How do search engines figure out the date of webpages that don't contain it in the metadata?
Poorly.
I have a blog so old I titled it an "online diary." It pre-dates search engines, so they tend to date the diary entries (blog posts) based on first crawl. Which means lot of the dates presented by the search engines are off by several years.
https://duckduckgo.com/?q=wordpress&t=h_&df=2000-01-01..2005...
I'm guessing you were looking for:
https://wordpress.org/news/2004/01/cruft-free-uris-in-wp-10/
I assume Google supports something similar, but I've stopped using it.