This article addresses a classic issue in the Metaphysics of Events concerning the conditions under which a specific event remains unchanged amid possible counterfactual changes. Multiplicators, who admit little or no variation in property categories for event individuation, contrast with unifiers, who are more tolerant of such variations. Two fundamental arguments are critically analyzed. Firstly, arguments based on common linguistic practices suggesting that we speak about the same events varying over time are discussed. Counterexamples demonstrate the unreliability of these practices (...) from the standpoint of the metaphysics of event individuation. Secondly, the argument that the non-fragilist position conflicts with the principle of alternative possibilities and the intuition that cases with very close action alternatives do not involve moral responsibility is addressed. The analysis indicates that such cases may involve numerically distinct actions, suggesting the need for a modification in the principle of alternative possibilities. It is concluded that there are no plausible reasons to reject the fragilist stance, and it is possible to decide abductively in its favor. (shrink)