Rating of Perceived Exertion and Velocity Relationships Among Trained Males and Females in the Front Squat and Hexagonal Bar Deadlift

Abstract

Odgers, JB, Zourdos, MC, Helms, ER, Candow, DG, Dahlstrom, B, Bruno, P, and Sousa, CA. Rating of perceived exertion and velocity relationships among trained males and females in the front squat and hexagonal bar deadlift. J Strength Cond Res 35(2S): S23–S30, 2021—This study examined the accuracy of intraset rating of perceived exertion (RPE) to predict repetitions in reserve (RIR) during sets to failure at 80% of 1 repetition maximum (1RM) on the front squat and high-handle hexagonal bar deadlift (HHBD). Furthermore, the relationship between RPE and average concentric velocity (ACV) during the sets to failure was also determined. Fourteen males (29 ± 6 years, front squat relative 1RM: 1.78 ± 0.2 kg·kg⁻¹, and HHBD relative 1RM: 3.0 ± 0.1 kg·kg⁻¹) and 13 females (30 ± 5 years, front squat relative 1RM: 1.60 ± 0.2 kg·kg⁻¹, and HHBD relative 1RM: 2.5 ± 0.3 kg·kg⁻¹) visited the laboratory 3 times. The first visit tested 1RM on both exercises. During visits 2 and 3, which were performed in a counterbalanced order, subjects performed 4 sets to failure at 80% of 1RM for both exercises. During each set, subjects verbally indicated when they believed they were at “6” and “9” on the RIR-based RPE scale, and ACV was assessed during every repetition. The difference between the actual and predicted repetitions performed was recorded as the RPE difference (RPEDIFF). The RPEDIFF was significantly (p < 0.001) lower at the called 9 RPE versus the called 6 RPE in the front squat for males (9 RPE: 0.09 ± 0.19 versus 6 RPE: 0.71 ± 0.70) and females (9 RPE: 0.19 ± 0.36 versus 6 RPE: 0.86 ± 0.88) and in the HHBD for males (9 RPE: 0.25 ± 0.46 versus 6 RPE: 1.00 ± 1.12) and females (9 RPE: 0.21 ± 0.44 versus 6 RPE: 1.19 ± 1.16). Significant inverse relationships existed between RPE and ACV during both exercises (r = −0.98 to −1.00). These results indicate that well-trained males and females can gauge intraset RPE accurately during moderate repetition sets on the front squat and HHBD.

Introduction

The most common methods of resistance training load prescription are percentage-based loading (i.e., 4 sets of 8 repetitions at 70% of 1 repetition maximum [1RM]) or the use of RM zones (¹⁹). However, a recent study from Cooke et al. (⁷) reported a large variation in repetitions (^1,7–26) performed on the back squat at 70% of 1RM in a cohort of 58 males and females. Thus, with percentage-based loading, some individuals could fail to complete prescribed training, while others may not receive a sufficient stimulus. Furthermore, fluctuations in an individual's day-to-day strength levels can cause missed repetitions within a predetermined percentage loading model. Because of these limitations, velocity-based training (VBT) (^10,40) and the repetitions in reserve (RIR)-based rating of perceived exertion (RPE) (^21,44) scale have emerged as alternative load prescription strategies. Specifically, VBT has various autoregulatory approaches: (²) performing repetitions with a specific load and terminating a set when a certain percentage of velocity loss (i.e., 10, 20, 30, or 40%) has occurred from the fastest repetition in the set (^{3,15,31–33,40}), performing repetitions in a set until a specific absolute velocity has been reached (e.g., terminating a set at 0.35 m·s⁻¹) (^12,28), and (⁴) performing a specific number of repetitions as determined by the velocity of the first repetition in a set (^3,12). In the RIR-based RPE scale, an RPE 10 corresponds to 0 RIR, RPE 9 corresponds to 1 RIR, RPE 8 corresponds to 2 RIR, etc. (⁴⁴), and load could be prescribed using RPE ranges (i.e., 3 sets of 10 repetitions between 6 and 8 RPE). Because VBT and RPE do not prescribe exact loads, these strategies account for both the interindividual variation in repetitions performed and the possibility of fluctuating daily strength levels in an effort to rectify the limitations of percentage loading.

