Yorick › When will I die

When will I die? Three kinds of answer, and only one worth having

Type the question into a search box and you get quizzes, countdowns and confident dates. They are not all doing the same thing. Here is what separates them.

Download Yorick on the App Store

Yorick answers it with life tables and cited studies. Free result.

The three answers you will find

1. The entertainment quiz

The oldest form, and still the most common. You answer a handful of questions, a number appears, and somewhere in the small print the app tells you it is for entertainment purposes only. Several of the best known death clock apps on the App Store are filed under Entertainment or Lifestyle rather than Health, which is the honest place for them. There is nothing wrong with this as a diversion. It is simply not information.

You can spot one quickly: it never tells you where a number came from, and the answer never carries a margin.

2. The AI black box

The newer form. An app promises a model trained on mortality research and hands you a date. Sometimes there is real work behind it. The problem is that you cannot tell, because the model does not show you which factor moved your number or by how much. If quitting smoking is worth seven years in the answer, you have no way to check whether that came from a published cohort study or from a number someone chose because it felt persuasive.

A model you cannot audit is a model you have to trust, and trust is a strange thing to ask for in exchange for a death date.

3. The actuarial estimate

The dull one, and the only one that is doing real arithmetic. Actuaries have been answering this question for insurers for two centuries, using a period life table: a table that says, for every single year of age and for each sex, how many further years a person of that age can expect on average. The United States Social Security Administration publishes one. In the 2022 table a newborn boy can expect 74.7 further years and a newborn girl 80.2.

That table is the honest starting point. Everything else is an adjustment to it.

Why the honest answer is a range

A life table is a statement about a population, not about you. It says that among a hundred thousand people who share your age and sex, the average remaining lifetime is such and such. Some of those hundred thousand die next year and some reach a hundred. Collapsing that spread into a single date and printing it in a large font is a presentation choice, not a mathematical result.

This is why a calculator worth using shows you the width of its own uncertainty. Yorick does it by attaching a likely range to every result. The range is never narrower than three years, and it grows with the amount of life you have left, because a projection forty years out deserves a much wider band than one made at eighty. No result is allowed past age 105.

A useful test. Ask a calculator what its answer would be if you were five years older with identical habits. A real model will move the date by roughly five years and narrow its range. A guess will move it in ways that make no sense.

What genuinely moves the number

The published effects are not evenly sized, and any calculator that treats them as if they were is wrong before it starts. Smoking dominates everything else. Social isolation and a poor diet are surprisingly large. Supplements, despite an entire industry, are worth nothing at all in the mortality trials.

Rough magnitudes from the research Yorick is built on, expressed as years of life expectancy around age 50:

  • Heavy smoking, about 10 years lost. Doll et al., BMJ 2004 and Jha et al., NEJM 2013.
  • Social isolation, about 3.7 years lost. Holt-Lunstad et al., PLoS Medicine 2010.
  • A western diet against an optimal one, a spread of about 7 years. Fadnes et al., PLoS Medicine 2022.
  • A sedentary life against meeting activity guidelines, about 3.4 years. Moore et al., PLoS Medicine 2012.
  • An irregular sleep schedule, about 2 years. Windred et al., Sleep 2024, where regularity predicted mortality better than duration did.
  • Vitamin D and omega 3 supplements, zero. Manson et al., NEJM 2019, the VITAL trial, found no mortality benefit.

These figures cannot simply be added. A person who smokes, eats badly, never moves and sleeps four hours is not missing thirty years, because those habits overlap and the research measures each one against a population that also contains the others. Handling that overlap correctly is most of the work, and it is covered on the levers page.

What no calculator knows

It does not know your genome beyond a crude question about how long your parents lived. It does not know your bloodwork. It does not know about the accident, the diagnosis nobody has made yet, or the drug that will exist in 2043. A death date calculator is a mirror held up to your current habits, and its real value is that it makes an abstract trade concrete: this behaviour, on average, costs this much time.

That is also why it must never be read as medical advice. Yorick states this on the result screen itself, not only in the small print.

How Yorick answers the question

Nineteen questions, two to three minutes. The estimate starts from the SSA 2022 life table, shifts to your country using United Nations data covering 234 countries and territories, and is then moved by your answers using effect sizes from named studies. Every factor in the app carries a tappable link to the study on PubMed. Your date, your age and the likely range are free, with nothing blurred.

Download Yorick on the App Store

Free download for iPhone, iOS 26.0 or later.