Evidence directory
Longevity interventions in the site’s house currency — years of life expectancy — with the confidence honesty the rest of the internet skips. The data supports what it supports.
The single most reliable dial you control: moving from unfit to fit buys years, and the biggest jump is off the bottom of the distribution.
≈ +4 years for the sedentary→fit shift, smaller with age
The largest single modifiable mortality effect we know of, and one of the few where the recovery on quitting is almost as dramatic as the damage.
~10 years lost by current smokers; most recoverable if you quit before 40
A J-shaped mortality curve where the cost of high BMI is real and graded, and the apparent benefit of being slightly heavy is mostly sick people who lost weight.
J-shaped; roughly −2 to −8 years across the obesity bands vs normal BMI
A U-shaped curve where both ends cost you, but the long-sleep arm is largely sick people sleeping more, not sleep making them sick.
U-shaped; short and long sleep each cost ~1–2 years
The famous J-curve was mostly a statistical mirage; the honest reading is that risk rises with dose and the "healthy glass of red" was wishful thinking.
heavy drinking ~−3 years; the 'moderate benefit' is largely confounding
A genuinely effective drug class whose mortality benefit is real, modest, and mostly routed through weight and metabolic risk the calculator already prices — so we do not double-count it.
no direct life-expectancy RCT yet; cardiovascular-mortality signal (SELECT trial) is real but modest, and it acts largely THROUGH weight and metabolic risk we already count
The most interesting molecule in the longevity conversation and the one with the least human evidence — genuinely impressive in mice, entirely unproven in people, and given zero calculator years on purpose.
no human life-expectancy evidence; strong in mice, unproven in people — shown here as directory content, NOT as a year-effect in the calculator