BackgroundAnthropometric measures such as the body mass index (BMI), reciprocal ponderal index (RPI), and waist-to-height ratio (WHR) have been proposed as predictors of physical fitness. This study aimed to identify the differences in explanatory capacity and fit of BMI, RPI, and WHR on physical fitness, which involves jumping, sprinting, change of direction, and aerobic capacity, by adjusting the polynomial regression.MethodsA sample of 297 healthy, recreationally active male university students between 18 and 20 years old was recruited for this study. Anthropometric measurements (height: 174.09 ± 6.27 cm, weight: 78.98 ± 20.27 kg, waist circumference: 93.74 ± 14.56 cm) were taken for each participant. Jumping tests (squat jump, countermovement jump), sprinting tests (20 m sprint), agility tests (agility T-test), and aerobic/endurance tests (6 min walk test, VAM-EVAL test) were performed. Nonlinear quadratic regression models were used to assess the relationship between the jump, sprint, and fitness test scores and the anthropometric indices. The models were compared based on R-squares and Bayesian Information Criterion (BIC). The significance level was set at p < 0.05.ResultsThe results showed that all the indices predicted a portion of the variance because all variables and index relationships were significant. Regarding the fitted models, the Bayesian Information Criterion showed that BMI was the best indicator of performance, although the RPI was better for VO2max.ConclusionThese findings may be of great interest to practitioners because it appears that anthropometric measures can be used to predict physical fitness in certain tests although the accuracy raises any concerns.