End-of-the-World Predictions: What Sacred Texts Actually Say — and What the Data Says

Every tradition carries an end-narrative, and almost every one ends in renewal. What Revelation, Ragnarök, the Kali Yuga and the Hour actually say, the record of dated doomsdays that passed, and what measurable civilizational stress reads in July 2026.

The scriptures refuse dates. The failed-date record belongs entirely to readers who overruled their own texts.

The oldest question in the search box

Typed into a search engine, the question reads as modern anxiety: when will the world end. It is in fact the oldest standing query on record — every civilization that left writing left an answer, and the answers agree with one another to a degree that deserves attention. This article does three things with that inheritance. First, it reads the four most consulted end-narratives — Revelation, the Hour, Ragnarök, the Kali Yuga — accurately, from what the texts actually say rather than from what panic culture says they say. Second, it identifies the structural feature nearly all of them share, which is not destruction. Third, it sets the complete record of dated doomsday predictions against the one thing a calibrated engine can honestly offer instead: measurable civilizational stress, with numbers that can be graded.

The universality is the first finding. Christianity carries the Apocalypse of John; Islam carries the Hour; Norse tradition carried Ragnarök; Hindu cosmology carries the running-down of the Kali Yuga. Beyond the four: Zoroastrianism awaits Frashokereti, literally the making wonderful of the world; Buddhist tradition describes the long decline of the dharma before the arrival of the future Buddha, Maitreya; Stoic philosophy taught ekpyrosis, the periodic conflagration and rebirth of the cosmos; Aztec cosmology counted our world as the fifth sun, four having ended before it. No major tradition lacks an end-narrative. The pattern is not a coincidence; it is a species-wide observation, encoded independently on every continent.

A note on posture before proceeding, because it governs every paragraph that follows. This observatory reads these texts with respect and without condescension — as pattern literature: the compressed, narrative memory of civilizations that watched systems end around them, in the fall of cities, dynasties, harvests and empires, and preserved the shape of what they saw. Grading scripture is not this observatory's business, and none of these traditions ever submitted itself for scoring. Dated claims made by later readers are a different matter entirely. Those entered the record voluntarily, with expiry dates attached, and the record has been keeping score.

What the four great end-narratives actually say

The Book of Revelation — in Greek, apokalypsis, an unveiling, not a destruction — was written near the end of the first century by an author who names himself John, in exile on the island of Patmos. Its machinery is famous: seven seals opened, seven trumpets sounded, seven bowls poured out; a beast; a final gathering of kings at Armageddon, the mound of Megiddo in northern Israel; a thousand-year reign; a last judgment before a great white throne. What is less famous is where the arc terminates — not in the burning world but two chapters past it, in a new heaven and a new earth, a city descending, and the line that closes the sequence: no more death, nor sorrow, nor crying. And the canon polices its own dating. Of that day and hour, the Gospel of Mark has Jesus say, no one knows — not the angels, not the Son.

Islamic eschatology is organized around the Hour, as-Sa'ah, and the Qur'an is unusually explicit about its scheduling: its knowledge rests with God alone. They ask you about the Hour, says surah 7:187 — say, its knowledge is only with my Lord; none but He shall disclose its time. The hadith literature describes what precedes it: lesser signs of moral and social disorder, then the great signs — the deceiver called the Dajjal, the descent of Jesus, the release of Gog and Magog, the sun rising from its setting place. Then the trumpet blast, the resurrection of every human being, and judgment. Note the destination, because it matters for the next section: not erasure but exchange. On that Day, says surah 14:48, the earth will be changed into another earth, and the heavens as well.

Ragnarök — the doom of the powers — survives chiefly in the Völuspá of the Poetic Edda and in Snorri Sturluson's thirteenth-century Prose Edda, and it is the most complete destruction in any major tradition: three winters with no summer between them, the wolf Fenrir devouring Odin, Thor and the world-serpent killing each other, the fire-giant Surtr burning the world, the earth sinking into the sea, the stars going out. And then the seeress who narrates the Völuspá keeps speaking. She sees the earth rise a second time out of the sea, green; fields grow unsown; the surviving gods gather where their hall once stood; Baldr comes back; two humans, Líf and Lífþrasir, emerge from shelter to repeople the world. The total apocalypse of the north is, by its own text, an intermission.

