Impact of the optimal minimum velocity threshold on the accuracy of one repetition maximum estimation-based on a mixed model

Abstract Purpose This study aimed at developing two mixed methods approaches for 1RM prediction (generalized equations together with the individual load–velocity relationship—LVR). The validity of such equations was explored and compared with the validity of individual LVRs. Methods The submaximal LVRs of seventy-six males was obtained for the free-weight back squat. Theoretical load at zero velocity (LD0), the slope of the individual LVR, and the individual (Model IND ) and optimal minimum velocity thresholds (MVT) (Model OPT ) were selected to develop the equations. Prediction accuracy was determined through mean differences, absolute percent errors and Bland–Altman plots. Results Model IND was predictive of 1RM ( p < 0.0001), explaining 89.7% of its variance. Model OPT increased the prediction power of 1RM to 98.7% ( p < 0.0001). No significant differences in the absolute percent errors were found (4.7 and 3.9%, for Model IND and Model OPT , respectively). The mean difference between actual and estimated 1RM was nearly zero for both models, while individual LVRs relying on the individual MVT displayed the contrary. The limits of agreement were 15.1 and 11.4 kg for Model IND and Model OPT , respectively. Conclusion The individual and optimal MVTs add predictive value to LD0 and the slope of the individual LVR for 1RM estimation on a group level and avoid error trends over the entire muscular strength spectrum. However, only Model IND increased the accuracy of estimations when compared with the more direct approach. From a practical standpoint, practitioners are encouraged to rely on the individual LVR instead of Model OPT, when using the optimal MVT for estimation purposes.