This paper provides a comparison of the Gazebo and ANVEL simulators and analyzes the aspects of vehicle modeling fidelity that are critical to the design of unmanned ground vehicle (UGV) control and estimation algorithms. The robotic simulators Gazebo [1], from the Open Source Robotics Foundation (OSRF), and Autonomous Navigation Virtual Environment Laboratory (ANVEL) [2], from Quantum Signal, are two popular new tools that are being used extensively in academic, commercial, and military development of perception, navigation, and control algorithms for UGVs. Despite the similarities between Gazebo and ANVEL there has been no direct comparison between the two simulators with respect to their validity as vehicle dynamics simulators.