I interpreted your premise as: (Things said about Y by a person who has a reputation for being an expert on Y) are likely to be (correct.) as opposed to (Things said about Y by a person who has a reputation for being an expert on Y) are (likely to be correct.)

The second was the intended meaning.

If, as you seem to be agreeing, a thing cannot be "likely to be correct" and "incorrect" (as known by the same reasoner), then the premise reduces to "Things said about Y by a person who has a reputation for being an expert on Y are correct".

Given the 'as known by the same reasoner' clause wouldn't that imply that it is '<...> cannot be known to be incorrect'? Either way it is clear that the encapsulation of the probabilistic parts is woefully inadequate here.

Is this really what you intended?

No, but it does seem to be the implication.