Studies on mate choice mainly focus on the evolution of signals that would maximize the probability of finding a good-quality partner. Most models of sexual selection rely on the implicit assumption that individuals can freely compare and spot the best mates in a heterogeneous population. Comparatively few studies have investigated the consequences of the mate-sampling process. Several sampling strategies have been studied from theoretical or experimental perspectives. They belong to two families of decision rules: best-of-n strategies (individuals sample n partners before choosing the best one within this pool) or threshold strategies (individuals sequentially sample the available partners and choose the first one whose quality exceeds a threshold criterion). Almost all models studying these strategies neglect the effect of scramble competition. If each paired individual is removed from the population of available partners, the distribution of partner quality dynamically changes as a function of the strategies of the other competitors. By means of simple simulations assuming opportunity costs, to the exclusion of all other costs, we show that scramble competition is a sufficient constraint to severely impair the evolution of choosy decision rules. In most cases, the evolutionarily stable strategy is to have a very low acceptance threshold or to sample two individuals at most in the population. This result may explain some discrepancies between predictions from previous models and their experimental validations. It also emphasizes the importance of considering the pairing process in studies of sexual selection.