I have read, here on lesswrong, that aging may basically stop somewhere around age 100: Your probability of death reaches 50% per year and doesn’t go much higher; the reason people don’t live much over 100 is just the improbability of the conjuction 50%*50%*50%, etc.
However, this table from the SSA seems to directly contradict that. Now I’m wondering what explains the seeming contradiction.
I’d point out that one would expect the SSA tables to overstate the number of centenarians etc, for the simple reason that they are linked to financial payments/checks. Japan recently had some interesting reports that its centenarian numbers were overstated… because other people were collecting their pension checks. From the BBC:
More than 230,000 elderly people in Japan who are listed as being aged 100 or over are unaccounted for, officials said following a nationwide inquiry. An audit of family registries was launched last month after the remains of the man thought to be Tokyo’s oldest were found at his family home. Relatives are accused of fraudulently receiving his pension for decades...Reports said he had received about 9.5m yen ($109,000; £70,000) in pension payments since his wife’s death six years ago, and some of the money had been withdrawn.
...Officials have found that hundreds of the missing would be at least 150 years old if still alive.
I have read, here on lesswrong, that aging may basically stop somewhere around age 100: Your probability of death reaches 50% per year and doesn’t go much higher; the reason people don’t live much over 100 is just the improbability of the conjuction 50%*50%*50%, etc.
However, this table from the SSA seems to directly contradict that. Now I’m wondering what explains the seeming contradiction.
FWIW, from a histogram I quickly made from the data in http://en.wikipedia.org/wiki/List_of_the_verified_oldest_people it doesn’t look like the probability of surviving to age x falls any faster than exponentially.
It could be that the table is not empirical past 100.
Maybe. But if anybody had empirical data on old people, I would expect it to be the SSA.
I’d point out that one would expect the SSA tables to overstate the number of centenarians etc, for the simple reason that they are linked to financial payments/checks. Japan recently had some interesting reports that its centenarian numbers were overstated… because other people were collecting their pension checks. From the BBC: