To accurately study pedestrians’ behaviors in Virtual Reality microscopic traffic simulators, the fidelity gaps between the tasks performed in the simulators and its real life equivalents must be reduced by increasing the simulator fidelity. Because of technical, financial, and usability constrains, targeting high levels of simulation fidelity might not be affordable and alternately an optimal level of simulator fidelity must be targeted. This optimal fidelity level can be achieved by trading off between the fidelity of the different simulator’s elements to achieve the targeted usability goals. Because VR road users’ behavior assessment simulators are a relatively new field, the impact of simulator fidelity on its usability is unknown. This research aims to explore the optimal fidelity level of pedestrians’ behavior assessment simulator which can afford high fidelity cognitive and motor crossing tasks hypothesizing that: affording high biomechanical symmetry level of the locomotion method is sufficient to achieve the required optimal fidelity. To validate this hypothesis, two usability pilot testings have been conducted on two prototypes of pedestrians’ behavior assessment simulator which studies the gap acceptance crossing behaviors. The testing results indicated that while increasing the fidelity of the locomotion method allowed users to conduct complete crossings without usability errors, intentionally reducing the fidelity of other simulator elements had a negative impact on the correctness and accuracy of the crossing tasks.