Performance Relationships in Timed and Mass-Start Events for Elite Omnium Cyclists.

PURPOSE: To explore the extent to which factors that determine performance transfer within and between time-trial and mass-start events in the track-cycling Omnium.

METHODS: Official finish rank in the 3 time-trial events, in the 3 mass-start events, and in the competition overall were collated in 20 international Omnium competitions between 2010 and 2014 for 196 male and 140 female cyclists. Linear mixed modeling of the log-transformed finish time for the time-trial events and of log-transformed finish rank for all events and final rank provided estimates of within-athlete race-to-race changes in performance and average betweenathletes differences across a season. These estimates were converted to various correlations representing relationships within and between the various events and final rank.

RESULTS: Intraclass correlation coefficients, representing race-to-race reproducibility of performance, were similar whether derived from finish rank or finish time for the time-trial events. Log-transformed finish ranks are therefore a suitable measure to assess and compare performance in time-trial and mass-start events. Omnium cyclists were more predictable in their performances from race to race in the timed events, whereas reduced predictability was observed in mass-start events. Interevent correlations indicated stronger links in performance between the timed disciplines, whereas performance in any of the mass-start events had only a slight positive relationship with performance in the other massstart events and little or no relationship with the timed events.

CONCLUSIONS: Further investigation is warranted to determine whether factors related to performance in mass-start events can be identified to improve reproducibility or whether variability in performance results from random chance.