Velocity-based training and RPE-based training research is limited compared with percentage- and RM-based load prescription (³⁸). However, preliminary longitudinal studies have shown additional strength benefits with both VBT and RPE-based loading (^9,16,23). Specifically, Dorrell et al. (⁹) observed significantly greater increases in bench press 1RM in a group of trained men using VBT for load prescription compared with percentage-based loading. Furthermore, Helms et al. (²³) reported a possibly meaningful effect size (ES) difference in favor of RPE loading versus percentage loading for back squat and bench press 1RM over 8 weeks, while Graham and Cleather (¹⁶) found significantly greater front squat and back squat 1RM increases with RIR-based RPE loading compared with percentage loading. Of the longitudinal studies, only Graham and Cleather (¹⁶) used a nonpowerlift (front squat), and none of the longitudinal studies used female subjects.

Despite the long-term benefits of autoregulation, the accuracy of RPE should be examined in individual exercises and specific populations since its rating is subjective. Indeed, trained males were not perfectly accurate when using intraset RPE to predict 1 RIR (±∼2 repetitions), 3 RIR (±∼3.5 repetitions), or 5 RIR (±∼5 repetitions) in the back squat (⁴⁵). However, similar to the longitudinal studies, the majority of the acute RPE/velocity relationship data have been relegated to the squat, bench press, and deadlift in male subjects. Helms et al. (²²) reported very strong inverse relationships between average concentric velocity (ACV) and percentage of 1RM (r = −0.90 to −0.92) on the back squat, bench press, and deadlift among mostly male powerlifters (males: n = 12; females: n = 3). Morán-Navarro et al. (²⁸) reported 95% confidence intervals for ACV for the back squat and bench press corresponding to 2 RIR among 30 trained males. Using data from Morán-Navarro et al. (²⁸), bench press could be prescribed for 4 sets of as many repetitions as possible until ACV reaches 0.26–0.28 m·s⁻¹, which would on average be 2 RIR. Although it is ideal to individualize RPE/velocity profiles, it is uncommon to do so in team sport settings. Therefore, establishing normative relationships would be a valuable tool. Training to failure has been shown to prolong recovery and fatigue more than nonfailure training with no greater improvements in strength or hypertrophy (³⁰); thus, RPE/velocity relationships can be used to avoid failure training and the subsequent elongated recovery process.

In team settings where group velocity profiles are common, the front squat and high-handle hexagonal bar deadlift (HHBD) are often used instead of the powerlifts because of their decreased erector spinae activity and upright torso position (^37,42). Thus, establishing RPE/velocity relationships on these exercises, in both males and females, would allow for greater specificity and individualization of programming. Therefore, the purpose of this study was to examine the accuracy of intraset RPE (i.e., RIR predictions) and establish RPE/velocity relationships in the front squat and HHBD in well-trained males and females. It was hypothesized that intraset RPE accuracy would improve closer to failure and that there would be a very strong inverse relationship between RPE and ACV in both sexes and in both the front squat and HHBD.

Methods

Experimental Approach to the Problem

Subjects performed 3 sessions over the course of the study. During session 1, subjects performed a standardized 5-minute dynamic warm-up, followed by validated (⁴³) 1RM testing for the front squat and HHBD. Forty-eight hours later, subjects completed session 2, in which they performed 4 sets to volitional failure in either the front squat or HHBD at 80% of 1RM. During each set, subjects verbally indicated when they believed they were at a “6” and “9” RPE (i.e., 4 and 1 RIR). Session 3, which occurred 48 hours after session 2, was the same as session 2 except subjects performed the exercise which had not yet been completed. The order in which the exercises were performed was counterbalanced during both the 1RM testing session and the experimental sessions. A timeline of the protocol can be seen in Table 1.

