Experience Bias (I don't know the technical term)
Despite having 3000+ apps available at launch, many from reputable developers and big-name companies, one of the main criticisms of 3rd-party Watch apps has been that they're slow, clunky, and buggy. Aside from the difficulties of not having access to the actual hardware, developers sharing their Watch app mistakes have attributed some missteps to making too many assumptions based on prior experience with iOS development for phones & tablets. They failed to alter their workflow or to really take into account many significant differences.
So designing for a "new" product or category can be hampered by habits and assumptions about how you'd design for an existing one. Just because something has been "the right way to do things" up until now, do not assume this will remain true. [Insert apocryphal/requisite Henry Ford or Steve Jobs quote here.]
Rumsfeldian Problem Domain Analysis
Even with new products there are known knowns, known unknowns, etc. There are assumptions, restrictions, and initial use cases that provide some boundaries and criteria to aid in design and testing. A critical requirement seems to be to clearly understand and force yourself to work within those boundaries, at least at first.
With WatchKit, developers have known for months that they would only have access to a limited subset of functionality and hardware; what the screen sizes, resolution, basic interaction patterns and storage restrictions would be, etc. All of this knowledge could be used to inform design and testing decisions. Be deliberate about asking and keeping track of what you know, what you assume, what you're clueless about, and use this information to refine and restrict what you can design and test.
Fake it until they make it
Try to reproduce conditions as accurately as possible. Work at scale, test prototypes at scale and in realistic environments. Sure, you can't really test a Bluetooth delay with paper, but you can at least take the time to tape UI sticky notes to a Timex and wander around (the 1st Apple Watch was allegedly an iPhone + a velcro strap). Despite wide availability of DIY UI mockups even before Apple released official ones, many devs still probably relied on iOS habits and the software simulator, which is too large, too fast, and too powerful to get accurate usability feedback.
Keeping focus on realistic use cases instead of relying on available tools can help you to at least build and prioritize a list of things to test more fully once the actual hardware becomes available. That way at least you're not starting from scratch & already have a head start.