Reliability and validity of the load-velocity relationship to predict the 1RM back squat

Document Type

Journal Article


National Strength and Conditioning Association


School of Medical and Health Sciences




Banyard, H. G., Nosaka, K., & Haff, G. G. (2017). Reliability and validity of the load–velocity relationship to predict the 1RM back squat. The Journal of Strength & Conditioning Research, 31(7), 1897-1904. Available here


This study investigated the reliability and validity of the load–velocity relationship to predict the free-weight back squat one repetition maximum (1RM). Seventeen strength-trained males performed three 1RM assessments on 3 separate days. All repetitions were performed to full depth with maximal concentric effort. Predicted 1RMs were calculated by entering the mean concentric velocity of the 1RM (V1RM) into an individualized linear regression equation, which was derived from the load–velocity relationship of 3 (20, 40, 60% of 1RM), 4 (20, 40, 60, 80% of 1RM), or 5 (20, 40, 60, 80, 90% of 1RM) incremental warm-up sets. The actual 1RM (140.3 ± 27.2 kg) was very stable between 3 trials (ICC = 0.99; SEM = 2.9 kg; CV = 2.1%; ES = 0.11). Predicted 1RM from 5 warm-up sets up to and including 90% of 1RM was the most reliable (ICC = 0.92; SEM = 8.6 kg; CV = 5.7%; ES = −0.02) and valid (r = 0.93; SEE = 10.6 kg; CV = 7.4%; ES = 0.71) of the predicted 1RM methods. However, all predicted 1RMs were significantly different (p ≤ 0.05; ES = 0.71–1.04) from the actual 1RM. Individual variation for the actual 1RM was small between trials ranging from −5.6 to 4.8% compared with the most accurate predictive method up to 90% of 1RM, which was more variable (−5.5 to 27.8%). Importantly, the V1RM (0.24 ± 0.06 m·s−1) was unreliable between trials (ICC = 0.42; SEM = 0.05 m·s−1; CV = 22.5%; ES = 0.14). The load–velocity relationship for the full depth free-weight back squat showed moderate reliability and validity but could not accurately predict 1RM, which was stable between trials. Thus, the load–velocity relationship 1RM prediction method used in this study cannot accurately modify sessional training loads because of large V1RM variability.



Access Rights

subscription content