Mental models can help decision-makers understand complex strategic environments and thus choose better strategies. Recent research in the theory-based view of strategy argues that theories– mental models that represent cause-and-effect relationships between strategic choices– are especially useful in strategy selection because they effectively guide experimentation and learning. In this paper, I use a formal and computational model to explain why the cause-and-effect links of theories outperform the associative links often used in mental models. I show that, by causally ordering choices, theories simplify how decision-makers see the conditional dependencies between them, making it easier to evaluate performance and improve strategies over time. My work suggests that even if causal models are less accurate representations of the strategic environment than associative models, their simplified representation consistently leads decision-makers to more successful strategies.