Instrumental gait analysis is increasingly recognized as a useful tool for the evaluation of movement disorders. The various assessment devices available to date have mostly been evaluated in healthy populations only. We aimed to explore whether reliability and validity seen in healthy subjects can also be assumed in subjects with cerebellar ataxic gait. Gait was recorded simultaneously with two devices - a sensor-embedded walkway and an inertial sensor based system - to explore test accuracy in two groups of subjects: one with mild to moderate cerebellar ataxia due to a subtype of autosomal-dominantly inherited neurodegenerative disorder (SCA14), the other were healthy subjects matched for age and height (CTR). Test precision was assessed by retest within session for each device. In conclusion, accuracy and repeatability of gait measurements were not compromised by ataxic gait disorder. The accuracy of spatial measures was speed-dependent and a direct comparison of stride length from both devices will be most reliably made at comfortable speed. Measures of stride variability had low agreement between methods in CTR and at retest in both groups. However, the marked increase of stride variability in ataxia outweighs the observed amount of imprecision.