The one-repetition maximum (1RM) is a well-established measure of an individual's maximal strength level for both practitioners and researchers (8). Maximum strength plays an integral part in most sports and is associated with enhanced force-time characteristics and general and sport-specific skill performance (37,40). Typically, the 1RM is used to precisely prescribe training intensities in the form of relative loads (% of 1RM), aiming to evoke specific neuromuscular adaptations (36). The 1RM can be determined either directly or indirectly. Direct assessment of 1RM can be described as a trial-and-error procedure where weight will be added until no further repetition can be performed (29). This approach, however, is time consuming, puts athletes at increased risk of injury when performed incorrectly, is impractical within large team settings, and can induce muscle damage and fatigue potentially deteriorating training performance on subsequent days (30). Therefore, regression equations to estimate the 1RM based on the maximum number of repetitions performed to muscular failure with submaximal loads have been established (28,33). Mean differences of 0–4% have been reported between the predicted 1RM and actual 1RM for bench press and back squat applying the repetition-to-failure approach (17,27,34). Although it is recommended to use less than 10 repetitions to increase accuracy (28), this approach also induces excessive levels of fatigue to the athletes, potentially attenuating power and strength adaptations (21). Moreover, the accuracy of these estimations seems to be specific to the given population, number of performed repetitions, exercise, and sex being tested (22). Consequently, there is a need of an alternative approach that accurately estimates 1RM while minimizing the accumulation of fatigue.
The conventional deadlift is a fundamental exercise in resistance training and often used as an assistance exercise by weightlifters or main lift by individual or team sport athletes (24). Research has shown that this traditional exercise can effectively promote neuromuscular adaptations and enhances performance in generic physical qualities (11,38,39). Despite inherent differences in kinematics (16), acute endocrine and neuromuscular responses (4) and tapering strategies (32) between the squat and deadlift, comparatively little research on the latter exercise has been conducted to date.
Over recent years, using commercially available devices to monitor velocity in strength training exercises have gained popularity in both practice and research (23). Because of advancements in technology, linear position transducer has been shown to accurately measure bar velocity for various exercises such as the back squats (2), bench press (13), and squat jumps (12,14). Interestingly, recent studies reported an almost perfect inverse linear relationship between mean concentric velocity and load with lighter loads lifted at higher velocities and vice versa (1,6,26). Therefore, it has been proposed that individualized regression equations could be used to estimate an athlete's daily 1RM based on submaximal warm-up sets enabling practitioners to adjust load prescription more frequently to reflect subtle changes in maximal strength and daily readiness to train (23). Evidence to support this practical application of velocity-based training is limited. When using a paused bench press performed on a Smith machine, Bosquet et al. (6) reported a moderate bias between the actual and estimated 1RM (typical error of estimate [TEE]: 5.8 kg) despite an almost certain correlation (r = 0.93). Similar results were reported from Jidovtseff et al. (22) using the maximal isometric force estimated from the force-velocity profile to predict 1RM. However, it is important to note that these results should be regarded with caution, given that the exercises were performed on a Smith machine and in a concentric-only manner. Consequently, these findings cannot be generalized to movements that are not restricting the barbell pathway or taking advantage of the stretch-shortening cycle, given that free-weight exercises likely change the kinetic and kinematic characteristics of the movement (10,35). More recently, Loturco et al. (25) investigated the accuracy in predicting the 1RM in the free-weight bench press and reported absolute differences of <5% between predicted and actual 1RM. By contrast, in the study of Banyard et al. (2), actual 1RM in free-weight back squat was largely overestimated (absolute difference: 20–30 kg) despite using the mean concentric velocity at 1RM (v at 1RM) of the same session in the linear regression equation. This was likely due to the high intraindividual variability (coefficient of variation [CV]: 22.5%) of the v at 1RM. However, these findings are of limited use for practitioners, given that the predicted 1RM was calculated by using the data of the entire load-velocity relationship of the actual 1RM and same session, respectively. Thus, for accurate daily predictions of 1RM, selected submaximal loads need to be considered and compared with a previously established actual 1RM and corresponding v at 1RM. Although a step forward, the reliability and validity of the load-velocity relationship to predict the 1RM for other free-weight exercises still needs to be investigated.
