This simulation study is a simplified one, as it considers only one multiplier and only two traits were included. Applied breeding programs for salmonids often include several multipliers and often target more traits (Rye et al., 2010). In this present study the economic weight in the nucleus and for the multiplier were kept constant at vCAND/vSIB = 50/50, however many more combinations of economic weights are possible also in the multiplier tier. Also, the heritabilities used for the two traits, chosen to reflect a typical situation with a higher heritability for the candidate trait than for the sib-trait, are not representative for all breeding programs. A realistic situation would be to have 6–10 traits in the breeding goal, with different heritabilities and accuracies, and then take two or three out of this for a special dissemination to the grow-out which would make it possible to tailor the fish even more to a specific marked. If a large economic weight is applied to the candidate trait and a small economic weight on the sib trait, it is expected to yield a much larger genetic response as compared to random sampling of parents among the candidates, thus the results from this paper will underestimate the realistic effect of applying a two-step selection strategy.