The sensitivity of biomechanical models to parameter estimation errors is crucial in determining the reliability of the dynamic estimates provided by these models. For evaluating the risk of head and neck injury from indirect impact (inertial loading), the forces and torques at key anatomical locations are important dynamic quantities. For human volunteers, these variables are estimated using head/neck models incorporating measured kinematic time traces and several indirectly measured mechanical and geometric parameters (e.g., the head center of gravity).In this paper, the sensitivity of estimated forces and torques at the occipital condyles to variations in head/neck geometric and mechanical properties, initial head positioning, and input kinematics is illustrated using a single fixed link model. Using anthropometric X-ray and fitted inertial response data from human volunteers, these forces and torques are estimated for two standard geometric/mechanical datasets. Monte Carlo techniques are employed to show how uncertainties in live subject head/neck geometric and mechanical properties and input kinematics severely limit the accurate assessment of head/neck dynamics even for a very simple mechanical model.The simulation database consists of six frontal (-Gx) impact tests at paired g-levels of 8 and 15g involving three anthropometrically dissimilar human research volunteers (HRVs). The results reflect the influence of head geometry, mass, moments of inertia, initial head positioning and acceleration pulse characteristics on the head/neck dynamic response. No attempt is made to assess the additional effects of measurement error. This analysis has implications for the usefulness of mechanical models in determining the effect of restraint systems and head protective devices on head/neck dynamics. These implications are also discussed.