At 2:57 in the afternoon on March 30, 1981, a thoracic surgeon named Benjamin Aaron stood over an open chest in trauma bay five of George Washington University Hospital and went looking for a bullet he could not find. The patient on the table was seventy years old, hypotensive, and bleeding into his left pleural cavity from a wound that had collapsed a lung. Outside the operating room, a small group of aides and Secret Service agents waited with a knowledge most of them would not have said aloud. The man on the table had been elected in 1980. The last six presidents elected in a year ending in zero had all died in office. So had the one before them. Seven straight, going back to 1840. The question hanging in that corridor, unspoken and absurd and impossible to fully dismiss, was whether the eighth was about to make it eight.

He did not. Ronald Reagan survived, and in surviving he broke something that had held for one hundred and forty years. The thing he broke has a name, two of them actually, and the names are where the trouble starts. People call it the Curse of Tippecanoe. They call it Tecumseh’s Curse. Both names assert a cause: that a Shawnee leader or his brother, defeated by William Henry Harrison at a creek in Indiana Territory in 1811, laid a death sentence on Harrison and on every president elected at twenty-year intervals thereafter. The pattern those names describe is one of the most genuinely arresting coincidences in American political history. The explanation those names smuggle in is one of the most thoroughly unsupported.

This is a myth-bust with a complication built into its core, and the complication is the reason the article exists. Most debunkings can afford to be dismissive because the thing being debunked is flimsy. The Lincoln-Kennedy coincidences collapse the moment anyone checks the dates, which is why the point-by-point grading of the fifteen most-circulated Lincoln-Kennedy claims reads as a clean takedown. The Tippecanoe pattern does not collapse that way, because the pattern is real. Seven presidents elected in zero-ending years really did die in office across seven consecutive twenty-year cycles. No one fabricated the deaths. The fabrication, the part that fails every test of evidence and mechanism, is the word “curse.” Pulling those two things apart, the true pattern and the false cause, is the whole job here.

What the Claim Actually Says

The claim circulates in several forms, and the forms matter because the looser versions are designed to be unfalsifiable. The tightest and most defensible version goes like this: every man elected president of the United States in a year ending in the digit zero, from 1840 through 1960, died while holding the office. That is a precise, checkable statement. It names a rule of selection (election in a zero year), a rule of outcome (death in office), and a window (1840 to 1960). Within that window it is true seven times out of seven, with no exceptions and no near-misses that have to be argued away.

The looser version drops the window and says presidents elected in zero years are doomed, full stop, which forces believers into awkward contortions when Reagan lives and George W. Bush serves eight years and walks out healthy. The loosest version of all attaches the supernatural agent, holding that the Shawnee leader Tecumseh, or more often his brother Tenskwatawa, called The Prophet, pronounced a curse after the 1811 defeat at Tippecanoe, and that the curse is the active cause of every subsequent death. This is the version that gives the phenomenon its names, and it is the version with the least evidence behind it. The progression from tight to loose is not accidental. A pattern that holds seven for seven invites a cause, the human mind supplies one, and the supplied cause then gets projected backward as if it had been there all along.

Grading the claim therefore means grading three separate things, and they earn three separate verdicts. The historical pattern, scored on whether the seven deaths happened as described, is true. The supernatural origin story, scored on whether any contemporary source records a curse pronounced at or after Tippecanoe, is false. The causal claim, scored on whether a curse offers any mechanism by which a duel pistol, a tubercular pneumonia, a cerebral hemorrhage, and three assassins’ bullets fired across one hundred and twenty years could share a single cause, is false and not even coherent. The interesting work happens in the space between the first verdict and the other two, because that space is where most people get lost. They notice the pattern is real, they reason that real patterns have real causes, and they accept the only cause anyone offered them. The error is not in seeing the pattern. The error is in assuming the pattern needs the kind of cause the curse pretends to be.

The Seven Cases, Walked One at a Time

A pattern is only as strong as its individual members, so the honest way to assess it is to take each of the seven cases on its own terms, with the documented cause of death and the documented circumstances. William A. DeGregorio’s reference compilation, The Complete Book of U.S. Presidents, supplies the spine of dates and medical particulars for the run, and the contemporary record fills in the rest. What emerges from walking the seven is a striking uniformity of outcome produced by a striking variety of means. Seven men died in office. They died of at least four medically and circumstantially unrelated things.

William Henry Harrison, Elected 1840, Died April 4, 1841

Harrison is the founding case, and his death is the one most often told as eerie because it was so fast. Elected in November 1840 after the most theatrical campaign the young republic had yet produced, the log-cabin-and-hard-cider spectacle that Paul F. Boller Jr. dissects at length in Presidential Campaigns, Harrison took the oath on March 4, 1841, in raw cold, delivered the longest inaugural address in American history without an overcoat, and was dead thirty-one days later. The standard account blames the speech for the pneumonia, and generations of schoolchildren learned that an old general talked himself to death. Modern reconstructions complicate that tidy story. A 2014 analysis in the medical literature argued that enteric fever from contaminated Washington water, the same sewage-tainted supply implicated in other White House deaths of the era, fits Harrison’s recorded symptoms better than a chill caught at the podium. Whatever the precise pathogen, the relevant fact for the pattern is plain: Harrison died of disease, in bed, of natural causes, thirty-one days into a term. No bullet, no assassin, no human hand. The first link in the chain is a man in his late sixties succumbing to an infection in an era before antibiotics, sanitation, or any reliable understanding of germ theory.

The aftermath deserves a moment, because it set a constitutional precedent that mattered far more than any superstition. Harrison’s death created the first presidential vacancy in American history, and the Constitution’s language about whether the vice president became president or merely acted as president was ambiguous enough to invite a power struggle. John Tyler resolved the ambiguity by sheer assertion, refusing to open mail addressed to “Acting President,” insisting on the full title and full powers, and establishing the precedent that would govern every subsequent succession. That precedent, born from the first link in the so-called curse, shaped the office for the following century and is the genuinely consequential legacy of Harrison’s brief tenure. The contemporary press treated the death as a national tragedy and a medical misfortune, not as a supernatural event. No newspaper of April 1841 reached for a Shawnee curse to explain why an aging general had caught a fatal illness in a Washington spring, because the curse story did not yet exist. It would be invented to explain a sequence that, in 1841, was not yet a sequence at all.

Abraham Lincoln, Elected 1860 and Reelected 1864, Died April 14, 1865

Lincoln is the second link, and his case is the first that involves violence. Elected in 1860 on a fractured field, reelected in 1864 as the Civil War ground toward its end, Lincoln was shot by John Wilkes Booth at Ford’s Theatre on the evening of April 14, 1865, five days after Lee surrendered at Appomattox, and died the following morning. The cause here is political assassination by a Confederate sympathizer acting as part of a wider conspiracy to decapitate the Union government. Booth’s motive was the lost cause of the Confederacy and a personal hatred of Lincoln sharpened by the president’s gestures toward Black suffrage. There is no plausible thread connecting Booth’s derringer to a chill at Harrison’s inauguration twenty years earlier except the calendar. Lincoln won his first election in a zero year. That is the entire link. The means of death, an actor’s pistol fired in revenge for a war’s outcome, has nothing in common with Harrison’s pneumonia beyond the fact that both men were presidents who died in office.

The Lincoln case also illustrates how the qualifying election gets chosen to suit the legend. Lincoln won the presidency in 1860, a zero year, and won reelection in 1864, a non-zero year. He was assassinated in 1865, during the term he had been elected to in 1864. By the strictest construction of the rule, the death falls in a term begun by a non-zero election, exactly the awkwardness that recurs with Franklin Roosevelt later in the sequence. Believers resolve it by treating the 1860 election as the binding one and the second term as a continuation, a defensible reading but a chosen one, and the choosing is the point. The contemporary nation that mourned Lincoln did not mourn a curse. The grief and rage of April 1865 fixed on Booth, on the Confederacy, on the conspiracy that also targeted the secretary of state and the vice president that same night. A bullet fired by a known assassin with a known motive in a known theater is the least mysterious death imaginable. Its later absorption into a supernatural pattern says nothing about 1865 and everything about the twentieth-century appetite for hidden design.

James A. Garfield, Elected 1880, Died September 19, 1881

Garfield extends the run to three, and his death belongs as much to the history of medicine as to the history of assassination. Shot on July 2, 1881, by Charles Guiteau, a disappointed and almost certainly delusional office-seeker who believed God had commanded the killing, Garfield did not die of the bullet. He died eleven weeks later of the massive infection his doctors introduced while probing the wound repeatedly with unsterilized fingers and instruments, searching for a projectile lodged harmlessly in his back. The bullet itself was survivable. The medical care of 1881, administered by physicians who scoffed at Joseph Lister’s theories of antisepsis, was not. Garfield wasted away over the summer, was moved to the New Jersey shore in a desperate hope that sea air would help, and died on September 19. The cause of this third link is a combination of an assassin’s bullet and lethally incompetent medicine, a circumstance specific to its decade and shared with none of the other six.

