The request was "a game that quickly sketches what kind of life we'd live, by the odds, if we were born." You can pick an era anywhere from 100,000 years ago to today, and the condition was that everything — where you're born, your sex, your family, the house you live in, all the way to when you die — had to come out close to reality. I made it play out automatically with no choices. It's a game that shows a life decided by probability, so the only thing the user gets to choose is the era. It's at ghmun.com/born.
Where you're born
About 117 billion people have been born so far (Population Reference Bureau estimate). When you pick an era, the birthplace is drawn according to each region's share of the children born in that era. Pick 1700 and one in three is born in East Asia, home to Qing China, Edo Japan, and Joseon Korea; pick 2025 and close to one in three is born in sub-Saharan Africa. I split the world into 10 regions, and within each region a specific birthplace like "Jiangnan, Qing China," "Joseon," or "Uttar Pradesh, India" is drawn again in proportion to population. For places where the state changes within an era, like the Yuan and Ming dynasties, only the name that matches the birth year comes up.
Before 1950, the shares are based on population estimates from HYDE 3.2 and McEvedy & Jones, adjusted for differences in birth rates; from 1950 on, they're calculated by grouping country-level births from the UN World Population Prospects (WPP 2024) into regions.
A life expectancy of 30 doesn't mean people died at thirty
Before the modern era, life expectancy at birth was around 25 to 30. The number is low because 40–50% of children died before age five. People who made it past five commonly lived to 50 or 60. Run the game a few times and you'll often get "Died at 8 months of diarrhea and dehydration." That's what the real rates look like.
So mortality isn't set by a single average lifespan. For each era and region, you enter just two numbers, the under-five mortality rate (q5) and life expectancy at birth (e0), and the engine fits the coefficients of a Siler model (a formula that adds three terms: infant mortality, adult background mortality, and aging mortality) by bisection to build an age-specific death probability table from 0 to 110. Each person's survival is decided against this table every year. Risk is multiplied up or down a little by class (slave, tenant farmer, landowner, and so on) and sex, and women roll a separate risk of dying in childbirth every time they have a child.
While a person is alive, the table used isn't the one for their birth era but the one for the era of that calendar year. Someone born in 1950 gets the medicine of the 1980s and 2000s, so they live longer than the 1950 life expectancy (53). Improvements in mortality after 2025 are not included.
I redefined each era's world life expectancy
At first I took each era's world life expectancy as the population-weighted value for the first year of that era. When I ran the numbers on modern data, it didn't add up. This game draws "a person born in that era," so the baseline has to be the era average weighted by number of births. World life expectancy in 2025 is commonly cited as 73, but weighted by births it's about 69. That's because 31% of children born today are born in sub-Saharan Africa.
What goes into a life
- Family. Your mother's age, how many siblings you have and how many of them died young, and when your parents die.
- Work. You start work suited to your class and sex at a set age. In hunter-gatherer societies it's from around age 6; in high-income societies, from around 20.
- Marriage and children. Age at first marriage and number of births follow the type of society. Each child gets its own age of death, so if a child dies young, it shows up on the timeline that year.
- 33 historical events. The Black Death, smallpox in the Americas, the Imjin War (the Japanese invasions of Korea in the 1590s), the Thirty Years' War, both World Wars, the Korean War, the famine after the Great Leap Forward, North Korea's Arduous March famine, the Khmer Rouge, HIV, COVID-19, and more. If you're alive in the affected region during that period, you experience it with some probability, and die of it with some probability.
- Daily life. For each era and region there are 3–5 sentences about food, sleeping, faith, and work. They don't appear in years earlier than the year written in the sentence, and age-bound sentences like "went to cram school" appear only at that age. If you're born in Silla, Goryeo, Joseon, Japanese-occupied Korea, South Korea, or North Korea, Korean Peninsula sentences come first.
Verification
tools/born/check.js checks the following: whether the computed death probability table reproduces the input q5 and e0 exactly for all 160 regional rows; whether the regional shares sum to 1 and the weighted average matches the era baseline; whether, simulating 20,000 people per era, life expectancy, under-five mortality, and regional proportions match the inputs; whether the place names in the event table exist among the actual birthplaces; and whether any generated sentence has an empty value.
Class, occupation, housing, and daily-life descriptions were chosen to reflect what was typical for each era and region, and the figures are estimates. Regional shares and mortality before 1850 in particular vary widely between studies.