Abstract Despite commitments to sustainability by essentially every national and international management agency, fisheries repeatedly fail to meet that objective. Major declines in stock status are common, and collapses have occurred in many jurisdictions. What makes unsustainability such a common feature of fisheries, when none of the parties—the managers, their science advisors, nor the industry itself—wish it? Historical trends in biomass and catch data for flatfish fisheries in New England, the Newfoundland-Labrador Shelf, west Coast of British Columbia, Bering Sea-Gulf of Alaska, and North Sea-Northern Shelf are used to test some general assumptions of governance and environmental variability and their role in sustainable fisheries management. The evidence suggests that sustainability is hard to achieve in flatfish fisheries, the majority of stocks (38 of 65) were overfished at some period during their history. However, some stocks in every category supported sustainable fisheries for the full period examined and this would appear to be achievable for a range of conditions. There were expected differences in the probability of management success between major geo-political regions and between different jurisdictional regimes. There were also contrary differences in sustainability between bycatch and target fisheries and in stock size. There was no support for the assumptions regarding harshness of the environment. More than any other factor examined, failure to comply with science advice greatly increases the risk of unsustainability. Finally, the analysis provides insight into the value of quantitative reference points as an objective tool in performance measurement.