Therefore, the first objective of this study was to investigate the reliability and validity of 1RM prediction models based on selected submaximal ranges of the load-velocity relationship during the deadlift. The second aim was to provide further information about the stability of the 1RM and associated v at 1RM.
Experimental Approach to the Problem
In this study, we investigated the short-term reliability and validity of the load-velocity profile assessed by a linear position transducer to predict the 1RM for the deadlift. Taking advantage of the linear load-velocity relationship, individualized linear regression equations were computed using the mean concentric velocity for 3 (20, 40, 60% of 1RM, 40, 60, 80% of 1RM, and 60, 80, 90% of 1RM), 4 (20, 40, 60, 80% and 40, 60, 80, 90% of 1RM), and 5 (20, 40, 60, 80, 90% of 1RM) incremental loads. Subjects reported to the laboratory on 3 occasions, each separated by at least 72 hours. The individual v at 1RM from session 2 was then used to calculate the predicted 1RMs based on the computed regression equations of session 3.
Eleven healthy male resistance-trained athletes were recruited for this study (SD ± age: 23.6 ± 1.4 years, height: 1.80 ± 0.06 m, and body mass: 85.6 ± 6.2 kg). All subjects were free from any musculoskeletal disorders, had at least 1 year of resistance training experience (mean training age: 3.2 ± 0.9 years and range: 1–6 years), and were able to perform the deadlift with at least 1.5 times their body mass (relative deadlift 1RM: 2.0 ± 0.3 kg·kg−1). Before the study, all subjects were informed of the purpose and risks of the investigation and provided signed informed consent. This study was approved by Cardiff Metropolitan University, Cardiff School of Sport Research Ethics Committee (acceptance no. 17/6/01P) and was conducted in accordance with the declaration of Helsinki.
Subjects performed 3 testing sessions separated by at least 72 hours. In the first session, subjects received information of the testing procedure, and individual characteristics (e.g., body mass, height, and grip type) were recorded. They then completed the standardized warm-up followed by the initial 1RM assessment based on which all relative loads (i.e., 20, 40, 60, 80, and 90% of 1RM) for the remaining two 1RM sessions were calculated. It is important to note that the load at each relative intensity was estimated for the initial 1RM assessment. The exact same warm-up and incremental 1RM assessment was repeated in session 2 and session 3 for each subject. Subjects abstained from alcohol, caffeine, and additional strenuous activity for 24 hours before each assessment period.
One-Repetition Maximum Assessment
Before each 1RM assessment, a standardized warm-up consisting of 8-minute dynamic lower-body mobilization exercises followed by 3 warm-up sets of 3 repetition with 25 kg of deadlift. Subsequently, subjects were provided a 3-minute passive rest. The 1RM assessment was performed in a custom-built power cage (Life Fitness; Hammer Strength, Rosemont, IL, USA) using a 20-kg barbell (Eleiko; Halmstad, Sweden, Europe). The 1RM protocol was adopted from Banyard et al. (2) and followed the guidelines proposed by McMaster et al. (29) comprising 3 repetitions at 20% 1RM, 3 repetitions at 40% 1RM, 3 repetitions at 60% 1RM, one repetition at 80% 1RM, and one repetition at 90% 1RM followed by 1RM attempts. After each successful 1RM attempt, barbell weight was increased in consultation with the subject by 1.0–5.0 kg until either no further weight could be lifted or the technique deviated significantly from the technical model (i.e., rounded lower or upper back, no full extension of hips and knees at top position, and initial full knee followed by hip extension). A maximum of five 1RM attempts excluding the submaximal repetitions up to and including 90% of 1RM was permitted for each subject. The best repetition within each set performed at 20, 40, and 60%, i.e. highest mean velocity, was considered for further analysis. Rest periods between submaximal sets and 1RM attempts comprised passive recovery of 3 and 5 minutes, respectively. Importantly, for each repetition, subjects were instructed to perform the concentric part of each repetition as fast and explosively as possible, but they were not allowed to lift their heels from the floor (23).
