The purpose of this study was to assess the reliability of a repeated-sprint test, specifically designed for field-hockey, as it was based directly on the time-motion analysis of elite level competition. The test consisted of 6 x 30-m over-ground sprints departing on 25s, with an active recovery (similar to 3.1-3.3ms(-1)) between sprints. Ten highly trained, mate, field-hockey players (mean +/- S.D.: age, 23 +/- 3 years; body mass, 78.1 +/- 7.1 kg) participated in this study. Following familiarisation, the subjects performed the repeated-sprint test on two occasions, 7 days apart. The reliability of the test variables was assessed by the typical error of measurement (TE). The total sprint time was very reliable (T-1: 26.79 +/- 0.76s versus T-2: 26.83 +/- 0.74 s), as the TE was 0.7% (95% CL, 0.5-1.2%). However, the percent sprint decrement was less reliable (T-1: 5.6 +/- 0.9% versus T-2: 5.8 +/- 1.0%), with the TE being 14.9% (95% CL, 10.8-31.3%). In summary, it is suggested that this field-hockey-specific, repeated-sprint test is very reliable when the results are presented as the total sprint time. (c) 2006 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.