Variation in performance times of elite flat-water canoeists from race to race.

PURPOSE: To estimate variability in performance time and smallest worthwhile changes for elite flat-water canoeists competing in 200-, 500- or 1000-m events at international regattas.

METHODS: The data came from A and B finals held at 7 to 13 regattas in 2003 to 2007. A linear mixed-model analysis of log-transformed official race times provided estimates of variability as coefficients of variation and included terms to account for changes in performance between years, venues, and A and B finals.

RESULTS: For men, the within-athlete variation in A finals was similar in canoeing and kayaking events, with the 200-m men's events demonstrating probably less variability than the longer events (by an overall factor of 0.75, x// 1.33) that may reflect differences in pacing strategies. In contrast, the within-athlete variation for women kayakers in A finals of the 500-m event was only half that of the other distances (ratio 0.54, x// 1.29), possibly because of differences in competitive experience or depth of competition. Predictability of performance in A finals was moderate to very high (interclass correlations 0.40 to 0.89). Within-athlete variation in the B finals was generally greater than in the A finals for the three distances for men, but there was no clear pattern for women.

CONCLUSION: The smallest worthwhile changes in performance time (0.3 x within-athlete variability) in canoeing and kayaking are approx. 0.3% to 0.6%. Effects of 1% to 2% in power output would be required to achieve such changes in this generally highly predictable sport.