INTENT: The intent of this paper is to examine race segment times and stroking variables in the 100 m breaststroke, see how these were related over the course of the race, determine how these contributed to the end-race results and how this differed among functional classes. METHODS: With the approval of the International Paralympic Committee and IPC Swimming, all swimming heats and finals were recorded at the Sydney 2000 Paralympic Games. Five digital video cameras were suspended 16 m above the pool surface. Cameras were connected via time code generators to videocassette recorders and the official timing system. All 50 m split times, end-race result, and block reaction times were automatically downloaded. Start time (start to 15 m), finish time (last 5 m of race), turn times in and out (7.5 m in and out), swimming speed, stroke length and rate in 4 race sections were measured on the tapes for 46 male and 36 female finalists. Individual impairment data was collected. Means, SD and coefficients of variation (CV) were calculated. Simple and multiple correlations, Cluster Analysis and ANOVA were performed (p<.01). RESULTS: In both men and women an R2 of .98 was achieved using only race section 3 speed to predict total race speed. The RMSE was, however, reduced to .0056m/s in men and .0055m/s in women by using speeds in sections 2, 3, and 4 and start speed. The changes in race segment speed between heat and final were also compared to changes in end-race result. The start speed was most responsible for changes in men (R2=.55), while changes in race section 3 were most responsible in women (R2=.59). In men, adding changes in section 2 and 4 to the equation along with start speed increased R2 to .89 (RMSE = .006m/s). In women, R2 reached a plateau at .93 by including changes in section speeds 2, 3, and 4, and finish speed (RMSE= .003m/s). A Cluster analysis was done for 3 changes in speed between race sections. In men, Cluster 1 (n=12) showed an even loss of speed over the entire race (~6 %). Cluster 2 (n=16) lost a little speed (~3.3 %) at the beginning and end of the race but a large amount in the middle (8.12 %). Cluster 3 (n=8) had a flat race model losing only (~2.5 %) speed per section. In women, Cluster 1 (n=21) had a flat race model, with a 3 % loss of swim speed over the race (= men`s Cluster 3). Cluster 2 had a steeper model (n=13), losing 4 % at the beginning of the race and then 8.6 % and 7 % in the later sections. There were no differences in functional class or performance between clusters in either men or women. The race is thus won by swimming faster and, to a lesser extent, by improving start and finishing speed. No race section speed contributed more to the end-race result than any other. Apparently a variety of strategies are used to win the race and there is no obvious advantage to any of the race models isolated here.
© Copyright 2003 Biomechanics and Medicine in Swimming IX. Published by University of Saint-Etienne. All rights reserved.