Multi-agent modeling is a computational approach to model behavior of complex systems in terms of simple micro level agent rules that result in macro level patterns and regularities. It has been argued that complex systems approaches provide distinct advantages over traditional equation-based mathematical modeling approaches in the process of scientific inquiry. We present a case study on how multi-agent modeling can be used to develop thought experiments in order to push theory forward. We develop a model of the evolution of gamete dimorphism (anisogamy), for which there are several competing theories in the evolutionary biology literature. We share the outcomes of our model and discuss how the model findings compare with, and contribute to previous work in the literature. The model clarifies mechanisms that can result in the evolution of anisogamy and offers a much simpler structure that is easier to understand, test, modify and extend.