How to recognise potential disasters is a question at the centre of risk analysis. Over-reliance on an incomplete, often epistemologically-biased, historical record, and a focus on quantified and quantifiable risks, have contributed to unanticipated disasters dominating both casualties and financial losses in the first part of the 21st century. Here we present the findings of an online workshop implementing a new scenario-planning method, called downward counterfactual analysis, which is designed to expand the range of risks considered. Interdisciplinary groups of disaster researchers constructed downward counterfactuals for a present-day version of the 365CE Cretan earthquake and tsunami, imagining how these events might have been worse. The resulting counterfactuals have trans-national, long-term impacts, particularly in terms of economic losses, and connect risks previously identified in separate sectors. Most counterfactuals involved socio-political factors, rather than intrinsic components of the hazard, consistent with the idea that there are “no natural disasters”. The prevalence of cascading counterfactuals in our workshop suggests that further work is required to give the appropriate weight to pre-existing economic and social conditions in scenario-planning methods, such as downward counterfactual analysis, which focus on the occurrence of a hazard as the temporal starting point for a disaster. Both proposed counterfactuals and their justifications reflect a bias towards contemporary issues and recent historical disasters. We suggest that interdisciplinary groups can expand the range of imagined risks. However, the setup used here would be improved by including local stakeholders. Qualitative forms of downward counterfactual analysis have potential applications for community engagement and education, as well as for risk analysis.</jats:p>