Ecopath with Ecosim (EwE) is a widely used modeling tool in fishery research and management. Ecopath requires a mass-balanced snapshot of a food web at a particular point in time, which Ecosim then uses to simulate changes in biomass over time. Initial inputs to Ecopath, including estimates for biomasses, production to biomass ratios, consumption to biomass ratios, and diets, rarely produce mass balance, and thus ad hoc changes to inputs are required to balance the model. There has been little previous research of whether ad hoc changes to achieve mass balance affect Ecosim simulations. We constructed an EwE model for the offshore community of Lake Huron, and balanced the model using four contrasting but realistic methods. The four balancing methods were based on two contrasting approaches; in the first approach, production of unbalanced groups was increased by increasing either biomass or the production to biomass ratio, while in the second approach, consumption of predators on unbalanced groups was decreased by decreasing either biomass or the consumption to biomass ratio. We compared six simulation scenarios based on three alternative assumptions about the extent to which mortality rates of prey can change in response to changes in predator biomass (i.e., vulnerabilities) under perturbations to either fishing mortality or environmental production. Changes in simulated biomass values over time were used in a principal components analysis to assess the comparative effect of balancing method, vulnerabilities, and perturbation types. Vulnerabilities explained the most variation in biomass, followed by the type of perturbation. Choice of balancing method explained little of the overall variation in biomass. Under scenarios where changes in predator biomass caused large changes in mortality rates of prey (i.e., high vulnerabilities), variation in biomass was greater than when changes in predator biomass caused only small changes in mortality rates of prey (i.e., low vulnerabilities), and was amplified when environmental production was increased. When standardized to mean changes in biomass within each scenario, scenarios when vulnerabilities were low and when fishing mortality was increased explained the most variation in biomass. Our findings suggested that approaches to balancing Ecopath models have relatively little effect on changes in biomass over time, especially when compared to assumptions about how mortality rates of prey change in response to changes in predator biomass. We concluded that when constructing food-web models using EwE, determining the effect of changes in predator biomass on mortality rates of prey should be prioritized over determining the best way to balance the model.