Peak power prediction in junior basketballers: Comparing linear and allometric models

Duncan, MJ, Hankey, J, Lyons, M, James, RS, and Nevill, AM. Peak power prediction in junior basketballers: Comparing linear and allometric models. J Strength Cond Res 27(3): 597-603, 2013-Equations, commonly used to predict peak power from jump height, have relied on linear additive models that are biologically unsound beyond the range of observations because of high negative intercept values. This study explored the utility of allometric multiplicative modeling to better predict peak power in adolescent basketball players. Seventy-seven elite junior basketball players (62 adolescent boys, 15 adolescent girls, age = 16.8 ± 0.8 years) performed 3 counter movement jumps (CMJs) on a force platform. Both linear and multiplicative models were then used to determine their efficacy. Four previously published linear equations were significantly associated with actual peak power (all p < 0.01), although here were significant differences between actual and estimated peak power using the SJ and CMJ equations by Sayers (both p < 0.001). Allometric modeling was used to determine an alternative biologically sound equation which was more strongly associated with (r = 0.886, p < 0.001), and not significantly different to (p > 0.05), actual peak power and predicted 77.9% of the variance in actual peak power (adjusted R2 = 0.779, p < 0.001). Exponents close to 1 for body mass and CMJ height indicated that peak power could also be determined from the product of body mass and CMJ height. This equation was significantly associated (r = 0.871, p < 0.001) with, and not significantly different to, actual peak power (adjusted R2 = 0.756, p > 0.05) and offered a more accurate estimation of peak power than previously validated linear additive models examined in this study. The allometric model determined from this study or the multiplicative model (body mass × CMJ height) provides biologically sound models to accurately estimate peak power in elite adolescent basketballers that are more accurate than equations based on linear additive models.