Barbell displacement with respect to time was monitored by a commercial linear position transducers (GymAware Power Tool; Kinetic Performance Technologies, Canberra, Australia, hardware version 6.1, firmware version 13.7523), which validity has previously been established (3,12,13). The main variable of interest was mean concentric velocity, which was obtained by differentiating the displacement-time curve. The tether of the GymAware device was attached to the middle of the barbell perpendicular to the ground and between the subject's feet. To record the measured data with the GymAware, the device was connected through Bluetooth to a freely available application (software version 1.2.1) on a tablet (iPad; Apple, Inc., CA, USA). Briefly, data captured from GymAware were sampled using a level-crossing detection method with changes in barbell position being time-stamped with a resolution of 35 microseconds, which was down sampled to 50 Hz for analysis (9). Furthermore, no filtering technique was applied.
Data in the text, figures, and tables are presented as mean values with SDs and 90% confidence limits (CLs), respectively. All data were first log transformed to reduce bias arising from nonuniformity. Data were analyzed for practical significance using magnitude-based inferences (5). This approach was used because traditional null hypothesis significance testing does not report the magnitude of a change/difference and produces higher inferential error rates (19). First, individualized regression equations were computed using Excel Software (Microsoft; Redmond, WA, USA). Validity was assessed by comparing the predicted 1RMs based on submaximal load ranges of session 3 with the actual 1RM established during session 2 by calculating the magnitude of correlation, the TEE, CV (%), and the standardized difference or effect size (ES, based on Cohen's effect size principle using pooled SDs). The intraclass coefficient (ICC, relative reliability), typical error of measurement (TE), CV (%, absolute reliability), and the ES were used as measures of short-term reliability of the actual 1RM (between sessions 1 and 2), v at 1RM (between sessions 1 and 2), and predicted 1RMs (between predicted 1RMs of sessions 2 and 3 based on v at 1RM of session 1). Threshold values for ES statistics were ≤0.2 (trivial), >0.2–0.6 (small), >0.6–1.2 (moderate), >1.2–2.0 (large), and ≥2.0 (very large) (20). Uncertainty in the differences were expressed as 90% CL and as probabilities that the true (unknown) difference was substantially greater or smaller than the smallest worthwhile change (SWC, 0.2 multiplied by the between-subject SD). Quantitative probabilities of beneficial/better or detrimental/poorer differences were evaluated qualitatively as follows: <5%, very unlikely; 5–25%, unlikely; 25–75%, possibly; 75–95%, likely; 95–99%, very likely; and >99%, almost certain. If the probabilities of a substantially greater and smaller differences were >5%, the effect was reported as unclear; otherwise, the effect was clear and interpreted as the magnitude of the observed value (20). The following scale was adopted to interpret the magnitude of the correlation coefficient (r): ≤0.1, trivial; >0.1–0.3, small; >0.3–0.5, moderate; >0.5–0.7, large; >0.7–0.9, very large; and >0.9–1.0, almost perfect. If the 90% CL overlapped trivial positive and negative values, the magnitude was reported as unclear; otherwise, the effect was clear and interpreted as the magnitude of the observed value (20). The magnitude of the ICC was assessed using the following thresholds: ≤0.20, very low; >0.20–0.50, low; >0.50–0.75, moderate; >0.75–0.90, high; >0.90–0.99, very high; and >0.99–1.00, extremely high. For clarity, all data presented in the text, tables, and figures are from the raw data set (Figure 1).
