How Mortmeter works
Every answer is one number, worked out one way: the chance an activity kills an average adult. Your age then turns that into time.
Two units
Micromort: a one-in-a-million chance of dying. The unit comes from decision analyst Ronald A. Howard (more on micromorts). Mortmeter uses it for sudden, one-off risks: a skydive is about 8.
Microlife: thirty minutes of life expectancy, from statistician David Spiegelhalter's "Using speed of ageing and microlives…" (BMJ, 2012). It's for slow, daily habits; Mortmeter uses it for one comparison: a cigarette costs about 15 minutes.
The odds
Each answer is priced in micromorts for an average adult. 8,000 micromorts is 8,000 in a million, or 1 in 125. That's the big number on every card, and it's the same for everyone who sees it.
What it costs you
Expected life lost = the chance of dying × the years you have left.
Kickflipping the Roadster at 45, for an average person worldwide: 8,000 ÷ 1,000,000 × about 33.7 years left = 14 weeks of life, on average. At 20 it's 24 weeks; at 75, 4 weeks. The odds don't change with age; the stake does. Most likely you lose nothing; rarely, you lose everything.
Your age and profile never leave your browser. The answer is priced once, and "what it costs you" is worked out on your device.
The life tables
Years left come from the WHO's life tables for 2023 (Global Health Estimates): how many people out of 100,000 are still alive at each age, for men, for women, and for both together ("neither, or skip"). Mortmeter has 31 countries and the world average. Pick yours in My meter. Until you do, it guesses from your browser's language and says so; anywhere not listed gets the world average.
WHO's tables are read every five years and smoothed in between. They stop at "85 and over", so 90 to 100 are extended: death rates keep rising at the pace they do in the late 70s and early 80s, set so that life expectancy at 85 matches WHO's. The years left at 20, 45 and 65 come out within about a quarter of a year of WHO's own figures for every country here.
Years left at 45
| Men | Women | |
| Argentina | 32.3 | 36.5 |
| Australia | 37.6 | 40.6 |
| Austria | 35.8 | 39.7 |
| Belgium | 36.4 | 39.7 |
| Brazil | 32.4 | 36.3 |
| Canada | 36.6 | 39.7 |
| China | 32.1 | 37.2 |
| Denmark | 36.3 | 39.4 |
| Finland | 35.2 | 39.3 |
| France | 36.4 | 40.7 |
| Germany | 35.0 | 38.9 |
| India | 29.2 | 32.2 |
| Ireland | 36.8 | 39.5 |
| Israel | 37.6 | 40.9 |
| Italy | 37.1 | 40.4 |
| Japan | 37.2 | 42.1 |
| Mexico | 31.9 | 35.6 |
| The Netherlands | 36.3 | 38.7 |
| New Zealand | 35.9 | 38.2 |
| Nigeria | 28.9 | 30.7 |
| Norway | 37.4 | 39.7 |
| Poland | 31.6 | 37.6 |
| Portugal | 35.3 | 40.2 |
| Singapore | 37.3 | 41.0 |
| South Africa | 25.7 | 31.7 |
| South Korea | 36.7 | 42.1 |
| Spain | 36.8 | 41.2 |
| Sweden | 37.5 | 40.0 |
| Switzerland | 38.2 | 41.0 |
| The UK | 35.5 | 38.5 |
| The US | 34.4 | 37.6 |
| World average | 31.6 | 35.8 |
Your profile
On top of your country's table, My meter scales the death rate at every age by two multipliers. Smoking has a solid source; fitness is a rough estimate.
Fitness
| Couch | ×1.3 |
| Average | ×1 |
| Active | ×0.8 |
| Very active | ×0.65 |
Smoking
| Not saying (population average) | ×1 |
| No | ×0.8 |
| Quit | ×1 |
| Yes | ×2.3 |
Current smokers die at about 2.8 to 3 times the rate of people who never smoked (Jha et al., NEJM 2013), scaled here so "not saying" is the population average.
How an answer is made
A new question costs one AI call, ever. The model (currently gemini-3.8-flash) gets the question and the Atlas, 87 sourced figures it uses as reference data, and works the price out as a sum: how much of the activity it assumed, times a risk per unit. It also gives a range, the assumption, what gets you, a verdict and a cheaper version. We check the numbers (a range in order, never above a million micromorts, which is certain death) and save the answer at its own page. Ask the same question again and you get the same answer.
Each answer is labelled with how sure it is:
- Measured: Counted deaths over a counted number of people, trips or hours.
- Estimate: Worked out from related data, not counted directly.
- Best guess: Little or no data, so reasoned from the closest thing we know.
Tap the label on any card to see its reasoning. The comparisons ("Like 19 BASE jumps") are worked out from the Atlas, not by the AI. The verdict is Mortmeter's opinion: how steep the price is, how rare the experience is, and whether a cheaper version gets most of it. The crowd's votes are shown next to it but never change it. Answers get checked by hand over time, and wrong ones are re-priced.
Known limits
- The AI can be wrong. The label says how sure it is; "Best guess" means take it lightly.
- An average hides a lot: experience, gear, weather and luck all move the real number.
- The activity's risk stays the same at every age here. In real life many things get riskier as you get older.
- Mostly sudden death, not slow harm: a binge's hangover and liver aren't in the price.
- Country tables are national averages for 2023, and the fitness multiplier is rough.
- Rare deaths mean small numbers, so some figures rest on a handful of cases.
Not advice
Mortmeter is for perspective, and for fun. It isn't medical, safety or financial advice. If you're weighing a real decision, talk to someone who knows you and the activity.