Let me echo the above sentiments about failure with advantages. Generally I adjudicate it such that they don't get the answer to the question they asked, but they get some equally (potentially more) valuable information that is tangentially related. In the example of the ship type, they can't identify it, but they know it's not heavily armed, lightly crewed and very fast. One useful piece of information per advantage (that doesn't answer the question asked).
As for difficulties, Whafrog had the right of it. If they're supposed to succeed, it should be 2 dice smaller than the pool the players are throwing. If it's to be a challenge, 1 die smaller. If it's a toss-up, same size. I suggest discounting setback and boosts from these equations. Those are supposed to add/remove difficulty situationally and the underlying difficulty of the check should be unchanged.
This gets back to my initial guide pertaining to character creation where I very emphatically suggest that players figure out a way to start with a 4 in their primary stat. That way they can handle average checks with some regular expectation of success, even if it's with threat.