Hindu cosmology does not end the world at all; it rotates it. Time runs through four ages — Satya, Treta, Dvapara, Kali — each darker than the last, with dharma standing on four legs in the first age and on one leg in ours. The Kali Yuga, in the traditional reckoning, began in 3102 BCE and runs 432,000 years, at the end of which Vishnu's final avatar, Kalki, appears on a white horse to close the age — and the Satya Yuga, the age of truth, begins again, as it always has. The four ages together form one mahayuga; a thousand mahayugas make a single day of Brahma. Two facts follow for the modern reader: the largest timescales in any scripture belong to a tradition in which nothing is ever final, and on its own arithmetic the end of this age sits more than 400,000 years out. Hindu cosmology has scheduled no nearby apocalypse for anyone.

The commonality: destruction is almost never terminal

Set the four endings side by side and the pattern announces itself. Revelation ends in a renewed heaven and earth and a city with open gates. The Hour ends with this earth exchanged for another. Ragnarök ends with the sea giving the land back, green. The Kali Yuga ends with the wheel turning to the age of truth. Zoroastrianism names its ending the making wonderful. Even the Stoic conflagration was followed, in doctrine, by palingenesis — the rebirth of the whole. Across unrelated languages, climates and centuries, humanity's end-narratives converge on the same structure: the end is real, the end is severe, and the end is not the end. Destruction is a chapter boundary, not a back cover.

Read as pattern literature, this is a remarkable artifact — arguably humanity's oldest recorded intuition about how systems behave. The civilizations that wrote these texts watched annual cycles, dynastic cycles, the death and refounding of cities: ends that were locally total and globally transitional. They encoded the shape of that observation at the largest scale their cosmologies could hold. Mircea Eliade documented the breadth of the pattern in The Myth of the Eternal Return: archaic cultures overwhelmingly understood time as cyclical, catastrophe as purgative, and renewal as structural. The end-narrative, in other words, is cycle theory in its mythic form: systems end, and systems restart.

The intuition has since survived peer review. Strip the imagery and the same shape organizes the secular frameworks THE CORE runs on: Kondratieff's long economic waves, Ibn Khaldun's asabiyyah cycle of cohesion and decay, Perez's technology surges, the long debt cycle — each a formalization of the claim that complex systems saturate, break and reconstitute rather than terminate. The difference between the mythic form and the calibrated form is not the shape; it is the exposure. A scripture describes the turn. An engine is obligated to put a probability and a date on the next one, in writing, where it can be wrong. That obligation is the entire subject of the second half of this article.

The failed-dates record

The moment a reader converts an undated scripture into a dated claim, the claim leaves literature and enters the record — and the record is unsparing. The founding modern case is the Millerite movement. William Miller, a New England Baptist preacher, computed from the 2,300 days of Daniel 8:14 that Christ would return around 1843–44; the movement's final, precise date, set by the preacher Samuel Snow, was October 22, 1844. Tens of thousands of Americans prepared, some giving away property and leaving crops unharvested. The day passed. The movement itself named what followed the Great Disappointment — which remains the most honest label any failed prediction has ever received — and out of the sober reckoning that followed, Seventh-day Adventism eventually emerged.

The modern file repeats the shape at higher production values. In May 1910, spectroscopy found cyanogen in the tail of Halley's Comet, an alarmed reading of an astronomer's remark did the rest, and anti-comet pills sold briskly before Earth passed through the tail without incident. July 1999 — the one explicitly dated verse in Nostradamus's corpus, quatrain X.72 — passed without its King of Terror. The millennium arrived carrying apocalyptic freight on both religious and secular channels; the secular half, the Y2K bug, resolved quietly because engineers had spent years and several hundred billion dollars repairing it. In 2011, Harold Camping's Family Radio computed Judgment Day for May 21, advertised it on more than five thousand billboards, re-dated it to October 21 — and then did something rare: in March 2012 Camping publicly called the attempt at date-setting sinful, and stopped. And December 21, 2012 was the rollover of the 13th b'ak'tun of the Maya Long Count — an odometer turning, as Maya scholars and NASA both said in advance; Maya inscriptions themselves reference dates thousands of years past 2012.

The score, then, on dated terminal predictions: every one on the record has failed. The observation carries an unavoidable survivorship clause — a correct entry would have interrupted the study — but the sample is not small, the claimants were diverse in method and sincerity, and the failure mode is uniform. In every case the dater was more confident than the source text. Miller's certainty exceeded Daniel's; Camping's arithmetic exceeded the Gospel's explicit refusal of days and hours; the 2012 industry's certainty exceeded anything a Maya inscription actually says. The pattern is precise enough to state as a finding: the scriptures refuse dates, and the failed-date record belongs entirely to readers who overruled their own texts.

