known failure mode — "no one empowered to adapt, discipline, or mobilize" — by making revision itself the written procedure. Is it unique? As a configuration, nearly; but it has one enormous precedent the census window is too young to contain: rabbinic Judaism, which after the year 70 moved authority out of a priesthood and into an indexed, argued, continuously-glossed corpus (Mishnah → Talmud → responsa) with interpretive procedure in place of prophet or throne — and rode that architecture through two millennia of statelessness and decapitation. That is the strongest long-run evidence anywhere for index-governance. The honest caveat travels with it: the rabbis' index did not work alone — it was embedded in congregations, high fertility, and a bounded interpreter class trained to apply the procedure. An index with no living body practiced at applying it drifts back into the text class's museum mode. The census's verdict on the configuration: best-in-class ancestry, genuinely untested as assembled.

The survivorship-bias test

The obvious objection: "you only see the winners." Here is the test that objection has to pass. If corpus-level foundings usually failed invisibly, the historical record should be full of famous, documented intellectuals who produced elaborate original metaphysical systems in writing that then went nowhere — no movement, no transmission. Famous people's manuscripts don't vanish; if the graveyard were large, we could name its occupants. So: name them. Across the entire 19th century, the hunt produces about a dozen —

Three things follow. First, the famous graveyard is small — a dozen names against a comparable-or-larger list of corpus-founders who got durable movements in the same century — and nearly every occupant had a prestigious secular container (academy, art world, occult orders) that absorbed the energy a church would otherwise have organized. Second, the objection survives one tier down: beneath fame there is a large movement-less population of complete written theologies — vanity-press systems, the clinical hypergraphia literature — the "audience cult" tier. A corpus alone reliably buys an audience, never a religion. Third, in every paired comparison the discriminating variable is organization, not corpus quality. Swedenborg had a printer-organizer; Fechner had a university; Fort refused. Which matches the sociology exactly: Melton — when new religions die it is from "lack of response of the public to the founder's ideas or the incompetence of the founder in organizing," and most that die do so in the first decade. Nobody in the scholarly literature has ever regressed survival on corpus-possession; this census is, as far as we can tell, the first attempt to measure it directly.

What were the founders actually doing?

Classify every founding by its most probable motivation — and state the criteria, so you can audit the judgment instead of taking it. The interesting category is the first one, so here is its test. We score a founding as a probable scam (>50%) when either or both hold: (i) the founder has a documented pre-founding history of running cons, or stated the intent to profit; (ii) the corpus is heavily plagiarized while its originality is denied. The concealment is the tell, not the borrowing — nearly every founder syncretizes and most say so (Cao Dai canonized Victor Hugo; Kardec cited his sources; Seicho-no-Ie openly fused Oomoto with New Thought). Concealed copying has no sincere explanation: if you believed you were transmitting truth from God, lying about where the words came from buys you nothing. Concealment only pays if the originality is the product being sold.

What motivation predicts: almost nothing

This is the section's real finding. The scam column contains the single biggest winner in the census (Mormonism) plus several stable survivors; the genuine-vision column contains both the most successful cluster (Tenrikyo, Baháʼí, Ahmadiyya, Hoa Hao) and a long death list (Oahspe, Koreshan, the Harmonial Philosophy, Southcott). Sincerity neither saves nor do scams kill. Second: exposure of constructed or fraudulent origins does not kill an established religion. Wicca's members learned the "ancient" origins were bogus and didn't care; Eckankar survived full documentation of its plagiarism and invented lineage; Mormonism survived the Book of Abraham papyri being translated. Adherents reframe. Third: the only motivation class that reliably fails is designed-with-the-supernatural-removed — the fatal variable rides on the metaphysics, not the motive. And a structural note: scam-founded systems tend to have thin canonical cores (Fard's memorized lessons) or product-corpora (Hubbard's 1,084 titles) — which makes them either easy to reinterpret wholesale (the Nation of Islam was converted to Sunni orthodoxy by its own heir in 1975) or collapse-prone; the scam-winners are the ones that built all four organs anyway.