The medical particulars of Garfield’s eleven-week ordeal are worth dwelling on, because they preview the contrast that would later break the entire sequence. The bullet that struck Garfield missed his spine and major organs and lodged in tissue where, left alone, it would have caused him no lasting harm. Doctors of 1881, led by a physician who literally went by the name Doctor Willard Bliss, refused to accept Joseph Lister’s published evidence that invisible organisms caused wound infection. They probed the entry wound repeatedly with unwashed fingers and unsterilized metal instruments, hunting for the bullet, and in doing so they seeded the massive infection that killed him. Alexander Graham Bell was summoned to try a primitive metal detector to locate the projectile, an effort confounded by the metal springs in Garfield’s mattress. The president lingered through a brutal Washington summer, lost more than half his body weight, and was carried by a specially built railway spur to a cottage on the New Jersey shore in a final attempt to ease his decline. He died on September 19, 1881. The lesson encoded in that death is precise and it will matter at the end of this story: a survivable wound becomes fatal when the medicine of the era cannot manage it. Change the medicine and you change the outcome, which is exactly what happened a century later.

William McKinley, Elected 1896 and Reelected 1900, Died September 14, 1901

McKinley is the fourth link, and the only one whose qualifying election was the zero-year reelection rather than a first win. He took the presidency in 1896, won a second term in 1900, and was shot on September 6, 1901, by Leon Czolgosz, an anarchist who approached the president in a receiving line at the Pan-American Exposition in Buffalo with a revolver concealed under a handkerchief. McKinley initially seemed to recover, then declined as gangrene spread along the path of the wound, and died on September 14. The cause was political assassination driven by anarchist ideology, a movement that had already killed a Russian tsar, an Austrian empress, an Italian king, and a French president in the preceding two decades. Czolgosz had no grievance against McKinley personally and no connection to Tippecanoe, Indians, or any of the previous three deaths. He was a symptom of a transnational political violence that targeted heads of state across the industrialized world. McKinley happened to have been reelected in a year ending in zero.

The international dimension of McKinley’s death is the detail the curse narrative cannot accommodate, because it points at a real cause that has nothing to do with Tippecanoe. The wave of anarchist assassinations that swept the Western world between roughly 1881 and 1901 followed a coherent ideological program known as propaganda of the deed, the conviction that spectacular acts of violence against rulers would catalyze revolution. That doctrine killed Tsar Alexander II of Russia in 1881, President Sadi Carnot of France in 1894, Empress Elisabeth of Austria in 1898, and King Umberto I of Italy in 1900, the last of which Czolgosz cited as his direct inspiration. He carried a newspaper clipping about the Umberto killing. McKinley’s murder was the American chapter of a global story driven by industrial dislocation, immigration, and radical politics, and it belongs to that story far more than to any sequence of American zero-year elections. The clustering of such violence in periods of economic and social strain is itself a documented and mechanism-bearing pattern, the kind explored in the study of how assassination attempts cluster around economic panics. McKinley died because he was a head of state during the high tide of anarchist terror, not because he had won a particular election year.

Warren G. Harding, Elected 1920, Died August 2, 1923

Harding breaks the assassination sub-pattern and returns the run to natural death. Elected in 1920 in a landslide promising a return to normalcy after the upheavals of war and pandemic, Harding died suddenly on August 2, 1923, in a San Francisco hotel room during a western speaking tour, almost certainly of a heart attack or a stroke. He had been in visibly declining health, exhausted and unwell, and the contemporary diagnosis of a cerebral hemorrhage or coronary event fits the evidence. His wife Florence declined to permit an autopsy, which fed decades of lurid speculation that she had poisoned him to spare him the unfolding scandals of his administration, but no credible evidence supports the poisoning theory, and the simplest explanation, cardiovascular failure in an overweight, overstressed man of fifty-seven, accounts fully for the death. The fifth link, then, is natural death from heart disease, sharing a category with Harrison’s infection only in the broadest sense that neither involved a weapon.

The Harding case carries its own attached myth, which is instructive because it shows the same machinery of suspicion operating on a smaller scale. Because Florence Harding refused to permit an autopsy, and because the Teapot Dome and Veterans Bureau scandals were beginning to surface around the administration as Harding died, a sensational 1930 book by a former Bureau of Investigation agent named Gaston Means alleged that the first lady had poisoned her husband to spare him disgrace. The poisoning theory had no evidence behind it, was advanced by a serial fraud later convicted of swindling, and has been rejected by every serious historian who has examined Harding’s death. Yet it persisted for decades for the same reason the larger curse persists: a sudden death plus a whiff of scandal plus an absent autopsy creates a vacuum that lurid explanation rushes to fill. The mundane truth, that an exhausted, overweight man of fifty-seven with untreated hypertension and a failing heart collapsed during a grueling cross-country tour, satisfies the evidence completely and the appetite for mystery not at all. Harding’s death is doubly useful for understanding the Tippecanoe legend, because it shows both a real natural death wrongly absorbed into the supernatural sequence and a separate poisoning fantasy spun from the same human refusal to accept that important people sometimes die of ordinary causes at inconvenient times.

Franklin D. Roosevelt, Elected 1932 and Three Times After, Died April 12, 1945

Roosevelt is the sixth link, and his inclusion required the pattern’s believers to perform a small act of selection that reveals how the rule actually works. FDR was elected four times, in 1932, 1936, 1940, and 1944. Only one of those, 1940, ends in zero. He died on April 12, 1945, of a cerebral hemorrhage at Warm Springs, Georgia, during his fourth term, his blood pressure having climbed for years to levels that modern medicine would treat as a screaming emergency and that 1945 medicine largely shrugged at. The death was natural, the product of untreated malignant hypertension and the accumulated strain of leading a nation through depression and world war while concealing the gravity of his own decline. For the pattern, the relevant election is 1940, the only zero year among his four. That FDR’s qualifying win came in the middle of his unprecedented run, and that he then survived more than four years before dying in a term he was elected to in 1944, illustrates how flexibly the rule gets applied. The believer counts the 1940 election as the trigger and the 1945 death as the payoff, quietly ignoring that the death occurred in the term following a non-zero election.

The medical story behind FDR’s death reinforces how thoroughly natural and how thoroughly era-bound it was. Roosevelt’s blood pressure had been climbing for years into ranges that would now trigger immediate aggressive intervention. Readings well above two hundred over one hundred were recorded in his final years, and the cardiologist Howard Bruenn, brought in quietly in 1944, documented a patient in advancing congestive heart failure and malignant hypertension. The pharmacological tools that today routinely control blood pressure simply did not exist in 1945; the available treatments were rest, digitalis, and a low-salt diet, none of which could reverse the underlying disease. Roosevelt’s decline was concealed from the public during the 1944 campaign, his gaunt appearance attributed to overwork rather than to the vascular catastrophe building inside him. When the hemorrhage came at Warm Springs on April 12, 1945, it was the predictable terminal event of an untreated chronic condition, not a bolt from a Shawnee beyond. That his qualifying zero-year election fell in 1940, in the middle of an unprecedented four-term run, and that he then lived more than four years past it before dying in a term won in 1944, makes him the clearest illustration in the whole sequence of how the rule is applied retroactively and selectively to whatever the record happens to provide.

John F. Kennedy, Elected 1960, Died November 22, 1963

Kennedy is the seventh and final link of the historical run, and his death is the one that fixed the pattern in modern memory and made the curse a household phrase rather than a magazine curiosity. Elected in 1960, Kennedy was shot in Dallas on November 22, 1963, by Lee Harvey Oswald, dying within an hour. The cause was assassination, the third such in the run after Lincoln, Garfield, and McKinley, by a man whose motives remain debated but whose act is not. With Kennedy, the count reached seven for seven across seven consecutive twenty-year cycles, and the coincidence acquired the weight that made it impossible for the public to ignore. The eerie symmetry of Lincoln and Kennedy, two zero-year-elected presidents shot in the head a century apart, generated its own separate body of folklore, but the Tippecanoe pattern is the larger frame into which that smaller coincidence fits.