Table 1 - Timeline of events.^*

Monday (session #1)	Tuesday	Wednesday (session #2)	Thursday	Friday (session #3)
1RM testing	Off	ES	Off	ES
AM
Familiarization
PAQ
PAR-Q

^*1RM = 1 repetition maximum; AM = anthropometric measurements. PAQ = Physical Activity Questionnaire; PAR-Q = Physical Activity Readiness Questionnaire; ES = experimental session (4 sets to volitional failure at 80% of 1RM).

Subjects

An a priori power analysis (G*Power v. 3.1.5.1) indicated that 27 subjects were required for the study. This calculation was based on a medium ES (Cohen's d = 0.50), an alpha level of 0.05, and a ß-value of 0.8 using a paired-sample t-test statistical approach. Twenty-seven subjects (14 male and 13 female) were recruited from local fitness centers in Regina, Saskatchewan, Canada (Table 2). To be eligible to participate in the study, subjects were performing resistance training for ≥6 months before the start of the study and were able to front squat 1RM ≥1.5x and HHBD 1RM ≥2x their body mass. Before 1RM testing, subjects who had any contraindications to exercise (e.g., heart disease, serious musculoskeletal disorders, etc.), as determined through the Health History Questionnaire, were excluded. The Research Ethics Board at the University of Regina approved this study, and subjects provided written consent before beginning participation.

Table 2 - Descriptive characteristics of male and female subjects.^*†

Sex	Age (y)	Body mass (kg)	Height (m)	FS 1RM (kg)	HHBD 1RM (kg)	Relative FS 1RM (kg·kg−1)	Relative HHBD 1RM (kg·kg−1)
Males (n = 14)	28.9 ± 5.7	89.9 ± 11.9	1.74 ± 0.1	160.2 ± 24.3	267.1 ± 38.0	1.78 ± 0.2	3.0 ± 0.3
Females (n = 13)	30.1 ± 5.4	64.2 ± 9.0	1.60 ± 0.1	102.9 ± 14.8	159.2 ± 22.3	1.6 ± 0.2	2.5 ± 0.3
Combined (n = 27)	29.5 ± 5.6	77.1 ± 10.5	1.70 ± 0.1	131.6 ± 19.6	213.2 ± 30.2	1.69 ± 0.2	2.75 ± 0.3

^*FS = front squat; HHBD = high-handle hexagonal bar deadlift; 1RM = 1 repetition maximum.

^†Values are mean ± SD.

Procedures

One Repetition Maximum

All 1RM testing was performed based on procedures previously described (⁴³). Front squat and HHBD testing took place on the same day and were judged by a Saskatchewan Powerlifting Association Provincial referee. Although the exercises in this study were not powerlifting competition lifts, the front squat depth and deadlift lockout position were performed in accordance with the International Powerlifting Federation (IPF) regulations, and only IPF-approved “unequipped” lifting material aids were permitted (knee sleeves and weightlifting belt) (¹). For the front squat, subjects were required to reach a depth where the hip crease passed below the top of the knee when viewed from the lateral aspect. To begin the descent for the front squat, subjects received the verbal command “squat” from the researcher, and after the completion of the concentric phase, the verbal command “rack” was given for the subject to return the barbell to the rack. Next, the deadlift was deemed successful if the subject stood erect, with their shoulders and hips locked out, and if there was no downward movement of the bar during the ascent. Subjects were given the verbal command “down” to return the hexagonal bar back to the floor after the completion of the concentric phase. Before 1RM testing commenced, subjects gave the investigator their previous best for weight and reps in each exercise, so the investigator could calculate an estimated 1RM (e1RM) using the Epley equation (⁸). Testing for each exercise began by the subjects performing 5 repetitions at 20% of their e1RM, followed by 3 repetitions at 50% of e1RM, then 2 repetitions at 70% of e1RM, 1 repetition at 80% of e1RM, and finally, 1 repetition at 90% of e1RM. After the final warm-up set at 90% e1RM, subsequent increases in 1RM attempt were determined at the investigator's discretion. To aid attempt selection, ACV and RPE were recorded on each repetition after the final warm-up set at 90% of e1RM. Furthermore, 5 minutes of rest was given between 1RM attempts and 10 minutes of rest was given before starting the 1RM procedure for the following exercise. A lift was accepted as a 1RM if (a) the subject reported an RPE (1–10; ^{Ref. 44}) score of “10” and the investigator determined that any increase in load would result in a failed attempt or jeopardize the subject's safety, (b) the subject reported a “9.5” on the RPE scale and missed the subsequent load increase of 2.5 kg or less, and (c) a subject reported an RPE of “9” or lower on the RPE scale and failed the subsequent attempt with a load increase of 5 kg or less. Each successive increase in load after the final warm-up set of 90% e1RM was required to be less than or equal to the previous attempts increase in load. A Rogue Ohio Bar was used for the front squat, and Legend Hexagonal Bar (Rogue Fitness, Columbus, OH) was used for the HHBD. Eleiko lifting discs (Eleiko USA, Chicago, IL) calibrated to the nearest 0.25 kg were used.