When the founder is caught — exposure, scandal, and what actually kills

Three different things can be "proven" against a founder, and the census treats them very differently. It matters enough to separate them, because the intuition — surely a religion dies when its origin is shown to be fabricated? — is almost exactly wrong.

1 · Fabrication exposed: the stock survives; the funnel is poisoned. The record is unanimous. Wicca's members learned the "ancient witch-cult" was assembled from bogus sources — and didn't care. Eckankar survived scholarly documentation of its founder's plagiarism and invented lineage, partially conceded by his successor, plus a hostile schism, without losing the office. Theosophy outlived the 1885 Hodgson Report declaring Blavatsky a fraud by four decades of growth. Mormonism absorbed the translation of the Book of Abraham papyri (they turned out to be ordinary funerary texts) by building an apologetics industry. The mechanism is fifty years old in the literature: Festinger's When Prophecy Fails (1956) showed disconfirmation can increase commitment among the invested — sunk costs, community embeddedness, and identity protection all point the same direction, and the sophisticated movements manufacture reframes on demand. But watch what exposure does to recruitment: Scientology's active membership collapsed in every census after 1995 — the year its secret doctrine and litigation record hit the open internet — while its asset base kept growing. Exposure rarely kills the believers you have; it reliably kills the believers you would have gotten. Funnel, not stock.

2 · Moral scandal — sex or crime — sorts entirely by institutional maturity, not by severity. Where leader = institution, it is fatal: Benjamin Purnell's trial broke the House of David; Sandford's manslaughter conviction broke Shiloh; Noyes fled and Oneida became a flatware company; York's conviction erased the Nuwaubians; Asahara's crimes got the movement surveilled into a remnant and himself executed. Where the office was already institutionalized, it is grimly survivable: La Luz del Mundo kept the bulk of its membership through its living Apostle's 2022 child-abuse conviction (reframed as persecution); the Unification Church sailed through Moon's 1982 federal tax-fraud imprisonment; Soka Gakkai absorbed decades of scandal coverage. Same crime class, opposite outcomes — the variable is whether there is a machine that exists independently of the man.

3 · The timing law that unifies both. Every exposure effect is gated on when it lands relative to institutionalization. Caught in the first decade — the Melton filter window — fabrication or scandal is usually terminal, because there is only a funnel and no stock (Bocheonism's failed enthronement; Moorish Science's founder dying mid-murder-feud at the three-way-split stage). Caught after the machinery exists, the same revelation is a growth tax, not a death sentence. Which yields the design corollary this page's own existence embodies: a constructed religion that documents its own construction has no exposure event in its future. The Wicca–Eckankar law applies trivially when there is nothing to expose — transparency isn't just more honest than concealment; it is strictly safer, because it converts the one attack that poisons funnels permanently into a page the movement itself published first.

Sophisticated vs. silly — two axes, and a surprise

Two ratings run through the census table, and they are deliberately separate axes because they do not correlate. Sophistication = the intellectual quality of the system: internal coherence, engagement with the philosophy and science of its day, argumentative depth. Silliness = the laugh test: would a person hearing the founding stories for the first time start laughing? Mormonism proves the axes are independent — a genuinely worked-out materialist theology of eternal progression, dictated from golden plates read through a seer stone in a hat. So: does either axis predict success? Cross-tabbing all 57 core cases:

Reading the table honestly

Sophistication has a hard floor effect. Not one low-sophistication religion ever reached major scale, and two-thirds of them are dead — regardless of silliness. Sophistication is what survives contact with the movement's own educated second generation: it builds the clergy, the apologetics, the seminary. Below the floor, extra silliness just decorates the grave (the low-soph/high-silly cell — I AM, Unarius, CUT, the Nuwaubians — is the deadest corner of the census). Above the floor, the headline actually runs opposite to the intuition: among sophisticated systems, the major-success rate rises as silliness falls (17% → 26% → 57%), and the dignified-sophisticated cell contains zero deaths. But three things stop that from being the last word. First, a confound: the dignified winners all ride maximal host-frame continuity (two are Islam-frame, one rode a Gospel commentary into Catholic Brazil, one is stripped-down delta Buddhism) — dignity and frame-continuity travel together, and the family analysis above suggests the frame is doing the work. Second, silliness is astonishingly unpenalized — that is the real finding. The laugh test predicts mockery, and mockery predicts nothing: the silliest cell still contains the census's single biggest winner (Mormonism) and its fastest (Falun Gong). Third, look at what silliness does predict: commitment depth. The silly winners extract the deepest per-member commitment in the dataset — Mormonism's two-year missions and 10% tithing, Scientology's $1.5B from ~30k members, the Nation of Islam's discipline from a tiny base. Dignity buys breadth; silliness buys depth.