Kennedy’s death is also the moment the legend fully entered mainstream consciousness rather than remaining a magazine oddity. By November 1963 the sequence stretched back through six predecessors, and the press and public, already primed by decades of Ripley’s-style retellings, immediately recognized the seventh fulfillment. The shock of Dallas, broadcast and rebroadcast to a television audience of unprecedented size, fused with the older legend to produce a cultural certainty that something uncanny governed zero-year presidents. The separate Lincoln-Kennedy parallels, the supposed coincidences of names, dates, secretaries, and successors, exploded into circulation in the months after the assassination, carried in newspaper columns and chain letters and eventually in the 1964 Republican congressional newsletter that gave the list its mainstream provenance. Those parallels, graded one by one in the point-by-point examination of the fifteen most-circulated Lincoln-Kennedy claims, turn out to be a mix of the true, the trivial, the exaggerated, and the outright fabricated. But the Tippecanoe sequence into which Kennedy’s death slotted was different in kind, because its members really had all died in office, and that authenticity is what gave the larger legend its grip. The seventh case did not need exaggeration. It needed only to happen, and it did.

The Variety Hidden Inside the Uniformity

Set the seven side by side and the pattern’s true character comes into focus. The outcomes are identical: death in office. The mechanisms are wildly disparate. Two of the seven, Harrison and Harding, died of natural disease. One, FDR, died of a chronic vascular condition that medicine of the day could not manage. Four, Lincoln, Garfield, McKinley, and Kennedy, were shot, and even within that group the killers acted from incompatible motives spread across a century: Confederate revenge, deranged office-seeking, anarchist ideology, and a lone gunman’s still-disputed reasons. The bullets came from four different decades and four different ideological universes. The diseases came from infection, from heart failure, from untreated hypertension. To bundle these seven deaths under a single causal agent is to claim that one force reached across one hundred and twenty years to arrange a Booth here, a pneumonia there, a hemorrhage somewhere else, all timed to land on men who shared nothing but the last digit of an election year.

That is the conceptual problem the curse cannot solve, and it is worth stating plainly before turning to the question of where the curse story even came from. A real cause produces effects through a mechanism. The bullet enters tissue, the infection multiplies, the artery ruptures. A curse, to be a cause rather than a label, would need a mechanism too, and no version of the Tippecanoe story has ever supplied one. The believer is left asserting that the supernatural agent operated through whatever means happened to be available in each era, pneumonia in 1841, an assassin in 1865, sepsis in 1881, and so on, which is not a mechanism but a description of the very outcomes the curse was invoked to explain. The reasoning runs in a circle: the deaths prove the curse, and the curse explains the deaths.

Where the Curse Story Came From

If the curse were a genuine 1811 pronouncement, the historical record would contain some trace of it close to the event. A defeated leader’s dramatic prophecy against a victorious general would be exactly the sort of thing nineteenth-century newspapers, memoirists, and campaign biographers seized on, especially during Harrison’s 1840 run, when his Tippecanoe credential was the centerpiece of his entire political identity. The campaign that gave us “Tippecanoe and Tyler Too” mined every ounce of drama from the 1811 battle. A curse laid by Tecumseh or The Prophet would have been a gift to Harrison’s opponents and a sensation in the penny press. It appears nowhere. Not in the 1840 campaign material, not in the obituaries of 1841, not in any source contemporary with the events it supposedly describes.

The historian who has done the most to dismantle the origin story is R. David Edmunds, whose Tecumseh and the Quest for Indian Leadership and companion study of The Prophet, Tenskwatawa, are the standard scholarly treatments of the two Shawnee brothers. Edmunds reconstructs the actual events of 1811 in detail, and the actual events leave no room for the legend. Tecumseh was not at the Battle of Tippecanoe. He was hundreds of miles to the south, recruiting among the Creeks and other nations for his pan-Indian confederacy, when Harrison marched on the settlement at Prophetstown in his absence. The brother who was present, Tenskwatawa, had assured his followers that his spiritual power would render them invulnerable to American bullets, and the warriors’ confidence in that promise contributed to the disastrous attack that scattered the confederacy and gutted The Prophet’s authority. Tenskwatawa emerged from Tippecanoe discredited, his claims to spiritual potency shattered by the very defeat the curse legend later credits him with avenging through supernatural means. A man whose prophetic reputation was destroyed because his promised protection failed is a peculiar candidate for the architect of a curse that worked flawlessly for a century and a quarter.

The actual history of 1811 is worth recovering in some detail, both because it is genuinely dramatic and because the drama of the real events is what makes the invented curse unnecessary. Tecumseh had spent years assembling a confederacy of Native nations across the Old Northwest and into the South, a political and military project of remarkable ambition aimed at halting American expansion onto Indian land by uniting peoples who had long acted separately. His brother Tenskwatawa supplied the spiritual dimension, preaching a religious revival that rejected American goods, alcohol, and assimilation, and the two together built Prophetstown, a multinational settlement near the confluence of the Tippecanoe and Wabash rivers, into the center of the movement. William Henry Harrison, governor of Indiana Territory, regarded the confederacy as an existential threat to the land cessions he had been aggressively negotiating, and he moved against Prophetstown in November 1811 while Tecumseh was away. Edmunds reconstructs what followed: Tenskwatawa, against his brother’s explicit instructions to avoid battle in his absence, was provoked into a pre-dawn attack on Harrison’s encamped force, having assured his warriors that his medicine would turn the soldiers’ bullets to soft mud and leave the attackers unharmed. The warriors fought hard, but the promised invulnerability proved illusory, casualties mounted, and the attackers withdrew. Harrison burned the abandoned town. The defeat did not destroy Tecumseh’s confederacy outright, but it badly damaged The Prophet’s credibility, since his central claim, supernatural protection, had failed in the most public and lethal way possible.

Tecumseh himself died two years later, in October 1813, fighting alongside the British at the Battle of the Thames in the War of 1812, where Harrison again commanded the American force. So the man most often named as the curse’s author was killed in battle by forces under the very general the curse supposedly targeted, and he left no recorded deathbed malediction against the presidency, an office Harrison would not win for another twenty-seven years. Edmunds, whose scholarship on both brothers is the standard against which popular claims must be measured, finds nothing in the documentary record, Native or American, contemporary or near-contemporary, to support the notion that either brother pronounced a curse on Harrison or on future presidents. The legend asks us to believe that a curse uttered by a discredited prophet, or by a leader who was absent from the battle and dead within two years, reached forward across more than a century to arrange assassinations and illnesses, while leaving no trace in any source until twentieth-century popularizers supplied one. That is not history. It is storytelling dressed in history’s clothes.

So where did the story originate, if not in 1811? The evidence points to a retrospective construction assembled in the twentieth century, decades after the pattern had begun to accumulate, and most likely crystallized into popular form through Ripley’s Believe It or Not, the syndicated newspaper feature and book franchise that specialized in exactly this kind of arresting, dubious, mass-market curiosity. The timing is the tell. The pattern only becomes noticeable once it has piled up several cases. After Harrison in 1841 there was nothing to notice. After Lincoln in 1865 there were two, separated by a span that might invite a glance but not a theory. Garfield in 1881 made three, McKinley in 1901 made four, Harding in 1923 made five, and only somewhere in this stretch does the twenty-year cadence become visible as a cadence. The curse narrative attaches the Tecumseh origin to a pattern that was identified long after Tecumseh was dead, reaching back to find a villain dramatic enough to carry the weight of the coincidence.

Ripley’s appears to be the vehicle that broadcast the story to a mass audience, and the most consequential moment came with FDR’s election in 1940. That election supplied the sixth zero year in the sequence and set up the prediction the feature could dangle before readers: the man elected in 1940 would, if the pattern held, die in office. When Roosevelt did die in 1945, the pattern had not merely continued, it had fulfilled a publicized forecast, which is the kind of confirmation that turns a curiosity into a conviction. Boller, whose Presidential Anecdotes catalogs the folklore of the office with a careful eye, treats the curse as exactly that, folklore, recounting the legend without endorsing it and noting its modern, popular-press character rather than any documentary root in 1811. The distinction Boller draws is the right one. The curse is real as a cultural artifact, a story people told and retold and believed. It is not real as history, because the event it claims as its origin produced no such curse and the pattern it claims to govern was assembled backward from outcomes.

How the Legend Traveled

The transmission history of the legend matters because it tracks the accumulation of cases rather than any 1811 origin, which is itself strong evidence that the story grew out of the deaths rather than predicting them. In the nineteenth century, with only Harrison and then Lincoln and Garfield and McKinley to point to, no widely circulated curse narrative existed. The deaths were mourned individually as the misfortunes they were. The twenty-year cadence required at least five or six cases before it could be perceived as a cadence at all, which places its earliest plausible articulation in the 1920s or later, well over a century after the battle it invokes. Ripley’s Believe It or Not, founded as a newspaper feature in 1918 and exploding into books, radio, and a media empire through the 1920s and 1930s, was the ideal vector: its entire business was the arresting, half-credible curiosity presented as astonishing fact, and a death sequence spanning a century was exactly its kind of material.

