Joe HallGeverifieerd account

I wonder if the data would shake out holding this as true as Google adopts more and more of a "user-centric" approach. How many users are going through the site map to get to the pages they want?

Not sure, but from a purely link based analysis its always true if the two have been linked the same. I run most of the crawls I do through internal PageRank and they are almost always the same internal PR.

But it is questionable how many site map pages have external links from other domains (though they do have internal Pagerank.) #VCBuzz

Yes that is true. Practically no one will link to an HTML site map.

Though sitemaps receive a lot of internal links, they are often "boiler plate" links, and I don't think G passes a lot of value to them? (Things like privacy, terms, sitemap etc. never seem to rank highly, despite internal-popularity and branding)

The reasonable surfer model tells us that certain anchor text, such as "terms of service". Will not receive much weight at all. That is likely true with the other pages that you mentioned also.

I stopped using html site maps ages ago. I remember simulating the effect of one on a pagerank distribution tool and they can completely change the PR distribution and siloing.

They can be a useful tool for the same reason.