Relationship Between Relative Load and Velocity
When analyzing the whole dataset (132 data points for session 2 and 3), there was an almost perfect linear correlation between mean concentric velocity and relative loads (Figure 1). Similarly, there were almost perfect linear correlations between mean concentric velocity and relative loads for all submaximal load ranges (Figure 2). Individual linear correlation coefficients across all submaximal load ranges and subjects for session 2 and 3 ranged from r = −0.983 to −0.999 with an average of r = −0.998. The different reliability variables for predicted 1RMs, actual 1RMs, and v at 1RM are presented in Figure 3.
There were trivial changes for the actual 1RM and predicted 1RMs based on submaximal measures between sessions 2 and 3 (Table 1). Furthermore, we observed very high ICCs (range ICC: 0.95–0.997) and small CVs <5% (range CV: 1.9–4.4%) for the actual and all predicted 1RMs. Typical error ranged between 3.4 and 7.5 kg, with the best (i.e., lowest) TE noted for the actual 1RM and the worst (i.e., highest) reported for 40–80% of 1RM.
Finally, there was a possibly small change between repeated sessions for v at 1RM (ES = −0.30 and 90% CL [−0.78 to 0.17]). Calculated ICC coefficient (0.63 [0.19−0.86]) and CV (15.7% [11.3–26.1]) were moderate and high, respectively, whereas the observed TE was 0.029 m·s−1 (0.022–0.047).
Results of validity analysis are presented in Figure 4. When assessing the agreement between predicted and actual 1RM, the standardized difference was likely small for 20–60% of 1RM (1RM prediction: 185.3 ± 26.7 kg), whereas all other prediction models were possibly overestimating the actual 1RM (20–80%: 182.0 ± 26.3 kg, 20–90%: 181.3 ± 25.4 kg, 40–80%: 180.5 ± 27.2 kg, 40–90%: 180.2 ± 25.6 kg, and 60–90%: 180.1 ± 25.0 kg). The correlation coefficients for predicted and actual 1RMs were very large to almost perfect (range r: 0.88–0.95) for all submaximal load ranges. Similarly, CVs for all prediction models were below 5% (range: 3.3–4.4) except for 20–60% (CV = 5.3% and 90% CL [3.9–8.5]). By contrast, absolute TEEs revealed absolute errors ranging between 9.1 and 13.7 kg with 40–90% of 1RM showing the highest (i.e., lowest TEE) and 20–60% the lowest accuracy (i.e., highest TEE).
Finally, there was considerable individual variation between and within athletes for all 1RM prediction models (Figure 5).
In this study, we examined the reproducibility and accuracy of selected submaximal ranges of the load-velocity relationship to predict the 1RM in deadlift. Our main results are as follows: (a) All prediction models and the actual 1RM showed high reliability, whereas v at 1RM showed considerable variation between sessions (i.e., small ES, large TE and CV, and moderate ICC); (b) All prediction models tended to overestimate the actual 1RM by ∼5–10 kg with submaximal ranges performed at higher loads showing slightly higher accuracy; and finally, (c) Individual variation in the accuracy to predict the 1RM was high, with only 4 subjects displaying sufficient levels of accuracy, especially at higher submaximal loads.