Franklin Roosevelt’s 1940 election supplied the legend with something rare and powerful, a forward-looking prediction. With five prior cases established, the sequence implied that the 1940 winner would die in office, and that implication could be dangled before readers as a forecast rather than a retrospective curiosity. When Roosevelt died in April 1945, the forecast appeared to have come true, and a confirmed prediction is psychologically far more compelling than a pattern noticed after the fact. Kennedy’s election in 1960 set up the same dynamic for a new generation, and his assassination in 1963, witnessed on live television by a nation, fulfilled the prediction a second time and burned the legend permanently into the culture. By 1980, when Reagan won, the curse had become common knowledge, discussed in newspapers and dinner-table conversations as a genuine question hanging over the new president. Reagan’s near-death in 1981 was therefore experienced by many as the curse striking on schedule, and only his survival, and then Bush’s two terms after 2000, drained the legend of its predictive menace, though not of its cultural persistence.

The presidency attracts numerological superstition more readily than most institutions, and the zero-year sequence is only the most successful example of a broader genre. Observers have at various times claimed significance in the recurrence of certain birth states, in the alleged tendency of left-handed presidents to cluster, in supposed patterns among presidents who were the eldest or youngest of their siblings, and in various coincidences of names and ages and military backgrounds. None of these has achieved the cultural durability of the Tippecanoe curse, for a simple reason: the others lack a tragic outcome and a ready-made villain. A pattern about handedness or sibling order is merely a curiosity. A pattern about death, complete with a wronged Native leader striking from the grave, is a story with stakes, a moral shape, and an antagonist, and stories with those features survive while bare curiosities fade. The success of the zero-year legend over its numerological cousins is a lesson in what makes a superstition stick: not the strength of the underlying coincidence, which is comparable across many such claims, but the narrative satisfaction of the explanation attached to it.

The Statistics: How Strange Is Seven for Seven?

Granting that the curse has no documentary origin and no mechanism, a fair-minded skeptic still has to face the pattern itself, because dismissing it as “just a coincidence” without doing the arithmetic is its own kind of intellectual laziness. Seven for seven across seven cycles sounds astronomically unlikely. Is it? The honest answer requires a back-of-the-envelope calculation and, more importantly, an understanding of why the calculation does not mean what the believer thinks it means.

Start with the base rate. Across the full sweep of American presidential history, the probability that any given president dies during his term in office, from any cause, is not small. Of the forty-odd men who have held the office, a substantial share died in office or came close, and the historical in-office mortality rate for a presidential term, counting all causes across the eras when medicine was primitive and security was lax, runs in the rough neighborhood of thirty percent if one is generous about the definition and the period. That figure is itself a product of the dangerous, unhealthy, pre-antibiotic, pre-Secret-Service-professionalization conditions under which most nineteenth and early twentieth century presidents served, and it would be far lower for the modern era. But take thirty percent as a working number for the relevant period, treat each zero-year president’s survival as roughly independent of the others, and the probability of seven consecutive in-office deaths is approximately 0.30 raised to the seventh power. That works out to about 0.0002, or roughly one chance in forty-five hundred.

One in forty-five hundred is small. It is not vanishingly small, and the difference between small and vanishing is the entire point. Events with probabilities around one in a few thousand happen constantly, because the world runs an enormous number of trials. A specific person being dealt a particular five-card poker hand has a probability far smaller than one in forty-five hundred, yet someone is dealt some specific hand every time the cards come out, and no one calls it a curse. The mistake the curse makes is the mistake of the retrospective coincidence: it calculates the odds of a specific pattern after the pattern has already been singled out from a vast field of patterns that could have been singled out instead.

The Retrospective Pattern Trap

This is the deepest point in the whole analysis, and it is the one Martin Gardner spent a career hammering home in books like Fads and Fallacies in the Name of Science and in his long campaign against numerological pattern-mongering. The human mind is a pattern-detection engine of extraordinary power and almost no discipline. It will find structure in noise, and once it finds a structure it will compute the odds of that exact structure as if the search had been specified in advance. It was not. The Tippecanoe pattern was not predicted in 1840 and then confirmed. It was noticed sometime in the twentieth century, after enough deaths had accumulated, and then the odds of that particular accumulation were calculated as though someone had called the shot before the first inauguration.

Consider how many other patterns could have been the one we marvel at. Suppose the deaths had clustered in even years rather than zero years, or in years divisible by some other number, or among presidents whose surnames began with certain letters, or among those who were tall, or those who were generals, or those from certain states. Any of these, had it happened to line up, would have generated its own legend with its own retrofitted cause. The space of possible eerie patterns is gigantic. With a few dozen presidents and a handful of in-office deaths, the chance that some striking pattern emerges somewhere in that space is not one in forty-five hundred. It is close to certain. The zero-year pattern is the one that happened to land, and having landed, it gets treated as if it were the only pattern that could have been searched for. The probability that matters is not “what were the odds of this exact run,” but “what were the odds that some equally striking run would appear somewhere,” and that second probability is high enough to make the pattern unremarkable as a matter of statistics, however remarkable it feels as a matter of psychology.

The selection problem compounds further when the rule itself is allowed to flex. FDR was elected four times; the believer keeps the one zero-year election and discards the three others. Lincoln and McKinley were each elected twice; for them the believer is willing to count either election as qualifying, since both happened to win a zero year at some point. This elasticity, counting a hit however it arrives and never penalizing the rule for the non-zero elections of multiply-elected presidents, is the same move that makes the Lincoln-Kennedy coincidence list feel uncanny: stretch the criteria until the data fits, then present the fit as astonishing. A rule that can be adjusted after the fact to capture whatever happened is not a rule that can be tested, and a pattern that survives only because the rule bends to protect it is not evidence of anything beyond the bending.

The Law of Truly Large Numbers

Statisticians have a name for the principle that makes coincidences inevitable: the law of truly large numbers, which holds that with a large enough number of opportunities, any outrageous thing is likely to happen. The presidency offers a surprisingly large number of opportunities once you count properly. There are dozens of presidents, each with a term or terms, each describable by a long list of attributes, the year of election, the month of birth, the state of origin, the prior occupation, the height, the religious affiliation, the number of children, and on and on. Cross-reference any of those attributes against the small set of in-office deaths and the number of possible patterns runs into the thousands. Searching that vast space for a striking alignment and then expressing astonishment when one turns up is like shuffling a deck, dealing a hand, and marveling that you received exactly those cards in exactly that order, an event whose probability is one in many trillions yet which happens every single time cards are dealt.

The persuasive force of the zero-year sequence comes from a cognitive trick that the law of truly large numbers exposes. People intuitively compute the odds of the specific alignment they are shown, never the odds of the search that found it. Shown seven for seven, the mind calculates roughly one in forty-five hundred and recoils at the improbability. What the mind never does is ask how many other equally arresting alignments were available to be discovered and weren’t, simply because they happened not to line up. Had no zero-year pattern existed, some other numerical or biographical regularity would almost certainly have surfaced to carry the same supernatural freight, and we would be discussing that one instead, equally convinced of its uncanniness. The improbability is real for the specific case and illusory for the phenomenon as a whole, because the phenomenon is not “this exact alignment occurred” but “some striking alignment occurred somewhere,” and the second is close to guaranteed.

History is littered with coincidences that feel as charged as the presidential sequence and that no one mistakes for the supernatural. Two of the founding generation’s giants, John Adams and Thomas Jefferson, died on the same day, the Fourth of July, 1826, exactly fifty years after the Declaration of Independence they had both shaped. James Monroe, a third founding president, also died on a Fourth of July, in 1831. Three of the first five presidents dying on Independence Day is a coincidence at least as eerie as the zero-year deaths, complete with patriotic poetry built right in, yet it spawned no curse and no theory of design, partly because the outcomes were positive rather than tragic and partly because no convenient villain stood ready to be blamed. The Adams-Jefferson coincidence shows the same machinery as Tippecanoe running in a register of wonder rather than dread, and the contrast is clarifying: when a coincidence lacks a sinister frame, we correctly file it as remarkable chance, and only when tragedy and a plausible scapegoat align do we reach for a curse.

