The Birthday Coin Toss

The Birthday Coin Toss

Eli: The statistician Anthony Davison frames extreme old age as an even wager. He puts it like this: "Think of it like a coin flip — when you reach 108, you flip a coin on your birthday. If comes up heads, you live to your next birthday. If it comes up tails, you die before turning 109,".

Mara: That bet points directly to our question. Why do statisticians like Holger Rootzén argue that human mortality rates level out past 110, instead of crashing into an absolute ceiling?

Eli: Today's angle is why statisticians argue human mortality rates plateau past age 110. They build on the disputed late-life mortality deceleration hypothesis. That idea claims our risk of dying stops accelerating in deep old age and flattens out.

Mara: And following that plateau, certain researchers argue our lifespan has no finite upper limit.

Eli: Biologists strongly dispute that claim, arguing our bodies face a hard natural barrier. But whether that wall exists matters for pension solvency, healthcare planning, and aging research.

Mara: Yet regular aging does not look like a 50-50 bet. In plain terms, how does normal adult mortality actually behave?

Eli: It comes down to mortality hazard, which means the immediate chance that an individual alive today dies across a specific window, like the coming year.

Mara: In 1825, British actuary Benjamin Gompertz showed that this risk surges over time. He formulated the Gompertz law: the mathematical pattern where, past early adulthood, your chance of dying doubles roughly every eight years.

Eli: Actuaries call that compounding the Gompertz doubling rule. If death rates keep multiplying at that pace, reaching extreme ages soon becomes mathematically impossible.

Mara: Then real records entered the picture. The International Database on Longevity launched in 2010 to track validated deaths of supercentenarians aged 110 or older.

Eli: Looking at verified deaths across Western countries and Japan, researchers found that the doubling stops.

Mara: It just flattens out! At age 110, proponents calculate your annual probability of dying settles around 47 percent. Survival turns into an annual birthday coin toss.

Eli: Plateau proponents calculate those annual odds of surviving hold near 50 percent each year.

Mara: A steady flip. How do statisticians actually prove that mortality stops climbing at the edge of life?

The Shape of Extremes

Eli: Statisticians reach for extreme value theory. That is a branch of statistics designed to model rare, extreme events, like once-in-a-century floods or record-breaking human lifespans.

Mara: Right, looking at the very tail of the distribution instead of the average crowd.

Eli: Plateau proponents like Rootzén and Dmitrii Zholud argued Gompertz doubling halts after 110. In their core analysis, they evaluated 566 records meeting the database's top validation level, category A.

Mara: They measured survival past that threshold using a generalized Pareto distribution. That is a flexible mathematical curve that models data points falling above an extreme threshold, showing whether they taper off to a hard stop or continue indefinitely.

Eli: The key is its shape parameter. That single number reveals whether longevity slams into an absolute wall or remains open-ended.

Mara: A negative value would signify a hard limit. But Rootzén and Zholud estimated the parameter at zero, which mathematically describes an unchanging hazard rate and a lifespan that is unbounded, though still short.

Eli: That mathematical model aligned with empirical findings in Italy. Demographers Elisabetta Barbi and Kenneth Wachter investigated 3,836 Italian citizens who reached extreme longevity.

Mara: All of them had survived to age 105 or older.

Eli: And their yearly death risk stopped climbing, leveling off at roughly 50 percent. That left an average remaining life expectancy of about 1.5 years.

Mara: And that expected lifespan stayed identical even if someone reached 110?

Eli: Yes, projected to remain at 1.5 years.

Mara: As Wachter put it, “our data tell us that there is no fixed limit to the human lifespan yet in sight.” He and James Vaupel argued that demographic selection explains this flat line: the progressive survival of only the hardiest, most disease-resistant individuals as frailer cohort members pass away earlier.

Eli: Statistician Adrian Raftery pointed out that “this is a very select group of very robust people.” In mathematical frailty models—statistical models accounting for hidden differences in vulnerability across individuals in a population—once more vulnerable people drop out, the survivor group's mortality hazard flattens into a constant.

Mara: That sounds startling on its face. Does the absence of a mathematical ceiling mean individual humans could live forever?

Why a Flat Line Is Not Immortality

Eli: Not in the slightest. Talking about having no limit easily confuses people into imagining eternal life, as if aging simply switched off. Back in 2021, Anthony Davison made clear that their findings do not suggest anyone will live forever.

Mara: Rootzén and Zholud suggested in 2017 that human life is unlimited, but short, though researchers still actively contest that view. An unbounded statistical distribution merely lacks a theoretical upper bound.

Eli: So having no ceiling in a mathematical distribution is completely different from an individual living on without an end?

Mara: Completely different. Under an exponential model—a statistical framework where the risk of an event remains constant over time—an unchanging hazard never produces an infinite lifespan. Beyond age 110, death risk remains level. That annual survival probability still sits at roughly forty-seven percent.

Eli: Facing an even coin toss on every birthday is certainly nobody's idea of eternal life. Picture a cohort of people at age 108. Start with 1,000 of them.

Mara: One thousand people at the starting line, all flipping for another year.

