RCS: I'm willing to believe that smoking is bad for your health, but beyond that, it's something of a puzzle. If the death rate from smoking is 2.7x nonsmoking, does it follow that, since American life expectancy is 80 years, that smokers only live to be 30? WDS: If your death rate changes to 1/K>0 times its old value, you do not live K times the expected lifespan. That would be true if death rate independent of age, but it is not. In fact, between ages 30 and 95, death rate is very well fit by an exponentially increasing function of age, "Gompertz's law" and death rate at age 95 in USA is about 200X death rate at age 30. As a result, if your death rate at each age changes to 1/K times its old value, you will live, in expectation, 12.3*ln(K) extra years. My chart gives reason to believe it is feasible to decrease your death rate to about 26% of the average person's, thus living about 16 years extra.