Experimental Sessions

During the experimental sessions (sessions 2 and 3), subjects performed 4 sets to volitional failure at 80% of 1RM in the front squat and HHBD. To enhance ecological validity, subjects were instructed to perform both the eccentric and concentric phases of each repetition at their own pace and performed each repetition within 5 seconds of the previous repetition. Only 1 exercise was performed in each session, and the order in which exercises were performed was counterbalanced. Each session began with a standardized 5-minute dynamic warm-up consisting of bodyweight movements to prepare them to squat and deadlift, followed by an exercise-specific warm-up, which consisted of 5 repetitions at 20% of 1RM, and then 3 repetitions at 50% of 1RM, which was followed by 5 minutes of rest. Next, the training protocol of 4 sets to volitional failure at 80% of 1RM was performed with 5 minutes of rest between sets. Repetitions were only considered valid if they were executed to the standard set out by the IPF (¹). During all sets to failure, subjects verbally indicated when they believed they had reached both a 6 and a 9 RPE (4 and 1 RIR). Furthermore, ACV was assessed on every repetition of every set to failure.

Intraset Rating of Perceived Exertion Accuracy

As previously stated, subjects indicated during all sets to failure when they believed they had reached a 6 RPE and 9 RPE and then continued to failure in accordance with previous procedures (⁴⁵). The investigator manually recorded both the predicted and actual amount of repetitions performed for analysis. The difference between the actual and predicted repetitions performed (actual repetitions − predicted repetitions) was recorded as the RPE difference (RPEDIFF) for both intraset RPEs as an absolute value. For example, if a subject reported a 9 RPE after the seventh repetition (i.e., predicted they could do 8 repetitions), but actually completed 10 repetitions, then the investigator recorded an RPEDIFF of 2 (10–8). In addition, the RIR-based RPE scale (⁴⁴) was shown and explained to all subjects before each exercise.

Average Concentric Velocity

The ACV (m·s⁻¹) of the barbell was assessed by the previously validated (¹⁴) Open Barbell System Version 3 (OBS3) on every repetition during the experimental sessions and all 1RM attempts. The OBS3 was synced with a tablet application to display the ACV of all repetitions performed. The OBS3 was used in accordance with the instructions set out by the manufacturer, so that when it was attached to the barbell just inside of the “sleeve,” and perpendicular angle was achieved during all repetitions.

Statistical Analyses

There was a 3-step procedure for analyses. First, to examine the accuracy of the called 6 RPE versus called 9 RPEs, a paired t-test was used within each set. This was followed by the fit of a generalized linear model to the data that accounted for the clustering of observations within subjects across sets and allowed for comparisons across both conditions (9 versus 6 RPE callouts). Furthermore, to determine the magnitude of difference between the RPEDIFF at the called 9 RPE versus called 6 RPE, Hedges g ES were calculated as ES = [(post-exercise mean − pre-exercise mean)/SD_pooled], then a small sample size correction was used (²⁰), which is recommended when n < 50. The ES magnitudes were interpreted as outlined by Cohen (⁶). Second, ES was calculated to examine the magnitude of difference between the RPEDIFF values at the called 6 and 9 RPEs.