Present findings indicate that 1RM is a highly reliable measure for maximum strength in deadlift in resistance-trained athletes as noted by trivial changes between sessions, alongside an almost perfect correlation and a low typical error of measurement expressed in both raw units and as a coefficient of variation. This is in line with previous findings examining maximum dynamic strength testing in various lifts such as back squat, bench press, squat jump, and power clean (29). For example, similar magnitudes in systematic change in the mean (ES = 0.03) and within-athlete day-to-day variation (TE = 2.9 kg and CV = 2.1%) were reported for the free-weight back squat by Banyard et al. (2). More importantly, however, this is the first study to report the reliability of maximum strength (i.e., 1RM) in the conventional deadlift. Present data also show good levels of reliability for all 1RM prediction models with the 20–90% of 1RM being the most reliable model. This is a very important and practical aspect because it directly determines the ability to monitor changes. Moreover, it is not the absolute TE (or CV) that matters, but rather the magnitude of the “noise” in relation to the usually observed change or smallest change that is considered meaningful in practice (i.e., SWC) (7). However, to date, results from this and other studies (2) focused on investigating the reliability of prediction models with athletes being tested in a nonfatigued state throughout all measurements making it difficult to draw any conclusions on the sensitivity of current prediction models. Nevertheless, mean propulsive velocity at relative loads seems to be unaffected by significant increases in strength levels (from 86.9 ± 15.2 kg to 94.5 ± 15.2 kg) after a 6-week strength program in the Smith machine bench press (15). More specifically, reliability measures revealed good long-term stability of mean propulsive velocity with high ICC (range: 0.81–0.91) and low CV (range: 0.0–3.6%) values attained at each relative load. It is important to highlight, however, that data from this study indicate that all predicted 1RMs were less reliable than the actual 1RM (i.e., TE of prediction models larger than for actual 1RM). Thus, if the 1RM prediction models were to be considered as a sensitive monitoring tool, changes in predicted 1RMs would need to mimic the change in the actual 1RM with the magnitude of change being bigger than the typical error of measurement of the predicted 1RM (i.e., 5.1 kg for the best prediction model 20–90%).
Recently, it has been suggested that the load-velocity relationship can be used to estimate daily 1RM for an individual by measuring the mean concentric velocity at 4–6 incremental loads ranging from 30 to 85% of the actual or estimated 1RM (23). Our results suggest, however, that 1RM predictions may overestimate the actual 1RM by a small but meaningful magnitude. More importantly, this bias was consistent across all selected submaximal loads, with typical errors of estimates ranging between 9.1 and 13.7 kg. Present data are in line with the very recent results of Banyard et al. (2), who reported predicted 1RMs in free-weight back squat across 3 different submaximal ranges (20–60%, 20–80%, and 20–90% of 1RM) to be moderately larger (ES: 0.74–1.09) than the actual 1RM. Similarly, prediction models were more precise in estimating the actual 1RM when lifts were performed at higher loads (i.e., 20–90% vs. 20–60% of 1RM). Greater accuracy in predictions at higher loads is well documented when applying the repetition-to-failure methods (34). However, asking the athlete to perform warm-up repetitions that are close to the 1RM for precise 1RM predictions contradicts somewhat the rationale of applying the load-velocity relationship method because the aim is to avoid the accumulation of fatigue by frequent maximal testing (23). Although the reasons for the differences in predicted and actual 1RM remain unclear, the high variability in v at 1RM might be related to these discrepancies. In fact, in highly standardized movements such as the bench press performed on the Smith machine with a 1.5-second pause between the eccentric and concentric phase, the v at 1RM seems to be a stable measure (15). It is important to highlight, however, that kinetic and kinematic parameters of movements performed in a concentric-only fashion or on a Smith machine are specific to that particular type of exercise, thus limiting the ecological validity of free-weight 1RMs. Nevertheless, Loturco et al. (26) have recently reported high levels of agreement between predicted and actual 1RM in both dynamic half-squat performed on the Smith machine for 4 different sports (range absolute bias: −0.9 to 1.3 kg and range CV: 0.30–0.75%) and dynamic free-weight bench press (absolute difference: 1.1 kg and CV = 1.15%) (24). The reasons for these differences to our study are more straightforward and can be explained by the different modeling processes used. Although we computed individualized prediction models based on velocity data of session 3 and using the previously established individual v at 1RM of session 2, Loturco et al. (25,26) used the v at 1RM from the same session to apply a group-based linear load-velocity regression including both submaximal loads and the actual 1RM. A limitation of the latter approach is that it does not take normal day-to-day variation of attained velocity at submaximal and maximal loads into account, thus limiting the practical usefulness of these findings. Therefore, if the load-velocity relationship was to be considered as a valid method in predicting daily 1RM, predictions would need to mimic the reliability statistics of the actual 1RM, which was not the case in this study.
