Kofi Osei
male1959 – 2026 · Ethiopia
Road injury
67 · Made it most of the way.
v1.1 · free · reverse life simulation · 173 countries · 20–60 min, skippable
“You think
life is
beautiful,
don’t you?”
Death is the only certain thing. Everything else is probability — where you are born, who you meet, how long you last. So here is the certain end, handed to you first. Then walk back. One year. One more. See which steps were yours, and which ones carried you.
0
people who died in the same time
0
babies born since you opened this page
Today's chainCardiovascular disease · about a third of all deaths
The numbers are real. The story is not.
one deatheach ring = one birth
At the end of a life — where would you land?
About 61 million people die each year, worldwide
Source: World Bank WDI 2023 (population × crude death rate)
China18.3%
India15.6%
United States5.1%
Nigeria4.4%
Indonesia3.5%
Russia2.9%
Japan2.7%
Pakistan2.6%
Brazil2.5%
the other N42.5%
weighted by deaths173 countries →
China, male, age 40 — 600 draws · 26 cause chains— three most common20–60 min per run
27%Cardiovascular disease
One of the most common ways to die here.
And this layer was locked long ago.
19%Lung cancer
One of the most common ways to die here.
And this layer was locked long ago.
11%COPD
One of the most common ways to die here.
And this layer was locked long ago.
one more life→
How it works
Few things can be changed.
And every one of them costs something.
01
The death date isn't random
It comes from your country, sex, current age and four habits. Life expectancy is the baseline, habits add or subtract, and undiagnosed illness is a hidden deduction.
02
Every layer has odds
The changeable layers carry a success rate. Gamble once: win and the date moves back; lose and it costs you years.
03
Some illnesses stay hidden
You may carry high blood pressure or diabetes. Without a check-up you won't know. It keeps deducting years until the ending tells you.
Steps
01
Birth
Weighted by births per country: a country, a sex, an age.
02
See the ending
First how many years are left, then what kills you.
03
Walk back
Year by year, back to birth. The changeable layers show their odds.
04
Take the record
A death certificate and your place in the cohort — save it as an image.
Random ending →No sign-up, no storage. Everything is computed in your browser.
Gallery
A few lives that already ended.
All generated by the model from real statistics. Not real people. Open to see all 24.
Huilan Li
female1951 – 2025 · China
A fall
74 · Outlived most.
Omar Hassan
male1955 – 2024 · Egypt
Cardiovascular disease
69 · Made it most of the way.
Linh Pham
female1948 – 2023 · Indonesia
Stroke
75 · Outlived most.
How to play
Six steps, one life.
Each step maps to something on screen. Read these six once and nothing will stump you.
01
Birth: not your choice
Weighted by annual births, the system draws a country, a sex, an age. India and China come up most; Iceland and Mongolia rarely — few children are born there. Then you see the ending first, and only then walk back.
On screen: “Random ending” on the start page
02
Ending: how you die, first
The date comes from that country's sex-specific life table, adjusted by your habits (smoking, few check-ups, late nights). The cause is drawn from 26 chains, weighted by real parameters: malaria where malaria is heavy, lung cancer where smoking is high.
On screen: “Cause” and “same cause” on the ID card
03
Walk back: one year at a time
Time runs backwards. Each year you step back adds a line to the log — where you were, what you did, who was with you. A farmer reads "you borrowed money for two pigs"; an office worker reads "you were overruled in a meeting".
On screen: “Life log” in the middle column
04
Cut: only a few layers
Layers marked with a success rate can be gambled once. Win and the date moves back; lose and it costs you years. About six such chances in a lifetime. Some layers are already locked.
On screen: the fork pop-up
05
Write a line: eight moments
At 5, 10, 15, 30, 40, 50, 60 and 70, time stops to ask you a question. Every line you write goes into the ending.
On screen: the question that pauses time
06
Pace: your call
Choose 4 / 20 / 40 / 60 minutes. At 40 minutes a year takes 30 seconds — slow enough that your mind wanders, and that is when you think about yourself. Pause or speed up anytime.
On screen: speed and “Next event” at the top
First time? Remember one thing:You are not playing a role. You are watching a life that has already ended. Very little can be changed — that is the whole point.
Data
Where these numbers come from
173 countries, 26 fields, 26 cause chains, 12 accident types, 795 life events — all derived from public statistics. The date comes from sex-specific life tables; cause weights are computed from each country's parameters — not looked up, not random.
Not medical prediction, diagnosis or health advice.
Life expectancy · under-5 mortalityWorld Bank WDI 2024
Urbanisation · out-of-pocket health spend · years of schoolingWorld Bank WDI 2024
Smoking · road deaths · maternal mortalityWHO · World Bank
Cause weights · accident oddsModelled from country parameters (not external data)
Income · savingsIncome = 90% of World Bank GDP per capita · savings modelled at 12%
173
countries
26
fields
26
cause chains
1630
life events
Fields that actually enter the model
Each country has 26 fields, all from public statistics. These are the ones that actually enter the model:
- Births per year
- decides your chance of being drawn
- Life expectancy by sex
- decides roughly how long you live
- Infant and under-5 mortality
- decides whether you reach five
- GDP per capita (PPP)
- decides income, accident risk, odds of leaving
- Gini coefficient
- decides unemployment risk
- Total fertility rate
- decides how many children you have
- Urbanisation
- decides city or countryside
- Share in agriculture
- decides occupation and work-injury risk
- Maternal mortality ratio
- decides extra risk for women aged 15–45
- HIV prevalence
- weights the HIV chain
- Malaria burden
- weights the malaria chain
- Homicide rate
- decides homicide and prison odds
- Road death rate
- decides yearly traffic risk
- Conflict intensity
- decides war and displacement
- Smoking prevalence
- weights lung cancer and COPD
- Out-of-pocket health spending
- decides the cost of not being diagnosed
Is the model any good
Same countries, 240 draws each, next to the real figures. Close — but not the same.
| Country | Share of births | Simulated mean lifespan | Actual life expectancy |
|---|---|---|---|
| India | 17.6% | 72.6 | 72.2 |
| China | 6.7% | 78.3 | 78.0 |
| Nigeria | 5.7% | 56.0 | 54.6 |
| Pakistan | 5.2% | 69.0 | 67.8 |
| Indonesia | 3.4% | 72.1 | 71.3 |
| DR Congo | 3.3% | 63.9 | 62.1 |
240 lives per country (half men, half women, no habits, starting at 30). Simulated lifespan runs slightly above actual because infant deaths are not counted here.
About
This is a simulation
The death date, cause, family and income are all computed by a model — not a prediction about any specific person. It does one thing: translates a statistical table into the second person.
On death and violence
It covers illness, accidents, suicide, homicide and war. No gore, no depiction of method. If you are going through something hard, this is not for you right now — take care of yourself first. You can tick “Skip uncomfortable content” at the start.
On originality
All text, illustration, music and code here is original. Data comes from public UN, World Bank and WHO statistics. The idea of being born and living a life by real statistics belongs to no one — we wrote our own version.
Who made it
Made by one person and one AI. Data from public UN, World Bank and WHO statistics; model, copy, illustration and music all written here. Free, no sign-up, no data collection.
This page is not medical prediction, diagnosis or health advice. All figures are simplified citations of public data and may contain error. Teachers may use it in class; to reduce adult content, use “Skip uncomfortable content”.