Despite well over a decade of research on dynamic combinatorial chemistry it is still unclear whether large libraries are more or less likely to yield strong binders than small libraries. We have now addressed this question by simulating a set of libraries containing from 65 to 4828 compounds under a range of different building block and template concentrations. We investigated the effect of library size on (i) the probability of detecting any amplification; (ii) the probability of detecting the strongest binding library member present; and (iii) the binding affinity of the most amplified detectable library member. The results indicate bigger libraries are more likely to produce better binders and that the affinity of the best binders identified rises more rapidly than expected statistically on the basis of the number of screened library members. This implies that it should be advantageous to work with DCLs that are much larger than the vast majority reported thus far.