Finally, prediction models showed substantial variation in accuracy between subjects across all submaximal loads (Figure 4). Irrespective of the submaximal load range, at least 5 subjects showed likely, very likely, or almost certain overestimation or underestimation compared with the actual 1RM. This is in line with the results of Banyard et al. (2), although individual differences for submaximal ranges performed at lower loads (e.g., 20–60% of 1RM) were larger in the study by Banyard et al. (−5.5 to 47.6% compared with −5.4 to 17.1%). Interestingly, our results reveal that there were no meaningful differences in prediction accuracy between subjects with respect to absolute and relative maximal strength levels, respectively. Thus, it seems that prediction accuracy is not directly affected by the athlete's strength level. This suggests that even subtle changes in movement execution might result in small changes in mean concentric velocity for a given submaximal load and thus can have a marked influence on the prediction accuracy (31).
This study presented a method for predicting the actual 1RM based on submaximal incremental loads. As such, conclusions about prediction models and results are limited to the characteristics of the cohort (i.e., male resistance-trained athletes), model (i.e., linear regression equations), and chosen exercise (i.e., conventional deadlift). Although we acknowledge that our proposed models were not able to accurately predict 1RM, it is not well understood whether contextual factors, such as age, sex, selected exercise, or training status, either affect reproducibility and accuracy of predictions or account for the observed individual variation in predictions.
The results of this study question the usefulness of linear prediction models based on submaximal loads to estimate daily 1RM in the free-weight deadlift. Although 1RM predictions were highly reliable between sessions, they consistently overestimated the actual 1RM on a group-based level by ∼5–10 kg (actual 1RM SWC ∼5 kg and TE ∼3 kg). Moreover, accuracy in 1RM predictions varied considerably between subjects with individual differences ranging from −14 to +29 kg. Thus, predictions for only 4 subjects displayed sufficient levels of agreement with differences between predicted and actual 1RM falling within the SWC of the actual 1RM. In this context, it is important to highlight the high within-subject variability of the v at 1RM between sessions making the computation of daily 1RM predictions to adjust training load according to daily readiness unlikely feasible. Future studies may examine the magnitude of changes in movement velocity at submaximal loads of free-weight exercises to determine the sensitivity of movement velocity as an indicator of acute and chronic fatigue. Importantly, the stability of the mean concentric bar velocity in a nonfatigued state must first be established to gather information about the normal within-athlete day-to-day variability.
The current results suggest that predictions of 1RM in well-trained strength athletes are highly variable when using mean concentric velocity of selected submaximal loads as predictors. As a result, prediction models moderately overestimated the actual 1RM. It is therefore recommended to not use actual 1RM and predicted 1RM interchangeably. Similarly, prediction models can unlikely be used as a monitoring tool to adjust daily training load aiming to account for daily readiness to train. It is therefore important for training prescription purposes to occasionally perform traditional 1RM assessments on which practitioners prescribe relative loads while monitoring potential changes for a given load through subjective markers such as rating of perceived exertion.
1. Balsalobre-Fernández C, Marchante D, Baz-Valle E, Alonso-Molero I, Jiménez SL, Muñóz-López M. Analysis of wearable and smartphone-based Technologies for the measurement of barbell velocity in different resistance training exercises. Front Physiol 8: 649, 2017.
2. Banyard HG, Nosaka K, Haff GG. Reliability and validity of the load-velocity relationship to predict the 1RM back squat. J Strength Cond Res 31: 1897–1904, 2017.