Why specific dates always fail

The mechanism is structural, and it is worth stating plainly because it will run again next year under a new date. Sacred end-narratives derive their durability from being undated: a text with no expiry can never be discredited, which is precisely why it survives twenty centuries. A dated claim works on the opposite economics — it borrows the ancient text's authority and attaches a modern deadline to it. This is the worst of both instruments. The confidence is inherited from a source that never stated it; the deadline is added by a reader who cannot back it; and when the date passes, the text's authority survives untouched while the reader absorbs the loss. Unfalsifiable scripture plus falsifiable date is not a synthesis. It is a category error with a countdown.

The arithmetic itself guarantees failure. Doomsday dates are derived by extracting numbers from texts — Daniel's 2,300 evenings and mornings, spans counted from the Flood, calendar cycles multiplied through one another — and such derivations have effectively unlimited degrees of freedom: with enough optional starting points, unit conversions and symbolic equivalences, any target date can be reached, which means the derivation carries no information about the target. This is also why failure never ends the enterprise. Leon Festinger's team documented the psychology in When Prophecy Fails, the 1956 study of a flying-saucer doomsday group: a failed date, publicly committed to, tends to intensify belief rather than extinguish it, because recalculation is cheaper than retraction. The Millerites recalculated; Camping re-dated within days; the mechanism is two centuries old and fully reproducible.

Underneath both problems sits the base-rate floor. Forecasting disciplines price events from reference classes — how often has this kind of thing happened, under these conditions. A terminal event has, by definition, an empty reference class: there is no frequency data on world-endings, and none can ever be collected by us. What does have a rich reference class is the prediction of world-endings, and that class resolves at zero for all recorded members. A calibrated engine therefore has exactly one honest move available, and it is not a cleverer date. It is to decline the terminal question as posed, and to answer the measurable question underneath it — which is the subject of the next section.

What a calibrated engine says instead

THE CORE — the convergence engine behind this observatory — seals probabilistic projections with four mandatory parts: an explicit probability, a hard window, an objective resolution criterion, and a written falsifiability condition, scored publicly by Brier's rule. Run the end of the world through that contract and it fails at the final step, and not for reasons of temperament. An end-of-the-world entry admits no scorer: its success is unverifiable by construction, and the scoring session it implies would be somewhat under-attended. The registry therefore carries no terminal-collapse projection at any probability — not because the engine is an optimist, but because the entry could never be graded, and this registry does not publish ungradable numbers.

What can be sealed is the stress underneath the anxiety, and it is sealed, component by component. The Historical Resonance Index scores July 2026 on five published inputs. Debt load reads 82: US federal debt near 100% of GDP, global debt at a record $348 trillion. Great-power rivalry reads 78: world military spending at $2.89 trillion in 2025, an eleventh consecutive annual rise. Internal polarization reads 84: US partisan antipathy at survey-record highs, trust in the federal government near two in ten. Tech-disruption velocity reads 88: the fastest technology adoption curve ever recorded. Monetary regime strain reads 72: central banks bought 863 tonnes of gold in 2025 after three straight years above 1,000. Against six past eras, the closest matches are 1968–74 at 83 and 1910–14 at 82, with 1929–33 at 79 and 1937–39 at 78 behind them.

Read that table honestly and it says two things the apocalypse genre cannot. First, the pressure is real and high — this is not a calm era, and the components are not reassuring. Second, the direction is undetermined: the two closest historical matches resolved in opposite directions, 1968–74 in inflation and institutional reform, 1910–14 in continental war. Resonance identifies the fork, not the branch. And the nearest thing to an end-input the registry prices at all is GEO-03 — battlefield use of a nuclear weapon by 2035 — sealed at 8%, ±5, ruin-class consequence, its falsifiability line published beside it. That is the data's full statement on collapse: stress measurable and elevated, tails priced without theater, apocalypse not scheduled.

The honest replacement for apocalypse anxiety

The search volume behind when will the world end is not foolishness; it is a demand signal from people reading the same stress the Resonance Index reads, and for centuries it has been met by two products — undated dread, which cannot fail, and dated panic, which always does. There is a third product now. On 3 July 2026 this observatory sealed a Calibration Ledger of ten predictions with locked probabilities, objective resolution criteria and hard resolve-by dates, every entry PENDING, first public Brier scoring in July 2027. It contains no doomsday. It contains the components of the anxiety, dated and priced: 76% that one of 2026–2029 sets a new global temperature record; 82% that US and Chinese forces exchange no live fire through 2029; 8% that the dollar falls below 45% of allocated reserves by 2030.

