Private by architecture: everything on this page is computed inside your browser. Nothing you enter is sent, stored, or logged — the page makes no network requests. Close the tab and it is gone.
What this is and is not: population statistics applied to one person. It cannot know your genome, your luck, or your future. The honest reading of the result is a central estimate with a wide range — not a prophecy, and not medical advice.
1. The base is the UN WPP 2024 World life table (2023,
single ages) — the same committed, hash-manifested input the index
settles on (fixture sha256 fbcbd9c9fe900408…). It gives q(a):
the probability of dying within a year at each age a.
2. Your answers each carry a published relative risk — a hazard multiplier from the epidemiology literature (source noted on every option). They multiply into one composite m, clamped to [0.35, 4.0] because questionnaire risks multiply badly at the extremes and honesty beats false precision. Sex uses the typical published mortality ratio (male ×1.28, female ×0.76) against the both-sexes table — an approximation, labeled as one.
3. Your adjusted hazard is q'(a) = min(1, m·q(a)). Walking it forward from your age gives your survival curve; summing survivors (a death counts half a year — the trapezoid rule) gives expected remaining years. The 10th and 90th percentiles are read off the same curve — that spread is the honest headline, not the single number.
4. Multiplying your remaining years by nothing gives your stake in the index: S is simply this number summed over every living person on earth. One token = one life-year of that total.