Anyway Firefox did make some effort to work around this common web design error (that being what I'd call it) which is probably why it behaves differently to Chrome: https://bugzilla.mozilla.org/show_bug.cgi?id=780345 but I still find it hit and miss between from one page to the next. As you mention there are many scenarios - what if the fixed header only goes across say half the width of the page? What if it's less than 100% opacity? What if it varies in height? The browser makers have to resort to "heuristics" and now we're off into the land of deep unpredictability.
While writing this I've also just noticed that - fixed header or no - pressing space / page down doesn't go exactly the same distance (in Firefox) as paging down by clicking on the scroll bar. Which is also a bit disconcerting.
http://consequenceofsound.net/2014/05/stream-robyn-and-royks...
http://www.thewire.com/culture/2014/05/7-eleven-just-killed-...
But I'm not sure exactly what the mechanism is that causes this; I thought it was simple a case of forgetting to add top padding to the page body, but that doesn't always seem to be the case. Last time I saw a site that did this, it already had the padding but still misbehaved in Chrome. Yet the paging worked correctly in Firefox! Further testing showed Safari worked, so (in this case, at least) it seemed to be webkit-related.
But that's just one scenario. I suspect there's a number of ways people get this wrong, and it's unfortunately not just a single simple fix for all cases.