The readiness potential (RP), a slow electroencephalographic (EEG) pre-movement potential, was used in earlier studies to determine the onset and order of neural processes preceding voluntary movement. Latencies in these studies were always calculated from the averaged RP, whereas onset times of individual trials remained inaccessible. The aim of this study was to use a different, statistical approach to examine how variable the onset of single-trial RPs within subjects is. We recorded RPs in 15 right-handed healthy subjects while they made self-paced repetitive unilateral button presses with their dominant right hand. Skewness, a measure of distribution asymmetry, was analysed in sets of single-trial RPs to discriminate between fixed onset and variable onset models. Results show that skewness has values around zero across all electrodes and pre-movement intervals without any significant deviation. This result obtained for the original data was replicated using modelled data with fixed onset times, whereas alternative models with variable onset times (i.e., including trials with exceptionally early onset) showed significant deviations of skewness from zero. In conclusion, for simple repetitive movements with the dominant hand these results confirm a fixed onset model of the RP with similar onset times of pre-movement cortical activation across trials. The methodology might be also applicable for other paradigms to test basic assumptions of mental chronometry.