Methodological warning — the dignified cell is a censored sample. The table above covers the 57-case core census, which by construction contains only supernatural foundings; the designed-naturalist class was pulled out as a separate control group. But that class — Comte, the Brahmo Samaj, Ethical Culture, the Monist League — is almost entirely made of high-sophistication, low-silliness systems, and it runs 92% dead. Merge the two samples and the "sophisticated + dignified" cell roughly triples in size while its death rate jumps from 0% to over 60% — making it one of the worst cells on the board, not the best. So the honest unconditional statement is: low silliness is where the death zone lives, because a system dignified enough to pass the laugh test entirely is usually one small step from having nothing supernatural at stake at all. The dignified-sophisticated strategy is only safe conditional on keeping real supernatural stakes — Kardec's spirits, Baháʼu'lláh's Manifestations, Ahmad's prophethood all pass that bar; Comte's Grand Being does not. Which restores a real edge to the original intuition that silly religions win: a founding story weird enough to laugh at is, among other things, unfakeable proof that the religion actually claims something.

Why doesn't the laugh test kill? (the mechanisms)

Four converging literatures explain why believing something laughable is functional rather than fatal. (1) Belief as a costly signal (Iannaccone's "Why Strict Churches Are Strong," 1994): professing Kolob or Xenu is a badge no outsider will wear casually — it screens out free-riders exactly the way dietary laws and dress codes do, raising average commitment. Silliness is strictness applied to cognition. (2) Memorability (Boyer's "minimally counterintuitive concepts"): stories that violate a couple of intuitive expectations — a hollow earth, a scientist named Yakub, golden plates — are the most memorable and retellable narratives humans produce, and crucially they are child-tellable: the silly bits are usually the concrete bits, and concrete stories are the interface to generational transmission. Dignified abstractions ("the Grand Being of Humanity") do not survive retelling to an eight-year-old. (3) Credibility displays (Henrich's CREDs): watching someone sacrifice dignity for a claim is among the strongest available evidence that they sincerely believe it — a missionary defending golden plates at your door is performing sincerity in a way no one defending a platitude can. (4) The commune data (Sosis & Bressler 2003): 19th-century religious communes outlived secular ones several-fold, and costlier requirements extended lifespan — but only for the religious ones. Sacrifice, including the sacrifice of dignity, is only cashable when something supernatural is at stake — which ties the silliness result directly back to the naturalism result above. And why does the intuition say silly religions win? Salience bias: the dignified winners exit the reference class — Baháʼí reads as "world religion," Ahmadiyya as "Islam," Kardecism as "Brazilian culture" — so the examples anyone can name of "weird new religion that made it" are all from the silly column.

Why Techno-Puritanism is built silly on purpose

This finding is not one this tradition observes from a distance — it is load-bearing in how the tradition presents itself, and the census is why. The name itself, Techno-Puritan, is chosen to make you smirk; the tracts lean on the future police — agents of humanity's distant descendant acting backward across time — precisely because it is the kind of idea you can retell to a child (or laugh at in a bar) in one sentence, which is what Boyer's mechanism says a doctrine must survive; the imagery war between the two competing futures of man — the Federation (gay space communism) versus the Terran Empire (a deeply religious, expansionist star empire) — gives the cosmology a picturable, meme-native form the way golden plates and hollow earths did for the 19th century; and the tract videos splice in comedy clips from anime and cartoons rather than reaching for stained-glass gravitas. Every one of those choices costs dignity on contact, and that is the point: per the mechanisms above, the willingness to say something laughable is the unfakeable signal that something is actually being claimed (the censored-sample warning in this section is exactly the grave that pure dignity digs), the laughable bits are the memorable and transmissible bits, and professing them is the membership badge no casual outsider will wear. The architecture is the Mormon one deliberately: a silly surface a first-hearer can laugh at, riding on a systematic argued core that survives the second look — M·H by this page's own ratings on the surface, aiming for the sophistication floor's right side underneath.