For velocity analysis, mean of the ACV for each of the last 4 repetitions was calculated in Excel (Microsoft Corporation, Redmond, WA) across all 4 sets for both males and females and for both exercises. For example, the ACV of the fourth to last successful repetition (i.e., the seventh repetition of a 10-repetition set) of each set for all males in the front squat was averaged to obtain the mean ACV at 7 RPE in the front squat among males. This analysis was conducted for both sexes in both exercises, as well as a combined average of both sexes for both lifts. The mean of the ACVs was also calculated at 100% of 1RM for both males, females, and the combined cohort. Finally, a Pearson's Product-Moment correlation was calculated between RPE and ACV for males, females, and their combined mean values. Significance was set at p ≤ 0.05.

Results

Repetitions Performed

The number of repetitions performed on each set, for both males and females, on both the front squat and HHBD is presented in Table 3.

Table 3 - Repetitions Performed.^*†

Set	Front squat			High-handle hexagonal bar deadlift
	Males	Females	Combined	Males	Females	Combined
1	8.42 ± 1.78	9.92 ± 2.40	9.14 ± 2.08	10.07 ± 2.76	9.08 ± 2.25	9.59 ± 2.51
2	7.00 ± 1.13	8.46 ± 2.15	7.70 ± 1.62	8.64 ± 2.98	8.69 ± 1.65	8.66 ± 2.34
3	5.92 ± 1.08	7.85 ± 2.30	6.85 ± 1.67	7.07 ± 2.40	8.54 ± 2.73	7.78 ± 2.56
4	5.42 ± 1.62	6.92 ± 1.98	6.14 ± 1.79	6.85 ± 2.08	8.00 ± 2.58	7.40 ± 2.32

^*Combined = weighted average of males and females.

^†Values are mean ± SD.

Rating of Perceived Exertion Difference (Females)

For females in the front squat, there was a significantly lower RPEDIFF (i.e., more accurate) at the called 9 versus 6 RPE for set 1 (p = 0.014, ES = 1.02), set 3 (p = 0.006, ES = 1.31), and set 4 (p = 0.038, 0.74), but not for set 2 (p = 0.082, ES = 0.50) (Table 4). Across all 4 sets, using the generalized linear model, the RPEDIFF was also significantly lower for the called 9 RPE versus 6 RPE (p < 0.001, ES = 0.94).

Table 4 - RPEDIFF results—females.^*†

Exercise	RPE	Set 1 RPEDIFF	Set 2 RPEDIFF	Set 3 RPEDIFF	Set 4 RPEDIFF	Mean RPEDIFF
FS	6	1.85 ± 1.79	0.38 ± 0.49	0.50 ± 0.50	0.69 ± 0.72	0.86 ± 0.88
	9	0.42 ± 0.49‡	0.15 ± 0.36	0 ± 0‡	0.18 ± 0.57‡	0.19 ± 0.36‡
HHBD	6	1.58 ± 1.26	1.17 ± 0.90	1.17 ± 1.46	0.85 ± 1.03	1.19 ± 1.16
	9	0.31 ± 0.46‡	0.08 ± 0.28‡	0.23 ± 0.42‡	0.23 ± 0.58‡	0.21 ± 0.44‡

^*FS = front squat. HHBD = high-handle hexagonal bar deadlift; RPE, rating of perceived exertion.

^†Values are mean ± SD. RPEDIFF = rating of perceived exertion difference (actual – predicted RPE).

^‡Significantly lower RPEDIFF (i.e., more accurate RPE call) compared with RPEDIFF at the called 6 RPE within the same exercise, p < 0.05. Mean RPEDIFF = average RPEDIFF across all 4 sets within each exercise.

For the HHBD, females exhibited a significantly lower RPEDIFF at the called 9 RPE versus 6 RPE for all individual sets (set 1: p = 0.003, ES = 1.26; set 2: p = 0.002, ES = 1.54; set 3: p = 0.042, ES = 0.83; and set 4: p = 0.005, ES = 0.70) and when all 4 sets were averaged together (p < 0.001, ES = 1.05) (Table 4).