Eli: Annual mortality runs at fifty percent. That means roughly 500 make it through that initial year. Over the following year, the group drops to 250.

Mara: That steep decline illustrates cumulative survival probability. That is the overall chance of surviving across several consecutive years, calculated by multiplying each year's survival odds together. Toss tails once, and the run finishes.

Eli: A single tails and your entire run is finished right then. Even if the annual hazard never climbs, the math compounds against you with every flip.

Mara: With survival chances cutting in half after every birthday, cumulative odds plummet. In 2021, Davison's team evaluated the likelihood of seeing 130. For any single person, the probability comes out below one in a million.

Eli: Biologist Steven Austad put that contrast between population equations and human reality into perspective. He observed, “It’s not really the limit that people who talk about limits think about, but it’s also not immortality.”

Mara: Austad pointed out that the chances of surviving past 125 are vanishingly slim. Rootzén examined supercentenarians in a 2017 study. Beyond age 110, remaining life is brief. The model found an average excess lifetime of just 1.34 years.

Eli: Barely sixteen extra months on average. What do biologists and skeptical demographers say to challenge this plateau theory?

The Pushback: Hard Limits and Dirty Data

Mara: Biologists argue our bodies cannot simply brush aside physical decay. Those who argue for a natural ceiling claim human lifespan is bounded by fixed biological limits.

Eli: Brandon Milholland and Jan Vijg noticed that survival gains shrink after people reach 100. In global demographic files, the oldest age recorded at death has stalled since the mid-1990s.

Mara: At what age did it stall?

Eli: The annual ceiling leveled off at 114.9 years. That limit held even outside of Jeanne Calment's 1997 record.

Mara: And that holds for supercentenarians too?

Eli: Milholland asked the obvious question: “Why would these processes suddenly halt at age 110?”

Mara: That barrier is cellular senescence: the biological process where aged cells lose the ability to divide and repair tissue, causing organs and systems to progressively fail.

Eli: Gerontologists see clear structural decay, including worn chromosome tips, damaged proteins, and exhausted stem cell reserves.

Mara: A 2024 paper in Nature Aging found longevity increases in developed nations are tapering off. S. Jay Olshansky put it directly: “If you live long enough, you run up against the biological process of aging.”

Eli: Researcher Saul Newman examined the records in 2018. He showed that clerical errors occurring less than once in 10,000 cases can produce a completely artificial plateau.

Mara: One mistake in 10,000 can fabricate an entire plateau?

Eli: At the far fringe, absolutely. When records get misreported, younger people accidentally inflate the oldest group. Newman concluded plateaus are “largely, if not entirely, artefacts of error processes.”

Mara: Demographic history shows how easily false milestones persist. For more than a century, record registries documented a Canadian named Pierre Joubert as an authentic supercentenarian.

Eli: His case withstood 100 years of examination. Yet when demographers finally pulled his original certificate, they discovered he actually passed away at 65.

Mara: Modern registries also disqualified Shigechiyo Izumi after scholars in Japan judged his case to be false. When every birth certificate is scrutinized, does the mortality plateau survive in bulletproof data?

The Outer Edge of Life

Eli: That depends on who you ask. Demographer Kenneth Wachter defended the Italian records, calling Newman's proposed error rate “wildly implausibly high.”

Mara: He argued that “these are the best data for extreme-age longevity yet assembled.” For his team, no fixed lifespan limit is yet in sight.

Eli: Still, I have to concede that even with strict audits, observations past age 114 become too scarce for reliable conclusions.

Mara: That is why ceiling proponents like Jay Olshansky say, “The plateau is a result of having too few people to generate reliable mortality statistics.”

Eli: Just how scarce are the numbers out at that fringe?

Mara: Across the entire World dataset, only nine people survived beyond 115, leaving hardly any evidence to test the model.

Eli: Just nine people.

Mara: Demographers need an extinct birth cohort, which means a group born in the same window where every single member has died, leaving no incomplete records.

Eli: That thin sample leaves Jeanne Calment as the ultimate benchmark. She passed away in 1997 at 122 years and 164 days.

Mara: And through the quarter-century that followed, nobody has managed to surpass that mark.

Eli: Not a single soul.

Mara: Yet under Anthony Davison's birthday coin toss comparison, her record is mathematically expected eventually. A steady 50 percent annual risk just demands an extraordinarily lucky streak.

Eli: That circles right back to the stakes: knowing if a ceiling exists directly affects the solvency of pension systems. Actuaries and biologists both depend on solid mortality figures at extreme ages.

Mara: That planning turns urgent very quickly. By 2100, the worldwide count of centenarians is projected to top 25 million.

Eli: So the dispute stays wide open.

Mara: Until thousands more people reach extreme age, we are left watching that yearly coin flip, waiting to see where the real frontier lies.

Eli: Read up on extreme value theory.

Mara: Follow science with us across more episodes as researchers debate whether human lifespan has a natural limit.

Eli: Thanks for spending this time with us. I'm Eli.

Mara: And I'm Mara. Until next time, this has been Angle.