3. Banyard HG, Nosaka K, Sato K, Haff GG. Validity of various methods for determining velocity, force and power in the back squat. Int J Sports Physiol Perform: 12: 1170–1176, 2017.
4. Barnes MJ, Miller A, Reeve D, Stewart RJ. Acute neuromuscular and endocrine responses to two different compound exercises: Squat versus deadlift. J Strength Cond Res 12: 1170–1176, 2017.
5. Batterham AM, Hopkins WG. Making meaningful inferences about magnitudes. Int J Sports Physiol Perform 1: 50–57, 2006.
6. Bosquet L, Porta-Benache J, Blais J. Validity of a commercial linear encoder to estimate bench press 1RM from the force-velocity relationship. J Sports Sci Med 9: 459–463, 2010.
7. Buchheit M, Manouvrier C, Cassirame J, Morin JB. Monitoring locomotor load in Soccer: Is metabolic power, powerful? Int J Sports Med 36: 1149–1155, 2015.
8. Buckner SL, Jessee MB, Mattocks KT, Mouser JG, Counts BR, Dankel SJ, Loenneke JP. Determining strength: A case for multiple methods of measurement. Sports Med 47: 193–195, 2017.
9. Cormie P, McBride JM, McCaulley GO. Validation of power measurement techniques in dynamic lower body resistance exercises. J Appl Biomech 23: 103–118, 2007.
10. Cotterman ML, Darby LA, Skelly WA. Comparison of muscle force production using the Smith machine and free weights for bench press and squat exercises. J Strength Cond Res 19: 169–176, 2005.
11. Cressey EM, West CA, Tiberio DP, Kraemer WJ, Maresh CM. The effects of ten weeks of lower-body unstable surface training on markers of athletic performance. J Strength Cond Res 21: 561–567, 2007.
12. Crewther BT, Kilduff LP, Cunningham DJ, Cook C, Owen N, Yang GZ. Validating two systems for estimating force and power. Int J Sports Med 32: 254–258, 2011.
13. Garnacho-Castaño MV, López-Lastra S, Maté-Muñoz JL. Reliability and validity assessment of a linear position transducer. J Sports Sci Med 14: 128–136, 2015.
14. Giroux C, Rabita G, Chollet D, Guilhem G. What is the best method for assessing lower limb force-velocity relationship? Int J Sports Med 36: 143–149, 2015.
15. González-Badillo JJ, Sánchez-Medina L. Movement velocity as a measure of loading intensity in resistance training. Int J Sports Med 31: 347–352, 2010.
16. Hales ME, Johnson BF, Johnson JT. Kinematic analysis of the powerlifting style squat and the conventional deadlift during Competition: Is there a cross-over effect between lifts? J Strength Cond Res 23: 2574–2580, 2009.
17. Hetzler RK, Schroeder BL, Wages JJ, Stickley CD, Kimura IF. Anthropometry increases 1 repetition maximum predictive ability of NFL-225 test for Division IA college football players. J Strength Cond Res 24: 1429–1439, 2010.
18. Hopkins WG. A spreadsheet for monitoring an individual's changes and trend. Sportscience 21: 5–9, 2017.
19. Hopkins WG, Batterham AM. Error rates, decisive outcomes and publication bias with several inferential methods. Sports Med 46: 1563–1573, 2016.
20. Hopkins WG, Marshall SW, Batterham AM, Hanin J. Progressive statistics for studies in sports medicine and exercise science. Med Sci Sports Exerc 41: 3–13, 2009.
21. Izquierdo M, Ibañez J, Calbet JA, González-Izal M, Navarro-Amézqueta I, Granados C, Malanda A, Idoate F, González-Badillo JJ, Häkkinen K, Kraemer WJ, Tirapu I, Gorostiaga EM. Neuromuscular fatigue after resistance training. Int J Sports Med 30: 614–623, 2009.
22. Jidovtseff B, Harris NK, Crielaard JM, Cronin JB. Using the load-velocity relationship for 1RM prediction. J Strength Cond Res 25: 267–270, 2011.