The honest skeptic concedes the genuine emotional pull of the sequence while refusing the inference it invites. Seven for seven is a real run, the arithmetic of its improbability under naive assumptions is real, and the feeling of significance it produces is a real feature of human cognition. None of that licenses the leap to supernatural causation, because the same arithmetic and the same feeling attach to coincidences we readily recognize as meaningless the moment no villain is on offer. The zero-year sequence is the Adams-Jefferson Fourth of July coincidence with a Shawnee scapegoat attached, and the scapegoat is the only difference between a fact filed under wonder and a fact filed under curse.

The Findable Artifact: The Zero-Year Timeline and the Probability Frame

The cleanest way to hold the whole phenomenon in view is to lay the seven cases against the calendar and then set the statistical frame beside them. The timeline below is the InsightCrunch Zero-Year Run, the seven qualifying cycles from 1840 through 1960 with the cause of each death, followed by the 1980 case that ended it. The point of the artifact is to show both things at once: that the pattern is genuine, every cell filled, no exceptions, and that the causes filling those cells share nothing but the calendar.

Election Year	President	Date Died in Office	Documented Cause	Era / Means
1840	William Henry Harrison	April 4, 1841	Pneumonia or enteric fever	Natural disease, pre-germ-theory medicine
1860	Abraham Lincoln	April 15, 1865	Assassination (gunshot)	Confederate revenge, Ford’s Theatre
1880	James A. Garfield	September 19, 1881	Gunshot plus fatal infection	Office-seeker, lethal medical care
1900	William McKinley	September 14, 1901	Assassination (gunshot, gangrene)	Anarchist ideology, Buffalo
1920	Warren G. Harding	August 2, 1923	Heart attack or stroke	Natural cardiovascular failure
1940	Franklin D. Roosevelt	April 12, 1945	Cerebral hemorrhage	Untreated malignant hypertension
1960	John F. Kennedy	November 22, 1963	Assassination (gunshot)	Lone gunman, Dallas
1980	Ronald Reagan	Survived; left office 1989	Wounded March 30, 1981, recovered	Pattern broken

The statistical companion to the timeline is a single sentence with a footnoted qualification. Under an independence assumption and a roughly thirty percent per-term in-office mortality rate for the relevant historical period, seven consecutive zero-year deaths carry a probability near one in forty-five hundred, which sounds decisive until one remembers that the pattern was selected after the fact from an enormous field of possible patterns, a procedure under which striking coincidences are not rare but expected. The artifact, in other words, is built to be self-correcting. It hands the reader the eerie pattern and the deflating context in the same glance, so that the pattern can be appreciated for what it is, a remarkable run, without being mistaken for what it is not, a cause.

The Pattern Breaks: Reagan, 1981

The run ended in the operating room described at the top of this article, and the manner of its ending is itself instructive. Reagan, elected in 1980, was the eighth zero-year president in the sequence and by every expectation of the curse should have been the eighth to die in office. On March 30, 1981, John Hinckley Jr., a young man whose motive was an obsessive fixation on the actress Jodie Foster rather than any political grievance, fired six shots outside the Washington Hilton. One ricocheted off the armored limousine, struck Reagan under the left arm, ricocheted off a rib, and lodged in his lung an inch from his heart. He very nearly died. He lost a great deal of blood, his blood pressure crashed, and the surgical team worked for hours. He lived because a fast Secret Service agent shoved him into the car, because the motorcade diverted to the hospital within minutes, and because the medicine of 1981, antiseptic, transfusion-equipped, staffed by trained trauma surgeons, was a different universe from the medicine that had failed Garfield a century earlier. Garfield was shot in a less immediately lethal location and died of his doctors. Reagan was shot in a more dangerous spot and lived because of his.

The specifics of the 1981 response read like a deliberate rebuttal of 1881. Agent Jerry Parr’s split-second decision to push Reagan into the limousine and then, noticing bright frothy blood at the president’s lips, to redirect the car to George Washington University Hospital rather than back to the White House, almost certainly saved Reagan’s life by minutes. At the hospital, a trauma team that practiced sterile technique as a matter of course opened the chest, drained the blood filling the pleural cavity, and located the flattened bullet lodged near the heart. Reagan received transfusions to replace the liters of blood he had lost. Antibiotics stood ready to prevent the kind of runaway infection that had killed Garfield. Every element that had been absent in 1881, antisepsis, transfusion, trained thoracic surgery, rapid transport, was present in 1981, and the difference between absence and presence was the difference between a dead president and a living one. The sequence did not break because a curse expired on schedule. It broke because, for the first time, a gravely wounded zero-year president had access to medicine equal to his wound.

That contrast is the quiet refutation the curse never recovers from. If a supernatural force had been arranging presidential deaths at twenty-year intervals, the force would not have been thwarted by the existence of trauma surgery. Curses are not supposed to be defeated by a competent thoracic team and a unit of packed red cells. The fact that the pattern broke precisely when medical capability had advanced enough to save a gunshot president points to the real common thread among the seven prior deaths, which was never a curse but a combination of the office’s exposure to violence and the limits of contemporaneous medicine. Four of the seven were shot in eras when security was thinner and protective doctrine less developed; three died of conditions that modern medicine routinely manages. Improve the security and improve the medicine, and the pattern that looked supernatural turns out to have been mortal all along.

The cases that followed Reagan confirm the break. George W. Bush, elected in 2000, the next zero-year president, served two full terms and left office in good health, surviving a notable choking incident on a pretzel and the ordinary hazards of the modern presidency without any brush with in-office death. The twenty-year cadence that the curse insisted upon would have placed the next death somewhere around the term beginning after the 2000 election, and nothing of the kind occurred. The chain that ran unbroken from Harrison to Kennedy simply stopped, first interrupted by Reagan’s survival and then plainly ended by Bush’s two uneventful terms. A curse that can be ended by a surgeon and then fails to reassert itself was never a curse.

Why the Mind Insists on a Cause

Understanding why the zero-year sequence feels supernatural requires a short detour into how human cognition handles randomness, because the curse is less a fact about presidents than a fact about minds. People are equipped with a powerful and largely involuntary tendency to perceive patterns, a trait that served our ancestors well when the cost of seeing a predator that was not there was trivial and the cost of missing one that was there was fatal. The same machinery that correctly detects the rustle of a real threat also fires at random noise, producing the perception of faces in clouds, voices in static, and meaning in coincidence. Psychologists call the perception of pattern in randomness apophenia, and it is not a defect so much as an overactive setting of an otherwise useful system. The zero-year sequence is a textbook trigger: a genuine regularity, surrounded by the emotional charge of presidential death, presented to a mind built to find significance.

Layered on top of pattern perception is a second tendency, the detection of agency. Humans do not merely see patterns; they attribute them to actors with intentions. When something significant happens, the mind reaches reflexively for a who rather than a what, for a deliberate cause rather than a blind process. A string of deaths cries out, to this instinct, for an author, and the curse supplies one in the figure of the wronged Shawnee leader. The agency instinct explains why the curse needs a villain at all. A coincidence with no one behind it offers nothing for the mind to grip; a coincidence authored by a vengeful spirit offers a complete narrative with motive, method, and meaning. The instinct that turns falling leaves into the work of a wind spirit is the same instinct that turns seven presidential deaths into the work of a curse.

Tragedy intensifies both tendencies, which is why curses cluster around disaster rather than good fortune. The Adams-Jefferson Fourth of July coincidence, positive and patriotic, triggered wonder but no curse, because nothing about it demanded a culprit. Presidential deaths are different. They are losses, often violent, often accompanied by national trauma, and loss activates a desperate search for cause and meaning, a refusal to accept that something so consequential could be the product of mere chance. To say that Kennedy died because a man in a window happened to fire, for reasons still disputed, and that the date of his election was an irrelevant accident, feels intolerably arbitrary to a grieving nation. The curse offers the consolation of design. It says the deaths were not random, that they fit a plan, that the universe is ordered even when the order is malign. That consolation is false, but it is powerful, and its power explains the legend’s persistence far better than any evidence about 1811 ever could.

The discipline this article tries to model is the conscious override of these instincts: the deliberate act of noticing a striking pattern, feeling the pull toward an authored cause, and then asking whether the pattern actually requires the cause the instinct supplies. Most of the time it does not. The pattern is real, the feeling is real, and the cause is a story the mind told itself to make the feeling bearable. Recognizing that sequence in oneself, the pattern, the pull, the manufactured cause, is the single most transferable skill a study of the Tippecanoe legend can teach, applicable far beyond presidents to every coincidence that ever tempted a person to believe that something must be behind it.

