In European lobster, Homarus gammarus, wide growth variation means that annual recruitment to a fishery (individuals reaching legal size in the same year) consists of at least six year-classes (individuals hatching in the same year). In this paper, a simple simulation analysis is used to explore the effects of uncertainty about the specifics of this protracted recruitment pattern on the way that we perceive the spawner–recruit relationship. In the simulation, if the age range of recruits is underestimated or a simple correction for growth variation is applied by averaging numbers of recruits across years, a spawner–recruit relationship with artefactual curvature and noise arises. Growth variability is typical in animal populations and problems with protracted recruitment may occur in any situation where recruitment is based on size. Asymptotic spawner–recruit curves may not necessarily reflect density-dependent compensatory mortality and resilience to harvesting. The evidence presented here has important management implications for lobster and other exploited species.