Coherence — do the systems survive their own premises?

A different axis from sophistication: a system can be erudite and self-contradictory, or simple and airtight. Coherence here means: do the parts entail each other; does the system's own epistemology license its own claims; how does it handle the classic stress test (theodicy — why does evil exist?); and how many ad-hoc patches has it accreted under pressure. Selected majors only — running this on all 57 would be a book. Each row: the core metaphysic in one line, the system's crown jewel, and its load-bearing crack.

And the uncomfortable census finding about coherence: it does not predict success at all. The most coherent system in the table (ACIM) built no community; the least coherent (Scientology) built a billion-dollar one; Mormonism thrives with its external cracks on public display. What coherence actually buys shows up a generation later, in the seminary rather than the street: incoherent systems fragment doctrinally at succession (the Nation of Islam's thin, unsystematic core is what let its own heir reinterpret it out of existence in 1975), while coherent ones revise gracefully under pressure (Baháʼí and Kardecism absorb science updates without crisis). Coherence is not a recruiting tool. It is what the interpreter class needs to keep the index alive — which, for a tradition whose succession plan is the index, makes it load-bearing here in a way it was optional for everyone else.

Sex — doctrine, exploitation, and the pattern

The census contains two different sexual phenomena, and they sort in opposite directions. The variable that matters is not the presence of sexual innovation — it's whether the innovation points at reproduction or at the founder.

Sexual doctrine coupled to marriage and reproduction behaves as a growth engine. Mormon plural marriage multiplied households through the persecution years, and — the truly instructive part — was strategically abandoned under state pressure (1890) without breaking the church, because it was doctrine about family structure, not about the founder's person. The Marange Apostles' polygyny powers the fastest demographic growth in the contrast class (1–3M). The Unification Church's mass-arranged Blessing weddings were precisely a machine for converting a convert movement into a lineage movement. Sexual-liberation doctrine, by contrast, is a demographic dead end wearing a scandal bomb. Oneida's complex marriage came with male continence — birth suppression by design — and collapsed in scandal and internal revolt over access, leaving a silverware company. Kerista's polyfidelity died with the commune. Raëlism's sensual meditation pairs with explicit anti-natalism (~14k members). Thomas Lake Harris's "celestial counterparts" ended in exposé. The Agapemone's "spiritual brides" scandalized their way to extinction by 1956. The Family International is the terrible limit case: it had both extreme fertility and sexualized doctrine ("Law of Love"), and the abuse the latter licensed destroyed second-generation retention so completely that the children became the movement's dedicated destroyers. Free-love systems suppress or poison exactly the intergenerational transmission that wins centuries. (Osho/Rajneesh, a census gap worth flagging — enormous discourse corpus, sex-positive practice — followed the same arc: post-founder it converted into a meditation-resort brand rather than a lineage.)

Founder sexual access is the single most reliable scandal generator in the failure record — and whether it kills depends entirely on how institutionalized the movement already is. Where leader = institution, it is fatal or crippling: House of David (Purnell's statutory-rape trial broke it), Oneida (Noyes fled to Canada), the Agapemone, the Nuwaubians (York: 135 years), Aum, TFI. Where the office had already been institutionalized, it is horrifyingly survivable: La Luz del Mundo retained the bulk of its membership straight through its living Apostle's 2022 child-abuse conviction, reframing it as persecution — grim proof that machinery outlives morality. The census's overall verdict is boring and biblical: every configuration of sex that wins at the century scale is the pro-family one, and celibacy (the mirror-image doctrine) is a 1–3 generation death sentence. Sexual innovation in either direction — more access or none — loses to households.

