Google regularly crawls websites. They know if a site still exists. It's not a 5 year old result, it's a website that has existed at least 5 years.

These pages die from starvation. If people can't get to your page even if they specifically google the name of your page, eventually you'll shut it down.

I understand why pages disappear. I'm just saying the expecting a search engine to return a result from 5 years ago is kind of a bad metric to be using in whatever point was attempting to be made.

I don't want results to be dropped arbitrarily. It kills all of the good content. I don't need a book to change every week to remain relevant.

edit: and to directly address what you said, excusing a search engine's loss of content because sites have disappeared because the search engine dropped them is circular.

I thought about this too.

After testing with marginalia I found that much of it is still there - it is just suppressed by Google and Bing.

That is interesting. I would have imagined honest link rot being a much larger problem.

My feeling is that the Lindy effect largely holds: The expected lifetime of any given document is proportional to its age. Old websites for the most part seem extremely stable. Young websites change drastically and often.