Last night I ran into a problem that took way longer to solve than I expected. The delay had two causes: I was working from an incorrect assumption, and I couldn't test each case of the solution space generated by my bad assumption. Because I was uncertain of how much of the solution space I had searched, I undoubtedly reduplicated efforts, which added delay. When I guessed I had searched enough, I examined my assumptions and found the error. All this added up to a big bunch of inefficiency.
Clearly, I need to work on my problem solving skills. In the future:
The probability level should be determined by cost to evaluate remaining cases versus cost to check assumptions. Both will be estimated costs; good luck. ;)