The aim of this study was to establish the validity of using shuttle run performance over 20 m to predict accumulated oxygen deficit. A new high-intensity shuttle run test (HIST) was devised, during which subjects ran to exhaustion at a speed equivalent to 120% of their performance attained during a progressive shuttle run test. The reliability of the new test was examined and found to be acceptable for 18 subjects who performed the test twice on separate days (r = 0.84, P < 0.01, study I). The discriminating ability of the new test was examined by comparing the distance covered by eight sprint- and eight endurance-trained athletes at 120% of their respective progressive shuttle run performances (615 +/- 111 vs 273 +/- 84 m, P < 0.01, study II). The strongest predictor of accumulated oxygen deficit for 27 subjects was found to be the geometric mean of the performances on the new test and on the progressive shuttle run test (r = 0.74, study III). The regression equation for this relationship was then used to estimate the accumulated oxygen deficit for a second group of 16 subjects (study IV). The correlation between the estimated and measured accumulated oxygen deficits was significant (r = 0.79, P < 0.01). The results from studies III and IV were therefore combined with the data from six new subjects to give a regression equation for predictive purposes based on 49 subjects.