The Complication: Do Not Throw Out the Pattern

Here is where a careless myth-bust would overreach, and where this one declines to. The temptation, having shown that the curse has no origin and no mechanism and that the statistics are less impressive than they first appear, is to wave the whole thing away as “mere coincidence,” as though saying the word “coincidence” dissolves the phenomenon. It does not, and pretending otherwise misunderstands what makes the Tippecanoe pattern worth a fourteen-thousand-word treatment in the first place.

The pattern is real. That sentence has to keep its full weight. Seven men, elected at perfectly regular twenty-year intervals, every one of them dead before his term or terms expired, with not a single exception across one hundred and twenty years. That actually happened. It is not a distortion, not a stretch, not an artifact of loose criteria like the Lincoln-Kennedy list. The dates are the dates. To call it coincidence is correct, but the word “coincidence” is doing a lot of quiet labor there, and the labor deserves to be made visible. A coincidence of this regularity and duration is a genuine and instructive thing: it is a real-world demonstration of how apparent order emerges from random processes when those processes are allowed to run long enough and are examined after the fact. The Tippecanoe pattern is not interesting in spite of being a coincidence. It is interesting because it is one of the most vivid coincidences in the documentary record, and studying it teaches the mind something durable about the difference between a pattern that means something and a pattern that only feels as though it must.

This is why the proper verdict grades three claims separately rather than collapsing them into a single thumbs-down. If the article simply declared “the curse is false, the end,” it would leave the reader with the impression that the underlying observation was bogus, and the underlying observation is not bogus. The deaths are documented. The twenty-year regularity is documented. What is bogus is the leap from the regularity to the supernatural agent, and the further leap from the agent to a causal force operating across a century. Keeping the true thing true while killing the false thing dead is harder than blanket dismissal, and it is the only intellectually honest way to handle a coincidence that happens to be both genuine and genuinely striking.

There is a related complication worth naming, because skeptics sometimes overcorrect into it. The thirty-percent base rate and the one-in-forty-five-hundred figure are themselves rough, and a determined believer can push back on the independence assumption, arguing that presidential deaths are not really independent trials. That pushback has some merit and ultimately strengthens the skeptical case rather than weakening it. In-office deaths cluster in eras of poor medicine and lax security precisely because those conditions, not the calendar, drove the outcomes. The deaths are correlated, but they are correlated with the dangers of the office in particular periods, not with the last digit of an election year. Once the real correlate is identified, the zero-year framing reveals itself as a coincidence riding on top of a genuine historical fact, that being president in the nineteenth and early twentieth centuries was a hazardous job. The hazard was real; its alignment with zero years was chance.

The Believer’s Strongest Case, Answered

Fairness requires steelmanning the position before burying it, so consider the most sophisticated version a defender of the sequence might mount. The defender concedes that the Tecumseh origin is fiction and that no spirit literally fired Booth’s pistol, but argues that the regularity itself is too clean to dismiss, that twenty-year spacing across seven cycles with zero exceptions is a precision that bare chance rarely produces, and that the skeptic’s one-in-forty-five-hundred figure actually concedes the point by acknowledging a genuinely small probability. This is the best the believer can do, and it deserves a direct answer rather than a wave of the hand.

The answer has three parts. First, the spacing is not as precise as it sounds, because the rule was defined to make it precise. The deaths did not occur at neat twenty-year intervals; they occurred at irregular intervals dictated by when each man happened to die, and only the elections were spaced twenty years apart by the fixed constitutional calendar. Harrison died after thirty-one days, Lincoln after more than four years, Garfield after a few months of his single term, and so on. The tidy cadence belongs to the election dates, which are mechanically regular for every president regardless of fate, not to the deaths, which were scattered. Second, the zero-exceptions claim depends on the flexible counting already exposed, the treatment of FDR’s 1940 win and Lincoln’s 1860 win as binding while their other elections are ignored. A rule that selects favorable cases and discards inconvenient ones will always look exception-free, because the exceptions have been defined away. Third, the small probability is the wrong probability, as the law of truly large numbers establishes; a one-in-forty-five-hundred event drawn from a search across thousands of possible patterns is not a marvel but an expectation. The believer’s strongest case, examined closely, rests on mistaking the regularity of the constitutional calendar for the regularity of fate, on counting rules that bend to fit, and on the wrong statistical question. It is the best available defense, and it fails on all three counts.

The Verdict

The Curse of Tippecanoe, graded as three distinct claims, earns three distinct marks. The historical pattern, that every president elected in a zero year from 1840 through 1960 died in office, is true, documented seven times without exception, and should be neither denied nor minimized. The supernatural origin, that Tecumseh or Tenskwatawa pronounced a curse at or after the 1811 Battle of Tippecanoe, is false, supported by no contemporary source and contradicted by the actual events of 1811, in which Tecumseh was absent and The Prophet was discredited by the defeat the legend later credits him with avenging. The causal claim, that a curse is the active force linking seven deaths of at least four medically and circumstantially unrelated kinds across one hundred and twenty years, is false and incoherent, lacking any mechanism and resting on circular reasoning in which the deaths prove the curse and the curse explains the deaths.

The statistical framing settles the residual unease. Seven for seven feels impossible, but the probability, around one in forty-five hundred under generous assumptions, is small without being extraordinary, and it is the wrong probability to fixate on in any case. The pattern was selected after the fact from an enormous field of patterns that could have emerged instead, and under after-the-fact selection, striking coincidences are not surprises but near-certainties. The break in 1981, when Reagan survived a wound that the medicine of Garfield’s era would have made fatal, exposes the real common thread among the prior deaths: the exposure of the office to violence in eras of thin security and the limits of contemporaneous medicine, neither of which is a curse and both of which improved enough to end the run. The Bush terms after 2000 confirmed that the chain had genuinely stopped.

The full verdict, then, is that the pattern deserves its fame and the curse deserves none. People are right to find the Tippecanoe run remarkable. They are wrong to find it supernatural. The correct response to a one-hundred-and-twenty-year coincidence is not to invent a Shawnee curse to explain it, nor to deny that anything happened, but to hold both truths at once: the run was real, and it was random, and the simultaneous reality and randomness is exactly what makes it one of the most useful object lessons American history offers about how the human mind manufactures meaning from chance.

Legacy: Why the Curse Outlived Its Own Refutation

The most telling fact about the Curse of Tippecanoe is that explaining it does not kill it. The pattern broke in 1981, the origin story was demolished by Edmunds, the statistics were laid bare by the Gardner tradition of skeptical analysis, and yet the curse persists in popular culture, trotted out every twenty years like clockwork whenever a zero-year election approaches. Its survival is itself the legacy, and the legacy connects to something larger than one piece of folklore.

Curses of this kind flourish in the same soil that grows conspiracy theories, and the soil is distrust. The Tippecanoe legend reached its cultural peak in the decades around the Kennedy assassination, a period when American confidence in official explanations was eroding fast under the pressure of Dallas, Vietnam, and eventually Watergate. A public primed to suspect that hidden forces shape national events found in the curse a tidy, narratively satisfying account of presidential deaths that the official record presented as a string of unrelated tragedies. The same impulse drives the enduring Lincoln-Kennedy coincidence folklore, which gained its widest circulation in the years after Dallas precisely because it offered pattern and meaning where the official story offered only the random horror of two lone gunmen a century apart. The curse and the coincidence list are cousins, both products of a hunger for design in events that resist it.

The pattern also intersects with a real structural feature of the presidency that the curse obscures rather than illuminates. Every one of the seven zero-year deaths triggered a succession, handing the office to a vice president who had not been elected to lead. The full audit of every vice president who inherited the presidency mid-crisis shows how consequential those unplanned successions were, from Tyler establishing the precedent that an inherited president is a full president, through Andrew Johnson, Arthur, Theodore Roosevelt, Coolidge, Truman, and Johnson again. The curse narrative treats these successions as a spooky side effect of a supernatural pattern. The institutional reality is the reverse: the successions were the most important consequences of the deaths, and they shaped the development of the office far more than any curse could pretend to. Reading the zero-year deaths as a curse trivializes what was actually a recurring stress test of the constitutional machinery of succession.

There is a further pattern worth distinguishing from the curse, because the two are easily confused. Several of the zero-year deaths were assassinations, and assassination attempts on American presidents are not randomly distributed across the calendar. The clustering of assassination attempts around economic panics and periods of social strain is a real pattern with a plausible mechanism, the elevation of political violence during times of dislocation, which is exactly the kind of cause the Tippecanoe curse lacks. Czolgosz acted in the wake of industrial upheaval and anarchist ferment; the conditions that produce assassins are social and economic, not numerological. Distinguishing the genuine, mechanism-bearing pattern of violence-and-strain from the spurious, mechanism-free pattern of zero-year-and-curse is precisely the discipline this entire analysis is meant to cultivate. Some patterns point at causes. Others only point at the mind that found them.

