Calling it a "mystery" gets suckers like me to click the link.

Yeah, it wasn’t a real mystery (as noted in the Wikipedia article, it already existed in the numerical literature). But practically it would have surprised a lot of programmers in the days before Wikipedia, when you’d have had to read a somewhat specialized textbook or a paper to learn about it.

Plus the exact constant selected and the method used to derive it remains a minor mystery, right? In the sense that it is good but non-optimal.

The "mystery" being referred to in the title is how the magic number was derived. This is what most of the article talks about.