Where new religions grow — the regional ecology

The language-market test — is "Europe founds more but stays small" just market size?

A plausible alternative to the regulation story: maybe US foundings get bigger simply because English is a ~1.5-billion-speaker market while European foundings are trapped in national-language ponds. That hypothesis makes a clean prediction — Spanish, with ~500 million speakers across 20+ countries, should produce English-scale winners. So we ran the census: the Spanish market fails the test decisively. Roughly fifteen Spanish-language corpus-bearing foundings catalogued — Samael Aun Weor's gnostic movement (60–70 books, the flagship Spanish corpus, now dozens of quarrelling acronym-groups totaling maybe tens of thousands), Trincado's Magnetic-Spiritual School, Escuela Basilio, Silo, Chile's Alfa y Omega (thousands of hand-drawn "telepathic scrolls," a movement so small it has no Wikipedia article) — and not one verifiably exceeded ~200k adherents. The only Spanish-language mega-successes carry no corpus at all: La Luz del Mundo (Bible-only, living Apostle) and the Santa Muerte devotion (10–12M devotees, zero organization). Meanwhile Portuguese — a market less than half Spanish's size — hosts Kardecism at 3.8M census plus Umbanda's millions.

The same-size comparison is even sharper. German ≈ Japanese in market size, and the outputs differ ~50-fold. German-language corpus giants: Jakob Lorber's 10,000-page channeled New Revelation → ~100k unorganized readers and a publishing house instead of a church; Steiner's ~350 volumes → 46k members (and 10,000 institutions); the Grail Movement → ~40k, now mostly in Nigeria. Japan, on the same-sized pond: Soka Gakkai at 2.5M+ real, Tenrikyo at 1.1M. The French record repeats it: Kardec's system died in its native 80-million-speaker market and became huge in Portuguese; Freytag's Geneva-founded Friends of Man grew in Italian; Raëlism's growth came in Japan and Korea; Antoinism, the one French-language healing religion to claim six figures (1920s Belgium), stayed francophone and died with its demographic. Tagalog — maybe 25–30M speakers at the founding — produced Iglesia ni Cristo, at 2.8M census bigger than every Spanish, German, and French corpus-founding combined. And the one German-founded body with nine million members, the New Apostolic Church, is 87% African.

Verdict: growth follows receptive markets, not the founder's language pool. Translation is cheap and insufficient (Lorber is available in 20+ languages; there is still no Lorber church), while every emigrant success story — Kardec→Brazil, Grail/NAC→Africa, Freytag→Italy, Raël→East Asia — tracks destination-market receptivity: open regulation, weak incumbent monopoly, crisis demand. Language-market size is real but enters only as the ceiling once penetration works: Falun Gong's zero-to-tens-of-millions in seven years required the largest single-language market on earth. Europe's signature shape — many foundings, all small — is a penetration failure, not a pond-size problem: state churches, sect commissioners, MIVILUDES, and the Nazi decapitation of the entire 1930s German cohort (Mazdaznan banned 1935, Anthroposophy 1935, Grail expropriated 1938) — plus a distinctly German-speaking habit of channeling charisma into a Verlag and institutions rather than congregations. Penetration = regulation × crisis × frame-continuity; absolute size = penetration × market. Both terms are real; only one of them is doing the Europe-vs-America work.

Does the parent family matter? (yes — almost as much as the state)

Classify every case by the religious family it grafts onto and the census quantifies what Stark could only propose: cultural continuity is close to destiny. Roughly 90% of the ★ outcomes sit in families that kept the host society's existing sacred frame and added new revelation on top of it. Invented frames pay for their novelty at exactly the rate the theory predicts.

Two corollaries. One-step hybrids scale; two-step exotics don't: Christianity+local substrate went big repeatedly (Kimbanguism, Cao Dai's Catholic-structured syncretism, Unificationism's Christianity+Korea), while frame-alien imports and invented frames cap at persistent-small. And the family predicts the failure mode: Islamic-frame dies by takfir, Chinese-syncretic by the state, esoteric by audience-drift, New Thought by its own falsifiable healing claims, designed-naturalist by naturalism. Note the inversion that matters for this tradition: the losing UFO family put a science frame around supernatural content; the winning move in every high-continuity family is the opposite — science-positive content riding inside the host culture's sacred frame.

