The structure of phd programs highly discourages doing important research. Important problems carry high risk of getting scooped, but do not offer proportionally higher upside. The rational course of action is to choose the most irrelevant and obscure topic possible so as to protect your 5+ year investment. Real research is the job of tenured professors.
If your goal is to get a Ph.D., then you are correct. If your goal is to start a research career, then it is not the rational course of action. The irrelevant and obscure thesis will not get you a research job.
Incidentally, getting "scooped" is not as bad as you seem to think. My thesis cites quite a few papers which "scooped" me, and the guy who hired me for my current job actually had a student who "scooped" me. My approach is different enough to be interesting.
Would it help to be entirely open about it? What would happen if you published every small piece of research that journals would accept, and blogged or self published the rest?
You might not be the first one to publish results, but you could reasonably claim that anybody that scooped you would be basing their experiment almost entirely off of yours, so it would be somewhat of a derivative work. And if you released a details on your experimental plan, I think that a scientist would have to design a new experiment if he wanted to say that he wasn't just testing your hypothesis.
And I think that even if someone did try to scoop you, you would have claim to enough original research to get a phd.
This strategy might cause you to change your thesis some. But if you are worried about getting scooped, it might be beneficial to release as quickly as possible to lay claim to as much research as possible.
I'm just comparing it to the patent system. Companies don't want their ip infringed, so they file as many patents as possible to get legal rights to their ip.