The curse persists, finally, because it is a good story, and good stories outlive their refutations. It has a villain with a grievance, a defeated leader striking back from beyond the grave. It has dramatic unity, seven deaths bound by a single calendrical thread. It has a satisfying payoff, the eerie fulfillment of a prophecy across a century and a quarter. And it has, crucially, a built-in resistance to disproof, because no amount of evidence about 1811 or about probability or about trauma surgery can quite erase the gut-level shock of seven for seven. The story survives not despite the facts but in a separate register from them, in the part of the mind that responds to narrative rather than evidence. The job of an honest account is not to expect the story to die, which it will not, but to make sure that anyone who wants the facts can find them stated clearly, with the pattern affirmed, the curse denied, and the difference between the two made permanent. The same instinct that makes the curse compelling is the instinct worth watching in oneself, the same one that turns a tavern brawl into a noble defense of a wronged wife in the Jackson dueling legends, the same one that finds destiny in coincidence and design in chance. Understanding why the Curse of Tippecanoe feels true is, in the end, more valuable than the simple fact that it is not.

A final thought ties the whole analysis to the wider purpose of grading myths at all. The point of dismantling a legend like this one is never merely to be correct. It is to leave the reader with a portable method, a way of meeting the next striking coincidence with the right questions already loaded. When some future arrangement of facts presents itself as too clean to be chance, the disciplined reader will now know to ask whether the pattern was predicted in advance or noticed after the fact, whether the rule defining it bends to absorb inconvenient cases, whether the proposed cause carries any mechanism or merely relabels the outcomes it claims to explain, and whether the small probability being waved around is the probability of this specific alignment or of some alignment somewhere. Those four questions dissolve the Tippecanoe curse, and they dissolve most of its cousins too. They are the lasting takeaway, worth far more than the verdict on any single legend, because legends are endless and the method is one. The deaths of seven presidents were real, the regularity of their election years was real, and the curse that supposedly bound them was a story the living told to make the dead make sense. Knowing the difference, and knowing how to find it again, is the whole reward.

Frequently Asked Questions

Q: What is the Curse of Tippecanoe?

The Curse of Tippecanoe, also called Tecumseh’s Curse, is the popular belief that every president elected in a year ending in zero is doomed to die in office. The claim holds that this fate traces to a curse supposedly laid by the Shawnee leader Tecumseh or his brother Tenskwatawa after William Henry Harrison defeated them at the 1811 Battle of Tippecanoe. The historical pattern the legend describes is real for the period from 1840 through 1960: seven consecutive presidents elected in zero years all died in office. The supernatural explanation, however, has no documentary basis. No contemporary source records any such curse, and the seven deaths resulted from entirely unrelated causes, including disease, heart failure, and four separate assassinations spread across more than a century.

Q: Which presidents died under the Curse of Tippecanoe?

Seven presidents make up the historical run. William Henry Harrison, elected in 1840, died of illness in 1841 after thirty-one days. Abraham Lincoln, elected in 1860, was assassinated in 1865. James Garfield, elected in 1880, was shot in 1881 and died of infection. William McKinley, reelected in 1900, was assassinated in 1901. Warren Harding, elected in 1920, died of heart failure in 1923. Franklin Roosevelt, elected to his third term in 1940, died of a cerebral hemorrhage in 1945. John Kennedy, elected in 1960, was assassinated in 1963. Each man won a zero-year election and died before completing his service, but the medical and circumstantial causes share nothing beyond the calendar.

Q: Did Tecumseh actually curse William Henry Harrison?

No contemporary evidence supports the idea that Tecumseh cursed Harrison. The historian R. David Edmunds, the leading scholar of both Tecumseh and his brother, establishes that Tecumseh was not even present at the 1811 Battle of Tippecanoe. He was far to the south recruiting allies when Harrison attacked the settlement at Prophetstown. The brother who was there, Tenskwatawa, known as The Prophet, had promised his warriors invulnerability to American bullets, and the defeat shattered his spiritual reputation rather than empowering him to issue effective curses. The curse story appears nowhere in the records of 1811 or in Harrison’s 1840 campaign, when his Tippecanoe fame was central. The legend was assembled in the twentieth century, long after the events it claims as its origin.

Q: Is the Curse of Tippecanoe real?

The pattern is real; the curse is not. From 1840 through 1960, every president elected in a zero year genuinely did die in office, with no exceptions across seven consecutive cycles. That historical fact is well documented and should not be denied. What is false is the explanation. There is no evidence of any curse pronounced in 1811, no mechanism by which a curse could cause deaths as varied as pneumonia, heart attack, cerebral hemorrhage, and gunshot wounds, and no reason to treat the regularity as anything other than a striking coincidence. The pattern broke when Ronald Reagan survived being shot in 1981, and it did not return, which a genuine supernatural force would not allow. Real pattern, false cause.

Q: How likely was the seven-for-seven pattern statistically?

Using a rough thirty percent in-office mortality rate for the relevant historical period and treating each president’s survival as independent, the probability of seven consecutive zero-year deaths is about 0.30 raised to the seventh power, roughly one in forty-five hundred. That sounds small, but it is the wrong number to focus on. The pattern was identified after the fact, selected from an enormous field of patterns that could have emerged instead, such as deaths clustering in even years or among generals or among tall presidents. When a pattern is chosen after it appears rather than predicted in advance, striking coincidences become not rare but expected. The genuine probability question is whether some eerie pattern would surface somewhere, and the answer to that is nearly certain.

Q: Why did the Curse of Tippecanoe break with Ronald Reagan?

Reagan, elected in 1980, was shot on March 30, 1981, by John Hinckley Jr. and very nearly died. A bullet lodged in his lung an inch from his heart, and he lost substantial blood. He survived because of fast Secret Service action, a quick diversion to a nearby hospital, and the trauma medicine of 1981, which included antisepsis, blood transfusion, and trained thoracic surgeons. James Garfield, by contrast, was shot in a less immediately lethal spot in 1881 but died because his doctors probed the wound with unsterilized hands and introduced fatal infection. The contrast reveals the real common thread among the seven prior deaths: the office’s exposure to violence and the limits of contemporary medicine. Better medicine, not the absence of a curse, saved Reagan.

Q: Was George W. Bush affected by the Curse of Tippecanoe?

George W. Bush, elected in 2000, was the next zero-year president after Reagan, and he served two full terms before leaving office in good health in 2009. His most famous brush with danger involved choking briefly on a pretzel, hardly the stuff of a supernatural death sentence. Bush’s uneventful eight years confirmed that the pattern had genuinely ended. The chain that ran unbroken from Harrison in 1841 through Kennedy in 1963 was first interrupted by Reagan’s survival in 1981 and then plainly closed by Bush’s two healthy terms. A curse capable of killing seven presidents across one hundred and twenty years does not simply give up after one survives surgery. The pattern stopped because it was always a coincidence, not a curse.

Q: Where did the Curse of Tippecanoe legend come from?

The legend appears to be a twentieth-century construction rather than an 1811 event. The pattern only becomes noticeable after several deaths accumulate, and the curse narrative reaches back across decades to attach a dramatic origin, the defeated Shawnee brothers, to a coincidence identified long after they died. The most likely vehicle for popularizing the story was Ripley’s Believe It or Not, the syndicated feature and book franchise that specialized in arresting, dubious curiosities. The timing centers on Franklin Roosevelt’s election in 1940, which set up a publicized prediction that the 1940 winner would die in office. When Roosevelt died in 1945, the fulfilled forecast turned a curiosity into a widespread conviction, cementing the curse in popular culture.

Q: Did Tenskwatawa, The Prophet, curse the presidency?

Tenskwatawa is the figure most often named as the curse’s author in versions that distinguish him from his more famous brother Tecumseh. He was present at the 1811 Battle of Tippecanoe and had assured his followers that his spiritual power would protect them from American bullets. The crushing defeat that followed discredited him precisely because his promised protection failed, scattering the pan-Indian confederacy and gutting his authority. A spiritual leader whose reputation collapsed because his powers visibly did not work is an odd candidate for the author of a curse that supposedly operated flawlessly for one hundred and twenty years. No contemporary record attributes any presidential curse to him, and the historian R. David Edmunds, who wrote the standard study of The Prophet, finds no basis for the claim.

Q: How did each of the seven presidents actually die?

