I don't think there's any sure way to know whether you'll have experimental questions or even if a particular question on your exam is experimental.
I also think that there probably is a tendency to speculate that a question one simply doesn't know is experimental.
That all being said, I can certainly imagine certain types of things happening on your exam where you have a strong suspicion the question was experimental. Obviously it would be stupid not to try on any given question on the wager that it's experimental, but it can be a justification for a testing strategy that simply does not waste time on questions that you don't know or that would take an inordinate amount of time to figure out.
The types of things I can imagine (of course I haven't taken the test) seeming experimental would be: strange new ways of interfacing with questions that you can find little or no mention of other people encountering in the past; questions about strange topics to which medical students are not routinely directly exposed (this is probably more out there than what I'm really imagining, but an example might be if a question seemed to require superficial but specific knowledge regarding comparative politics related to health care law), etc.
I don't think there's ever a way of really knowing whether you will get or have gotten an experimental question, but it's possible that people in the past have had their suspicions confirmed by later changes to the test content or structure. For instance, I bet you that within the last few years there have been people who got questions about the new "safety" stuff that they thought were kind of odd. They can probably rest assured now that those were experimental.