Life simulation / serious game
Data-Driven Life Simulator (Birth Lottery)
Tested with Claude · produced 1 in 8 Billion
The Prompt
Build a free browser life simulation in which the player is born into a random country and lives from age 0 to death. Single-page web app, TypeScript preferred, no 3D, no accounts, no runtime LLM calls. Data model: - A JSON dataset for at least 100 countries. Per country: annual births, life expectancy by sex, infant and under-5 mortality, GDP per capita (PPP), Gini coefficient, mean years of schooling, university enrolment, agricultural employment, fertility rate, maternal mortality, HIV prevalence, homicide rate, smoking rate, out-of-pocket health spending, child marriage rate and life satisfaction. Cite the source and vintage for every field. - Make the country draw weighted by annual births by default, with a toggle for population weighting. Make the data year selectable, including one historical year and one UN projection. - Simulate a life table by sex and country for the age of death. Sample causes of death from WHO regional patterns by age band; never invent precise individual predictions. Game loop: - Five lifecycle stages. At ages 5, 15, 30, 50 and 70, pause and show a reflection prompt about the simulated person and the player. - Each year, the player picks a focus: school, work, family, health, migration or rest. Decisions about education, marriage, moving abroad and medical care appear at plausible ages. - Choices shift probabilities, but do not erase the starting country’s baseline. Show the trade-off explicitly rather than pretending agency is unlimited. - Simulate four other babies born in the same second as a comparison cohort, with their outcomes visible beside the player’s life. - Auto mode plays a full life with no input. A full run should take 30-90 minutes at normal speed. Safety and tone: - Rate 15+. Death, violence, addiction and suicide occur at real statistical rates; never show a method and always surface a help link in the relevant region. Add a “leave out frightening scenes” toggle that preserves the statistics without graphic text. - No score, no leaderboard, no “good ending”. The point is comparison, not winning. UX: - Mobile-first, high contrast, readable at 360px. Save the current life and the country the player was actually born in locally only. Export a share card with the life summary, no personal data. Testing: - Run 10,000 simulated lives headlessly and report the distribution of ages at death by country against the source life expectancy. Flag any country whose simulated mean is off by more than 2 years. - Print the data sources, known gaps and the exact assumptions used to convert country statistics into individual life events.
What This Prompt Produces
1 in 8 Billion is the reference implementation: 172 countries, UN/World Bank/WHO data, four parallel lives and five pause points, built in about four hours with Claude. The emotional weight comes from restraint, not graphics — no 3D, no score, no runtime model.
Modification Tips
- Start with the data table and the life table; the game is the data. A beautiful UI over invented numbers is worse than a plain one with cited sources.
- Simulate a comparison cohort. One life is a statistic; five lives born at the same second make inequality visible.
- Put content warnings and help links in the data layer, not as an afterthought in the UI.
- Run the 10,000-life test before launch. Small sampling bugs produce countries where everyone dies at 40.
The Game This Prompt Produced
Want to try it? Open Claude, paste the prompt, and you'll have a playable game in minutes. When you're happy with the result, submit it to Daily AI Games.