Agreed that not all years are equal, and the impacts of early deaths can be large.

For measuring N, I wonder how much of this is availability bias. I know a number of old people who have outlived half of their friends, as well as seeing a few friends lost early in life. My weak suspicion is that they would favor a lower N, just by more familiarity with death due to old age and a better idea of what old age deaths do to the culture / friend groups / family.