Building damage probabilities are invaluable for assessing short-term losses from natural hazards. In many countries however, the individual building level data required for assessing reliable damage are usually unavailable. This paper shows how the post-processing of aggregate HAZUS earthquake damage assessments can yield building-level damage probabilities. On the basis of three plausible scenarios for Northern Israel, we generate and visualize a building-level combined damage probability index. We use the tools of exploratory spatial data analysis to purge any causal influences in the spatial pattern of these calculated damage probabilities. The costs and benefits of our approach are discussed.