Neuromotor processes are inherently noisy, which results in variability during movement and fluctuations in motor control. Although controversial, low levels of variability are traditionally considered healthy, while increased levels are thought to be pathological. This systematic review and meta-analysis of the literature investigates the thresholds between healthy and pathological task variability. After examining 13,195 publications, 109 studies were included. Results from over 3000 healthy subjects and 2775 patients revealed an overall positive effect size of pathology on variability of 0.59 for walking and 0.80 for sway. For the coefficient of variation of stride time (ST) and sway area (SA), upper thresholds of 2.6% and 265mm(2) discriminated pathological from asymptomatic performance, while 1.1% and 62mm(2) identified the lower thresholds for pathological variability. This window of healthy performance now provides science based evidence for the discrimination of both extremely low and extremely high levels of variability in the identification as well as standardised monitoring of functional status in neurological cases.