The seven died of strikingly different causes. Harrison succumbed to pneumonia or possibly enteric fever from contaminated water, a natural death in an era before antibiotics. Lincoln was shot by John Wilkes Booth in an act of Confederate revenge. Garfield was shot by a delusional office-seeker and died of the infection his doctors caused. McKinley was shot by an anarchist and died of gangrene. Harding died of an apparent heart attack or stroke from cardiovascular disease. Roosevelt died of a cerebral hemorrhage caused by untreated severe hypertension. Kennedy was shot by a lone gunman in Dallas. Two natural disease deaths, one chronic vascular death, and four assassinations by killers with incompatible motives across a century share only one feature: each man won a zero-year presidential election.

Q: Is the Curse of Tippecanoe the same as the Lincoln-Kennedy coincidences?

They are related but distinct pieces of folklore. The Tippecanoe curse covers seven presidents across one hundred and twenty years and centers on the twenty-year death pattern. The Lincoln-Kennedy coincidences focus narrowly on the eerie parallels between two of those seven presidents, both elected in zero years and both assassinated by gunshot a century apart. The Lincoln-Kennedy list is far flimsier, relying on stretched and sometimes fabricated parallels that collapse under scrutiny, as the detailed grading of those fifteen claims demonstrates. The Tippecanoe pattern, by contrast, rests on genuine documented deaths and is a real coincidence even after debunking. Both legends share the same psychological root, a hunger for hidden design behind national tragedy, but the Tippecanoe pattern is the sturdier of the two.

Q: Why do people still believe in the Curse of Tippecanoe?

Belief persists because the curse is a compelling story that operates in a register separate from evidence. It has a villain with a grievance, a defeated leader striking from beyond the grave; dramatic unity, seven deaths bound by a single thread; and a satisfying payoff, a prophecy fulfilled across a century. It also resists disproof, since no fact about 1811 or about probability erases the visceral shock of seven for seven. The legend flourished during the era of the Kennedy assassination, Vietnam, and Watergate, when distrust of official explanations ran high and a public primed to suspect hidden forces found the curse more satisfying than a string of unrelated tragedies. Good stories outlive their refutations, and the curse is a very good story.

Q: Did Franklin Roosevelt’s death really fit the pattern?

Roosevelt fits the pattern only through a selective reading that reveals how the rule actually works. He was elected four times, in 1932, 1936, 1940, and 1944, and only the 1940 election ends in zero. He died in April 1945, during the term he won in 1944, a non-zero year. Believers count the 1940 election as the trigger and the 1945 death as the payoff, quietly ignoring that the death occurred in the term following a non-zero election. This elasticity, keeping the one zero-year win and discarding the three others, exposes the after-the-fact flexibility that lets the rule capture whatever happened. A genuine rule cannot be bent to fit the data after the data arrives, yet that bending is exactly what Roosevelt’s inclusion requires.

Q: What does the Curse of Tippecanoe teach about coincidences?

The curse is one of the best object lessons available about how the mind manufactures meaning from chance. It demonstrates that apparent order can emerge from random processes when those processes run long enough and are examined after the fact. The crucial insight is the retrospective pattern trap: when a pattern is noticed and then its odds are calculated as though it had been predicted in advance, the calculation badly overstates how surprising the pattern is. The relevant question is never the odds of one specific pattern but the odds that some striking pattern would appear somewhere in a large field of possibilities, and that second probability is high. The curse trains the skeptical reflex to ask not whether a pattern is real, which it often is, but whether it requires the cause being offered.

Q: Were the zero-year presidential deaths really independent events?

Strictly speaking, no, and acknowledging the dependence actually strengthens the skeptical case rather than the believer’s. In-office presidential deaths cluster in eras of primitive medicine and lax security, conditions that drove the outcomes far more than any calendar did. The deaths are correlated, but with the genuine hazards of the office in particular historical periods, not with the last digit of an election year. Four of the seven were shot in eras of thinner protection, and three died of conditions modern medicine routinely manages. Once the real correlate is identified, that being president in the nineteenth and early twentieth centuries was a dangerous job, the zero-year framing reveals itself as a coincidence riding atop a real historical fact. The hazard was genuine; its alignment with zero years was chance.

Q: How does the Curse of Tippecanoe relate to presidential succession?

Every one of the seven zero-year deaths handed the presidency to a vice president who had not been elected to lead, making the curse narrative obscure something genuinely important. The successions, not any supernatural force, were the most consequential results of the deaths. Tyler established the precedent that an inherited president holds full authority; later successions brought Andrew Johnson, Arthur, Theodore Roosevelt, Coolidge, Truman, and Lyndon Johnson to power, each reshaping the office in turn. The complete audit of vice presidents who inherited the presidency mid-crisis shows how repeatedly these unplanned transitions stress-tested the constitutional machinery. Reading the deaths as a curse trivializes what was actually a recurring trial of the succession system, the real and lasting institutional consequence of the pattern.

Q: Which historians have studied or debunked the Curse of Tippecanoe?

Several scholars and analysts have addressed the legend from different angles. R. David Edmunds, author of the standard studies of Tecumseh and Tenskwatawa, debunks the supposed 1811 origin by establishing that Tecumseh was absent from the battle and The Prophet was discredited by it. Paul F. Boller Jr., in Presidential Anecdotes and Presidential Campaigns, treats the curse as folklore, recounting it without endorsement and noting its modern popular-press character. William A. DeGregorio’s reference work documents the medical and circumstantial facts of each death. Martin Gardner, in the skeptical tradition of Fads and Fallacies in the Name of Science, supplies the framework for understanding retrospective pattern-mongering. These figures disagree on emphasis, with Boller cataloging the folklore and Edmunds and Gardner dismantling the origin and the statistics, but none endorses the curse as historical fact.

Q: Could the Curse of Tippecanoe return in a future zero-year election?

There is no curse to return, so the question answers itself, but the framing is worth addressing because it surfaces every twenty years. The pattern that held from 1840 to 1960 was a coincidence produced by the office’s historical exposure to violence and the limits of older medicine, not by any supernatural agent. Both of those underlying conditions have improved dramatically: presidential security is now professionalized and intensive, and trauma medicine can save wounds that once killed. Reagan’s survival in 1981 and Bush’s two healthy terms after 2000 demonstrate the change. A future zero-year president faces the ordinary risks of the office, no more and no less than a president elected in any other year. The twenty-year cadence was never a rule, only a run, and runs end.

Q: What is the single best way to summarize the verdict on the curse?

The pattern is real and the curse is not, and holding both truths at once is the entire point. Seven presidents elected in zero years from 1840 through 1960 genuinely died in office, an unbroken run across seven cycles that deserves its fame as one of the most striking coincidences in American history. The supernatural explanation fails completely: no contemporary source records the 1811 curse, the seven deaths share no common mechanism, and the statistics, while initially impressive at roughly one in forty-five hundred, dissolve under the recognition that the pattern was selected after the fact from a vast field of possible patterns. The run broke in 1981 because trauma surgery saved Reagan, revealing that the real thread was always mortal danger and primitive medicine, never a Shawnee curse. People are right to be amazed and wrong to be spooked.

Q: Why didn’t the curse appear in newspapers before the twentieth century?

The legend’s absence from nineteenth-century sources is one of the strongest pieces of evidence against its authenticity. If a curse had been pronounced in 1811, it would have surfaced during Harrison’s 1840 campaign, when his Tippecanoe fame was the centerpiece of his political identity and his opponents would have seized on any dramatic Native prophecy against him. It appears in none of the campaign material, none of the 1841 obituaries, and no contemporary record. The reason is simple: the legend could not exist until the pattern existed, and the pattern required several accumulated deaths to become visible. A curse story attached to a sequence cannot predate the sequence it explains. Its first appearances trace to twentieth-century popularizers, decades after the events it claims as its origin, which is precisely the signature of a story built backward from outcomes rather than forward from a cause.

Q: Does the Curse of Tippecanoe prove anything about the supernatural?

It proves nothing about the supernatural and a great deal about human cognition. The sequence is a coincidence, genuine and striking, produced by chance operating on an office that was historically dangerous in eras of poor medicine and thin security. What the legend really demonstrates is how reliably the mind manufactures supernatural explanations for emotionally charged coincidences, especially tragic ones with a convenient scapegoat available. The same machinery that built the Tecumseh curse builds countless other superstitions, from lucky numbers to omens to conspiracy theories, all of them attempts to impose authored meaning on the random. Studying the curse teaches skepticism not by mocking belief but by revealing the cognitive steps that lead a reasonable person from a real pattern to a false cause. The supernatural is not the lesson. The lesson is the pattern-seeking, agency-detecting, meaning-hungry mind that finds the supernatural so easy to believe.