Is What is Worse More Likely?—The Probabilistic Explanation of the Epistemic Side-Effect Effect