The conquest and plunder of major cities is among the most potent symbols of political decline for a defender and vitality — or barbarity — for the aggressor. The Visigothic sack of Rome in 410 CE, for example, is regarded as a milestone moment in the decline of the Western Roman Empire. Likewise, the Mongols put an end to Bagdad’s golden age with a gruesome frenzy of destruction in 1258 CE. These conquerors qualified as “roving bandits” in the most literal sense, but also insofar as they exhibited qualities predicted by Mancur Olson for rulers (and their retinues) with short time horizons. Mainly, that maximizing loot extraction takes priority over prospective long-term economic returns. We test this proposition using a novel dataset of actor-oriented (i.e., not “natural”) episodes of rapacious extraction and urban destruction in Eurasia from 1 to 1799 CE. To construct this dataset, we rely on a combination of existing sources related to premodern urbanization and major battles throughout Europe and Asia. On one hand, we find support for a historical link between mobility and rapacious extraction episodes in which conquerors subjected their new possessions to plunder and even destruction. However, in contrast to stylized models of time horizons and extraction, we identify military force mobility, not the faction’s (and thus ruler’s) mobility as a robust predictor of rapacious extraction tendencies. Accordingly, we propose a modification of Olson’s model in which aggregate regime (polity, faction, etc.) extractive behavior is primarily a function of principal–agent dynamics grounded in elite-level incentives and time horizons — elite mobility counteracted rulers’ stationary bandit tendencies.