Fish recruitment can vary by an order of magnitude between years, and the larval stage is a key determining factor. Zooplankton, the main source of larval food, are temporally and spatially heterogeneous, and this could contribute to recruitment variability and ultimately stock sustainability. Here we use simple stochastic models of larval growth and zooplankton dynamics, together with an evolutionary algorithm, to investigate the role of transient peaks in zooplankton abundance and the match/mismatch hypothesis in recruitment success and variability. We draw four main conclusions. (i) Stochasticity in individual growth is more beneficial to recruitment when larvae experience high food availability early on in their growth. (ii) When the timing of peak prey abundance is stochastic, recruitment probabilities are greatest for hatching days just before the expected timing of peak prey abundance. (iii) When the timing of periods of high prey density is held fixed, the evolved optimum hatching day becomes earlier as the length of the high density period increases. (iv) When both the timing and length of the periods of high density are allowed to (co-)vary, we find no evidence of strong selection pressure for specific hatching days, only for a hatching "window" around the expected prey peak.