Students estimated the chance that President Clinton would be convicted or acquitted in the U.S. Senate impeachment trial, at 4 points in time: (a) 22 days before the verdict (distal pretest), (b) 3 days before the verdict (proximal pretest), (c) 4 days after the verdict (immediate posttest), and (d) 11 days after the verdict (delayed posttest). Within-subjects analyses revealed that both foresight and hindsight changed over time. With respect to foresight, the perceived likelihood of conviction significantly decreased and the perceived likelihood of acquittal significantly increased between the 2 pretests. Content analyses of archived national news reports supported the notion that media predictions about the impending vote produced this temporal shift toward greater prospective accuracy. Confirming hypotheses, hindsight bias emerged for estimates of the prior probability of conviction, but not for estimates of the prior probability of acquittal. Within-subjects analyses revealed that 4 days after the verdict, participants accurately recalled that their prospective estimates of the likelihood of conviction had shifted toward greater accuracy over time; but 1 week later, they incorrectly believed that they had been more certain all along that Clinton would not be convicted. This article suggests that the potential losses associated with conviction were more salient than the potential gains associated with acquittal, and speculates that cognitive assimilation of the past did not occur until people had forgotten their earlier beliefs. These findings are important because they demonstrate that whether or not hindsight bias emerges depends not only on how long after the event people make retrospective judgments, but also how long before the event people focus these retrospective judgments. Supporting a self-serving interpretation, hindsight bias held for one's own estimates, but not for one's beliefs about the estimates of the average American or one's best friend. Consistent with Hawkins and Hastie's (1990) conclusions, a hybrid of cognitive and motivational accounts thus best explains the results.