23. Jovanovic M, Flanagan EP. Researched applications of velocity based strength training. J Aust Strength Cond 22: 58–69, 2014.
24. Kompf J, Arandjelović O. The sticking point in the bench press, the squat, and the deadlift: Similarities and differences, and their significance for research and practice. Sports Med 47: 631–640, 2017.
25. Loturco I, Kobal R, Moraes JE, Kitamura K, Cal Abad CC, Pereira LA, Nakamura FY. Predicting the maximum dynamic strength in bench press: The high precision of the bar velocity approach. J Strength Cond Res 31: 1127–1131, 2017.
26. Loturco I, Pereira LA, Cal Abad CC, Gil S, Kitamura K, Kobal R, Nakamura FY. Using bar velocity to predict the maximum dynamic strength in the half-squat exercise. Int J Sports Physiol Perform 11: 697–700, 2016.
27. Macht JW, Abel MG, Mullineaux DR, Yates JW. Development of 1RM prediction equations for bench press in moderately trained men. J Strength Cond Res 30: 2901–2906, 2016.
28. Mayhew JL, Johnson BD, Lamonte MJ, Lauber D, Kemmler W. Accuracy of prediction equations for determining one repetition maximum bench press in women before and after resistance training. J Strength Cond Res 22: 1570–1577, 2008.
29. McMaster DT, Gill N, Cronin J, McGuigan M. A brief review of strength and ballistic assessment methodologies in sport. Sports Med 44: 603–623, 2014.
30. Niewiadomski W, Laskowska D, Gąsiorowska A, Cybulski G, Strasz A, Langfort J. Determination and prediction of one repetition maximum (1RM): Safety considerations. J Hum Kinet 19: 109–120, 2008.
31. Preatoni E, Hamill J, Harrison AJ, Hayes K, Van Emmerik RE, Wilson C, Rodano R. Movement variability and skills monitoring in sports. Sports Biomech 12: 69–92, 2013.
32. Pritchard HJ, Tod DA, Barnes MJ, Keogh JW, McGuigan MR. Tapering practices of new Zealand's elite raw powerlifters. J Strength Cond Res 30: 1796–1804, 2016.
33. Reilly T, Morris T, Whyte G. The specificity of training prescription and physiological assessment: A review. J Sports Sci 27: 575–589, 2009.
34. Reynolds JM, Gordon TJ, Robergs RA. Prediction of one repetition maximum strength from multiple repetition maximum testing and anthropometry. J Strength Cond Res 20: 584–592, 2006.
35. Schwanbeck S, Chilibeck PD, Binsted G. A comparison of free weight squat to Smith machine squat using electromyography. J Strength Cond Res 23: 2588–2591, 2009.
36. Scott BR, Duthie GM, Thornton HR, Dascombe BJ. Training monitoring for resistance Exercise: Theory and applications. Sports Med 46: 687–698, 2016.
37. Suchomel TJ, Nimphius S, Stone MH. The importance of muscular strength in athletic performance. Sports Med 46: 1419–1449, 2016.
38. Swinton PA, Stewart AD, Keogh JWL, Agouris I, Lloyd R. Kinematic and kinetic analysis of maximal velocity deadlifts performed with and without the inclusion of chain resistance. J Strength Cond Res 25: 3163–3174, 2011.
39. Thompson BJ, Stock MS, Shields JE, Luera MJ, Munayer IK, Mota JA, Carrillo EC, Olinghouse KD. Barbell deadlift training increases the rate of torque development and vertical jump performance in novices. J Strength Cond Res 29: 1–10, 2015.
40. Wisloff U, Castagna C, Helgerud J, Jones R, Hoff J. Strong correlation of maximal squat strength with sprint performance and vertical jump height in elite soccer players. Br J Sports Med 38: 285–288, 2004.