Inference to the best explanation only works when the candidate explanations are of a kind we can assess. “A miracle happened” has no background rate, no mechanism, and can explain anything. Doesn’t introducing supernatural hypotheses break the method?

This is the strongest objection to using abduction in apologetics. The reply is that God did it is not offered as a bare hypothesis but as part of a broader view, theism, that makes specific predictions: that the universe has a beginning and is orderly, that moral facts exist, that a certain sort of religious history would occur. Its explanatory power is assessed by comparison with the naturalistic view on all those data together, and it can be disconfirmed: a universe with no moral facts, an eternal past, or a Jesus who stayed dead would each count against it. Whether it wins the comparison is the debate; that it can be entered into the comparison is what this claim defends.