This seems to match my experience. I'll pile on a hypothesis of my own: When we encounter a new problem, our brain spans a decision tree of assumptions from th

This seems to match my experience. I'll pile on a hypothesis of my own: When we encounter a new problem, our brain spans a decision tree of assumptions from the closest known solution to a similar problem. The area around this tree is searched depth-first, which often brings good results (e.g. thinking about whether you'll put some threshold value at 0.2 or 0.25), but it's best at refinement. The more time passes and the more different things you do, the more the search is moved into breadth-first (which makes intuitive sense - the deeper the tree, the more processing power would have to be put towards it). Often times we literally cannot conceive a good solution to a problem, because our brain is focusing on finding the right deep leaf of the wrong subtree!