Since the 1980s, there has been a shift in energy research. It has shifted from approaches that forecast or project the future to approaches which make more tentative claims and which explore several plausible scenarios. Due to multiple uncertainties in energy systems, there is an infinite amount of plausible scenarios that could be constructed and scenario developers therefore choose smaller, more tangible sets of scenarios to analyse. Yet, it is often unclear how and why this scenario choice is made and how such choices might be improved. This paper presents a retrospective analysis of twelve UK energy scenarios developed between 1978 and 2002. It investigates how specific scenarios were chosen and whether these choices captured the actual UK energy system transition. It finds that scenario choice reflected contemporary debates, leading to a focus on certain issues and limiting the insights gleaned from these exercises. The paper argues for multi-organisation, multi-method and multi-scenario approaches to the development of energy scenarios to capture the wide range of insights on offer. Rather than focus on uncertainty in model parameters, greater reflection on structural uncertainties, such as shifts in energy governance, is also required.