Life expectancy is such a consistently deceptive measurement that I often wonder why it's not been replaced. Surely there must be more useful ways of quantifying lifespans, something like a "median age of currently living people"?
Point estimates in general are deceptive. It's difficult to capture the nuance of a full distribution with a single number.
I've had some luck with upper and lower percentiles, e.g. 5 % and 95 %. These cover 9 out of 10 people, which is large enough to be meaningful, yet not so small it's difficult to estimate.