The verdict — and what it means for this tradition

The thesis survives quantification. People who sit down and write out a complete new religious system succeed at roughly two orders of magnitude above the base rate for religious foundings — about 1-in-3 against 1-in-1,000 on Stark's bar. But the census also says why, and it is not the page count: the corpus is a marker, not an engine. Writing a complete system is the costliest possible signal of founder seriousness and systematizing capacity — and the founders who do it mostly also build the other three organs. Where they don't, the corpus alone buys an audience (Oahspe, the 1,000-volume Racional canon, 3,100 books of Happy Science) and the audience dies childless.

Read against this census, Techno-Puritanism is a deliberate bet on the untested quadrant: the method of the designed class (Comte's — religion as engineered artifact, sources cited, construction admitted) with the three organs every designed religion amputated deliberately installed — a supernatural-scale metaphysics whose falsifiable claims sit in the archaeological past (graded, misses kept), a natalist household engine, and a no-prophet, iterative-revelation structure that belongs to the succession class with the cleanest record in the data (written covenant / elected council / text-over-lineage: Baháʼí, Won Buddhism, Ahmadiyya). The census's one warning to any such project is the Wicca–Eckankar law run in reverse: concealment is pointless as well as corrosive — constructed origins, even exposed, don't kill a religion, so there is never a reason to hide the construction. This page is itself the compliance: the full failure record of the class we belong to, published by us, misses kept.

The scorecard — where this tradition sits in every category, and the odds

Having built the instrument, we point it at ourselves. Misses kept, as always:

Odds, from the census's own base rates. Three nested bets, stated so history can grade them. Persistence — surviving as a self-reproducing tradition two-plus generations at any scale: roughly 2 in 3. The base rate for corpus-level foundings is 65%, and TP's profile beats the base on nearly every conditioning variable; even machinery-less family-transmission traditions ran centuries (the Muggletonians took an indexed corpus and zero outreach 328 years). Major success — Stark's bar, 100,000 adherents held for a century: about 1 in 10 as configured today, because the missing fourth organ dominates every other advantage — no corpus-without-congregations case in two hundred and fifty years has ever crossed into mass membership; but rising toward the census's ~1-in-3, plausibly better, if congregational machinery and the interpreter class are built while the founders are alive — foundings holding all four organs inside the top family cluster succeed at rates no other configuration touches, and the Melton filter says the window for building them is the early decades, not later. World-religion scale (10M+): ~1 in 20 — three or four cases in the whole census got there, and every one of them was a four-organ machine within a generation of founding. The instrument's summary judgment on its own maker: the design is in the best quadrant history offers, the wager is legitimate, and the single variable that will decide it is not the corpus, the metaphysics, or the laugh test — it is whether the readers ever become congregations.

Full texts: the Techno-Puritan corpus · The Genesis Ledger · machine index at llms.txt.

The buildout — why the camps, the meetings, the Season, and the sending-out must actually happen

The scorecard above says one variable decides everything: the missing fourth organ. For a natalist, household-first tradition, that organ does not begin as Sunday pews — historically the congregation's irreplaceable, load-bearing function was always the marriage market and the youth-retention pipeline; the pews condensed around those, not the reverse. The network's planned institutions — marriage meetings, children's summer camps, an annual Season, and the sending-out of young people to live with other network families during dating — are therefore not auxiliary programs or nice-to-haves. Graded against the census, they are organ four itself, in household form. Here is what the precedents say about each, and how to run them at minimum cost:

The stakes, in this page's own numbers: these four institutions are where the odds move — the difference between the 1-in-10 of a corpus with readers and the ~1-in-3 of a four-organ founding in the best family cluster is exactly whether the children of the network marry, camp, host, and foster inside it. And Melton's first-decade filter sets the deadline: every case in the census that built its machinery did so while the founding generation was alive to staff it. The tracts are written; the ledger is public; the laugh test is passed. What history is waiting to grade is the Season's first calendar.