A felicitous but unproved conjecture may be of much more consequence for mathematics than the proof of many a respectable theorem.
The thing is, it's very dangerous to have a fixed idea. A person with a fixed idea will always find some way of convincing himself in the end that he is right