Is the Death Clock Real? The Model Behind the Number
This death clock is a real actuarial calculation: a published survival model applied to public life tables, which returns an average for people with your inputs and not a prediction for one person.
This death clock does not state when any one person dies: this page answers whether a death clock is real or fake by showing the method, never by an accuracy percentage. Every death clock result is computed by the equation and tables described on this page.
- Author
- Jacob V.
- Last reviewed
- Next review
Reviewed against the published methodology and UN World Population Prospects 2024.
Is the death clock real or fake?
A death clock is real as arithmetic and not real as prophecy: it computes an average age at death from life tables, and it cannot know when any one person dies.
This death clock is evidenced by three public parts: the Gompertz–Makeham law of mortality (Gompertz 1825 and Makeham 1860, summarised on Wikipedia), the United Nations figure for life expectancy at birth for the 96 countries the calculator covers, and the calculator’s own constants.
This death clock does not state an accuracy percentage, a validation or the age at which one particular person dies.
Partly: a death clock is true as arithmetic on life tables and not true as a forecast for one person.
What the model computes and from which data
The model computes the result in four steps: it takes the Gompertz–Makeham hazard, calibrates it to the United Nations (UN) life expectancy at birth for the country and sex, conditions it on the attained age, and applies the lifestyle adjustments.
- Take the Gompertz–Makeham hazard, which is defined as the hazard rate
μ(x) = A + B·e^(b·x)at age x (Gompertz 1825, Makeham 1860). - Solve B for each country and sex, so that B is calibrated to life expectancy at birth in UN World Population Prospects 2024, within 0.001 years.
- Read the survival curve from the attained age, so that the rest of the curve is computed from the age already reached.
- Add the lifestyle adjustments after the country-and-sex baseline: the result is adjusted by body mass index (BMI) and smoking from named studies (Lancet 2016, and Doll and Peto, BMJ 2004), then by alcohol, outlook, activity and diet as calculator weights.
The model returns the result of the last step: the country-and-sex baseline shifted by the lifestyle adjustments.
The first step is the Gompertz–Makeham hazard equation, with its three constants.
The model is a published equation, not a trained model, and the page on AI death predictors explains the difference.
Why no accuracy percentage is published
No accuracy percentage is published because the model’s fit to individual outcomes has not been measured, and a figure without that measurement would be invented.
The middle-half range, which the calculator publishes instead, is defined as the 25th to 75th percentile of age at death on the same survival curve that gives the result, shifted by the lifestyle adjustments.
The middle-half range is bounded by two ages: a quarter of people with the same inputs die before the lower age, and a quarter die after the upper age.
The middle-half range does not state how sure the clock is about one person: it describes how ages at death spread among people with the same inputs, not a forecast for one person, and it is not an accuracy figure.
The model does not read sleep, disease, medication, family-history or genetic input, so the result cannot reflect any of them for one person.
Two clocks rarely agree, and the reasons are set out in why death clocks disagree.
What is the formula for life expectancy?
Life expectancy at age x is the total years lived beyond age x divided by the number of people alive at age x: e(x) = T(x) ÷ l(x).
In the life expectancy formula, each of T(x), l(x) and e(x) is defined as one quantity, with T(x) the total years lived beyond age x, l(x) the number alive at age x and e(x) life expectancy at age x.
At x = 0, e(0) equals life expectancy at birth, which in the 2023 period life table was 76.9 years for US men and 81.8 years for US women (UN World Population Prospects 2024, 2023 life table).
The baseline is calibrated to UN life tables, which give the probability of dying at each age.
The equation in full
The equation behind the curve is μ(x) = A + B·e^(b·x).
The full form, with each constant and its value, is the Gompertz–Makeham hazard equation page.
What this clock is not: AI, astrology, novelty
This clock is not three things: an AI predictor, an astrology or numerology reading, or a novelty clock.
- An AI predictor differs from this clock, which applies the Gompertz–Makeham law of mortality (Gompertz 1825, Makeham 1860), a published equation, not a trained model.
- An astrology or numerology reading differs from this clock, whose survival curve is calibrated to the UN figure at birth for each country and sex.
- A novelty clock differs from this clock, which states a middle-half range around its result.