What do we know about how people predict heretofore unobserved but extreme events? How about a war or pandemic that kills 50 million persons? Or indeed, how likely do people think human extinction is? Despite their prevalence as themes in popular culture, there have been few investigations into the psychology behind beliefs about the likelihood of such events. There is some evidence that people over-weight extreme, low-probability events. However, there also is evidence that they under-weight the probability of events that haven’t yet occurred. We present two studies of adults’ probability assignments to both specific and unspecified catastrophes. The findings indicate that their assignments are strongly suggestible to boosting via descriptions of specific examples and/or causes of such catastrophes. There also is evidence of some polarization in these assignments, and they appear to be influenced by trait optimism, faith in science, and political orientation.