In contrast to the extant quantitative studies on the hindsight effect, the present narrative analysis looks at it from a rare angle of talk-in-interaction. Fifty one-to-one interviews were done with five student groups, each of which was presented with a scenario ending with one factual outcome and three alternative outcomes that actually did not happen. Confirming the already proven role of the provided event outcome in overestimating the probability of its occurrence, this study expands the current understanding of the processes neglected by the research on the hindsight effect. It does so by highlighting the strategic use of vagueness, self-empowerment and selective perspective-taking that question the assessment of the past for its own sake and emphasize the importance of self-presentation. (Contains 7 footnotes.)