> you have chance 1 in 1000-2000 that you'll die next year.
You have a 1000-2000 chance that you'll die when you are no longer an infant but still young. Your chances of dying next year go up as your get older. If you had only 1 chance in 1000 to die each year that would mean that people would live to be 1000 years old on average.
Your chances of living to be 1000 are approximately 1 in e, 37%, yes. But that has nothing to do with the average age people live to. I don't get 1000 as the average lifespan, though, I get 999. Can someone explain where I'm going wrong with this reasoning?
If there is a 0.001 chance of dying every year, the average lifespan will be 0(0.001) + 1(0.001)(0.999) + 2(0.001)(0.999)^2 + 3(0.001)(0.999)^3 + ... , or \sum_{i=0}^\infty (i(0.001)(0.999)^i). The idea is that to live to the age of 0, you have to die in your first year; to live to be three, you have to live through exactly three years and then die, and so on. Then, \sum_{i=0}^\infty ix^i is x / (1-x)^2 .
So the sum of interest is 0.001 times (0.999) / (0.000001), which is exactly 999.
You have a 1000-2000 chance that you'll die when you are no longer an infant but still young. Your chances of dying next year go up as your get older. If you had only 1 chance in 1000 to die each year that would mean that people would live to be 1000 years old on average.
A good estimate of human mortality rates can be calculated from Gompertz law: http://forio.com/simulate/mbean/death-probability-calculator...