Dynamic biplane radiographic (DBR) imaging measures continuous vertebral motion during in vivo, functional tasks with sub-millimeter accuracy, offering the potential to develop novel biomechanical markers for lower back disorders based on true dynamic motion rather than metrics based on static end-range of motion. Nevertheless, the reliability of DBR metrics is unclear due to the inherent variability in movement over multiple repetitions and a need to minimize radiation exposure associated with each movement repetition. The objectives of this study were to determine the margin of uncertainty (MOU) in estimating the typical intervertebral kinematics waveforms based upon only a small number of movement repetitions, and to determine the day-to-day repeatability of intervertebral kinematics waveforms measured using DBR. Lumbar spine kinematics data were collected from two participant groups who performed multiple trials of flexion-extension or lateral bending to assess the uncertainty in the mean estimated waveform. The first group performed ten repetitions on the same day. Data from that group was used to estimate MOU as a function of the number of repetitions. The second group performed five repetitions on each of two separate days. MOU was not only movement-specific, but also motion segment-specific. Using just one or two trials yielded a relatively high MOU (e.g. > 4& or 4 mm), however, collecting at least three repetitions reduced the MOU by 40% or more. Results demonstrate the reproducibility of DBR-derived measurements is greatly improved by collecting at least three repetitions, while simultaneously minimizing radiation exposure.