The difference in kind is the exposure. Each of those numbers can be wrong in public, on a date already fixed, by a criterion adversaries would accept, and the misses will be published at the same size as the hits. A reader tracking that ledger holds something no prophecy consumer has ever held: a falsifiable account of the near future that is obligated to confess its errors on schedule. Anxiety attaches poorly to arithmetic. The scored probability does to dread what the almanac once did to weather panic — it does not abolish the storm; it replaces the storm's theology with a number that improves as it is graded.

The traditions, read carefully, were right about two things all along. They were right about the shape — that ends are turns, that destruction precedes renewal, that systems restart; the calibrated frameworks that replaced them formalize the same curve. And they were right about humility: the texts themselves, from the Gospel of Mark to surah 7:187, refuse the date with a firmness their readers keep failing to match. The failure was never in the scriptures. It was in the daters. Read the old texts for the pattern, which they preserved better than anyone; read the registry for the probabilities, which they never claimed to hold; and come back in July 2027, when the first grades are published. Nothing here is advice. All of it is checkable, and that is the entire difference.

Frequently asked

Did the Bible predict the end of the world?

The Bible describes an end, refuses a date, and finishes in renewal. Revelation's sequence terminates in a new heaven and a new earth and the New Jerusalem (Revelation 21), and the canon explicitly bars scheduling: of that day and hour no one knows (Mark 13:32; Acts 1:7). Every dated Christian doomsday on record — 1844 most famously — came from readers overriding that refusal, and every one has passed.

What did Nostradamus say about the end of the world?

Less than his reputation suggests. He set no doomsday date for our era; the preface to Les Prophéties (1555) states that his prophecies extend to the year 3797 — a horizon, not an appointment. For the record, 3797 sits roughly 1,770 years beyond the longest resolve-by date in this registry, and no scoring session that far out is currently staffed. His one explicitly dated near-term verse — the seventh month of 1999 — passed without event.

Has any doomsday prediction ever come true?

No. Every dated end-of-the-world prediction on the historical record — October 1844, May 1910, July 1999, January 2000, May and October 2011, December 2012, among dozens of others — has failed. The observation carries a survivorship clause, since a success would have ended the record-keeping, but the recorded reference class resolves at zero for all members.

What does the data say about civilizational collapse risk?

That stress is measurable and elevated, and that collapse is not scheduled. The Resonance Index reads July 2026 at 72–88 across its five components — debt, rivalry, polarization, tech velocity, monetary strain — and its two closest historical matches, 1968–74 and 1910–14, resolved in opposite directions: reform and war. The registry's nearest ruin-class entry, battlefield nuclear use by 2035, is sealed at 8% ±5. No terminal-collapse probability is published, because such an entry could never be scored.

Sources

The Revelation of John (Apocalypse), c. 95 CE — Revelation 21:1–4, the new heaven, new earth and New Jerusalem
Gospel of Mark 13:32; Acts 1:7 — the explicit refusal of dates within the Christian canon
Völuspá, Poetic Edda (Codex Regius); Snorri Sturluson, Prose Edda (13th c.) — Ragnarök and the earth rising green from the sea
Mahabharata; Vishnu Purana; Bhagavata Purana — the yuga cycle, the Kalki avatar and the return of the Satya Yuga
Qur'an 7:187, 31:34, 14:48 — knowledge of the Hour reserved to God; the earth exchanged for another earth
Sahih Muslim, hadith of Hudhayfa — the major signs of the Hour
Mircea Eliade, The Myth of the Eternal Return, 1949
Leon Festinger, Henry Riecken & Stanley Schachter, When Prophecy Fails, 1956
Contemporary accounts of the Great Disappointment, October 22, 1844 (Millerite movement; Daniel 8:14 computation)
NASA, 'Beyond 2012: Why the World Didn't End' (December 2012); Maya Long Count scholarship on the 13th b'ak'tun
Family Radio / Harold Camping — public statement of March 2012 retracting date-setting
Nostradamus, preface to Les Prophéties, Lyon, 1555 — the horizon of 3797
SIPRI Military Expenditure Database (2026) — world military spending $2.89T in 2025, an 11th consecutive annual rise
IIF Global Debt Monitor — global debt at a record $348T at end-2025
Glenn W. Brier, "Verification of Forecasts Expressed in Terms of Probability", Monthly Weather Review, 1950