If I interpret JGW's comment correctly, the rhetorical question wouldn't suffer much were it phrased "If X were true, what would you expect debates with people who believed X was true to look like?"

The answer is that as humans speaking colloquially, they would first say "X is true" and then rattle off reasons, in the same format apologists use. This pattern of speaking does not strongly imply that the pattern of speaking was the pattern of thinking, it's just how people speak.

Some people do think in this pattern, including many theists, so one can lose sight of the fact that the mode of speaking and mode of thinking are not perfectly correlated.

explain how this correctness can be tested experimentally

As hard as they try, I don't think religions can avoid making testable claims. The untestable claim X is implicitly paired with the testable claim that one should believe X.

Even probabilistic beliefs are held because belief systems lead people to expect things. when confronted with inputs.

If a Unitarian Universalist says "(One ought to believe that) there is a 99.9% chance Jesus existed," and the scientific consensus is "(One ought to believe that) there is a 99.5% chance Jesus existed," and we fire up the ol' AIXI and it outputs the latter, the UU is wrong even if Jesus existed as one historical character.

The UU might as well claim that the Noah's ark tale literally happened, if he isn't his belief system is in one way worse than the fundamentalist's, as his contains the proposition "To hell with reality when it contradicts my religion, if I can defy it without doing so in a flagrant enough way that people notice, including myself", whereas the latter's contains the proposition "To hell with reality when it contradicts my religion." Much simpler.