I remember reading somewhere that good speculative fiction (including horror, sci-fi, fantasy, etc) takes a widely understood backdrop, whether that is modern day, Tolkien-style fantasy, or something else, and tweaks one, or at most a handful, of key factors, and then works out the consequences of the tweaks. I have forgotten the source, but the idea stuck with me.
Various traditions of vanilla fantasy serve as examples of potentially well-understood backdrops, depending on particular audience. The Tolkien-derivative or swords and sorcery in the Leiber/Howard style are two examples. There is something to the formula of a known baseline modified just enough to add interest without becoming overwhelming. Maybe unworthy of being considered an iron rule, but something.
This seems similar to the figure/ground distinction in Gestalt psychology. If everything is ground, nothing stands out, leading to boredom. If everything is figure, all is confusion and nothing makes sense. Also: to the neophyte, even vanilla fantasy can seem strange enough while the seasoned player may require excessive marginal weirdness to get a successful hit of strange. Of the people that play fantasy games at least partly for the pleasure of exploring an imaginary world, most seem to be between these extremes, wanting a bit more uniqueness than Tolkien sans serial numbers but a bit less than a setting which avoids all pop culture or mythological landmarks.