Progression and variability of competitive performance of Olympic swimmers

Abstract

Estimates of progression and variability of athletic performance in competitions are useful for researchers and practitioners interested in factors that affect performance. We used repeated-measures mixed modelling to analyse 676 official race times of 26 US and 25 Australian Olympic swimmers in the 12-month period leading up to the 2000 Olympic Games. Progression was expressed as percent changes in mean performance; variability was expressed as the coefficient of variation in performance of an individual swimmer between races. Within competitions, both nations showed similar improvements in mean time from heats through finals (overall 1.2%; 95% confidence limits 1.1 to 1.3%). Mean competition time also improved over 12 months by a similar amount in both nations (0.9%; 95% confidence limits 0.6 to 1.2%). The US swimmers showed a greater improvement between the finals (a difference of 0.5%; 95% confidence limits −0.2 to 1.1%), which paralleled changes in the medal haul of the two nations. The coefficient of variation in performance time for a swimmer between races was 0.60% (95% confidence limits 0.56 to 0.65%) within a competition and 0.80% (95% confidence limits 0.73 to 0.86%) between competitions. Our results show that: (a) to stay in contention for a medal, an Olympic swimmer should improve his or her performance by ∼1% within a competition and by ∼1% within the year leading up to the Olympics; (b) an additional enhancement of ∼0.4% (one-half the between-competition variability) would substantially increase the swimmer's chances of a medal.

Keywords:

• coefficient of variation
• elite athletes
• reliability
• swimming