Mortality rates follow a Gompertz curve. This blog post explains it pretty well: http://gravityandlevity.wordpress.com/2009/07/08/your-body-w... There are many suspected specific reasons for why this is (telomeres shortening, mtDNA mutations, protein cross-links, etc), but it boils down to this: accumulating damage accumulates faster as self-repair mechanisms are damaged by accumulating damage.
That's not strictly true, you are more likely to die at age 2 than 10. And even if you have a ~50% chance of dieing at 100 there are people who make it to 120 or so.
About one in 45 billions. How many people in history did reliably make it to 120? exactly one, Jeanne Calment. One other did it to 119, and that's about it.