Longitudinal Monitoring of Load-Velocity Variables in Preferred-Stroke and Front-Crawl with National and International Swimmers

Load-velocity (LV) profiling in swimming provides key metrics, including theoretical maximal velocity (V ₀) and theoretical maximal load (L ₀); however, longitudinal studies tracking these variables across competitive seasons are limited. This study investigated LV profiling and competition performance in national and international-level swimmers (Level 1–3) over a 15-month period. Twenty-six swimmers participated (16 males: age: 19.8 ± 3.9 years, body mass: 80.3 ± 7.9 kg, height: 1.84 ± 0.07 m; 10 females: age: 20.7 ± 3.6 years, body mass: 68.2 ± 5.7 kg, height: 1.74 ± 0.03 m), all specializing in 50–200 m events. Swimmers completed 4–6 testing sessions, each involving 3 × 10 m sprints against resistances of 1, 5, and 9 kg (males) and 1, 3, and 5 kg (females), in both front-crawl and their preferred-stroke. Linear mixed-effects models assessed changes in LV outputs—V ₀, L ₀ (absolute and relative to body mass), relative slope (−V ₀/rL ₀), and active drag (AD). Smallest worthwhile change (SWC) assessed within-athlete variation, while Pearson's correlations evaluated relationships between race performance and LV outputs. Analysis of preferred-stroke found males exhibited significantly higher values across all variables except the slope (p = 0.607). National-level swimmers had lower L ₀ (−2.8 kg, p = 0.019), but no statistical difference in rL ₀ (−1.5%, p = 0.244) or slope (−0.002 m/s/%, p = 0.558). AD remained stable across observations, though males produced greater drag (+30.2 N, p < 0.001), while national-level swimmers produced less (−12.8 N, p = 0.045). Analysis of front-crawl performance found males presented higher values across all variables (p ≤ 0.05) while national-level swimmers were lower (p < 0.005). SWC analysis revealed that most within-athlete changes in V ₀ and L ₀ were trivial or unclear, with only isolated meaningful changes observed. Large to very large correlations existed between race performance and L ₀ (r = 0.67, p < 0.05), V ₀ (r = 0.73, p < 0.05), and AD (r = 0.58–0.7, p < 0.05) at select observations. These findings highlight the stability of LV profiling metrics over time while reinforcing their relevance in distinguishing between performance levels. This suggests their potential utility in talent identification and informing training prescription.