We're talking about the probability that a person can accurately predict something about which he admittedly hasn't the first fragment of a conception.
Your scenario has to be rewritten:
Scientist A : I think X blahblah
Scientist B : Interesting. That flies in the face of everything we've learned up till now. Do you have any evidence?
Scientist A : Absolutely none! I just made it up right now.
Scientist B : Well how can it possibly be correct then? Let's stick to building things up from what we know and what's likely based on that.
Scientist C : Now hold on, B! A may have just invented that idea out of thin air based on no evidence whatsoever, but we can't conclude that it's wrong.
Scientist B : Well by that logic I could come up with something out of thin air myself, and would you say that we should waste even an instant considering it?
Scientist C : I try to keep my mind considering all things with equal possibility until they are definitively proven false.
Scientist B : Well I'm more interested in theories based on evidence.
Scientist C : Oh, B. You fool. If we'd concentrated only on things that were more likely than not then where would we be today? Why we wouldn't even have math.
Scientist A : Well I'm going to go develop my theory of X. It will require invalidating much of what we know so far about nature so I'd better get on it.
Scientist B : Enjoy yourself. I'll be ignoring you due to the virtual certainty that your theory is wrong. If you try to teach your theories to the children in place of the reason-based work I've accomplished I'll write editorials. Scathing editorials.
Scientist C : I shall split my time equally between observing the actions of both. If B writes scathing editorials I'll try to play the part of the peacemaker. We're all equal under the eyes of what may be a higher being, and no theory is any worse than any others unless 100% proof negative is achieved.