Rating of Perceived Exertion Difference (Males)

Males exhibited a significantly lower RPEDIFF (i.e., more accurate) at the called 9 RPE versus 6 RPE in all individual sets in the front squat (set 1: p = 0.046, ES = 0.83; set 2: p < 0.001, ES = 1.95; set 3: p = 0.017, ES = 0.60; and set 4: p = 0.012, ES = 1.07) and across all 4 sets averaged together (p < 0.001, ES = 1.05) (Table 5).

Table 5 - Rating of perceived exertion callout accuracy—males.^*†

Exercise	RPE	Set 1 RPEDIFF	Set 2 RPEDIFF	Set 3 RPEDIFF	Set 4 RPEDIFF	Mean RPEDIFF
FS	6	0.58 ± 0.86	0.75 ± 0.43	0.58 ± 0.76	0.92 ± 0.76	0.71 ± 0.70
	9	0 ± 0‡	0 ± 0‡	0.17 ± 0.37‡	0.18 ± 0.39‡	0.09 ± 0.19‡
HHBD	6	1.57 ± 1.40	1.29 ± 1.58	0.58 ± 0.76	0.54 ± 0.75	1.00 ± 1.12
	9	0.29 ± 0.45‡	0.31 ± 0.46‡	0.23 ± 0.58	0.15 ± 0.36	0.25 ± 0.46‡

^*FS = front squat. HHBD = high-handle hexagonal bar deadlift; RPE = rating of perceived exertion.

^†Values are mean ± SD. RPEDIFF = rating of perceived exertion difference (actual – predicted RPE).

^‡Significantly lower RPEDIFF (i.e., more accurate RPE call) compared with RPEDIFF at the called 6 RPE within the same exercise, p < 0.05. Mean RPEDIFF = average RPEDIFF across all 4 sets within each exercise.

For the HHBD in males, there was a significantly lower RPEDIFF in sets 1 (p = 0.008, ES = 1.07) and 2 (p = 0.033, ES = 0.73) at the called 9 RPE versus the called 6 RPE; however, there was no significant difference in RPEDIFF in sets 3 (p = 0.054, ES = 0.45) and 4 (p = 0.137, ES = 0.57). Furthermore, averaging across all sets, the RPEDIFF was significantly lower at the called 9 RPE versus 6 RPE (p = 0.004, ES = 0.77) (Table 5).

Rating of Perceived Exertion/Average Concentric Velocity Relationships

Both males and females, and the combined cohort, had nearly perfect inverse relationships (r = −0.97 −1.00) between RPE and ACV (Table 6). On the front squat, males displayed faster ACVs at the actual 7 RPE (p < 0.001), 8 RPE (p = 0.002), 9 RPE (p = 0.002), and 10 RPE (p = 0.016) repetitions. However, for the HHBD, the ACV at 10 RPE was significantly faster in females versus males (p = 0.027), while ACVs of the 7 (p = 0.612), 8 (p = 0.608), and 9 (p = 0.132) RPE repetitions were not different between the sexes. The ACV at 1RM was similar between the sexes for both the front squat (p = 0.309) and HHBD (p = 0.215).

Table 6 - Rating of perceived exertion/ACV profiling.^*†

RPE	Front squat, ACV (m·s−1)			High-handle hexagonal bar deadlift, ACV (m·s−1)
	Males	Females	Combined	Males	Females	Combined
7	0.51 ± 0.05‡	0.43 ± 0.06	0.47 ± 0.07	0.33 ± 0.04	0.34 ± 0.06	0.33 ± 0.05
8	0.50 ± 0.06‡	0.42 ± 0.06	0.46 ± 0.07	0.32 ± 0.05	0.33 ± 0.05	0.32 ± 0.05
9	0.46 ± 0.06‡	0.38 ± 0.06	0.42 ± 0.07	0.29 ± 0.05	0.32 ± 0.05	0.30 ± 0.05
10	0.40 ± 0.06‡	0.34 ± 0.06	0.37 ± 0.07	0.24 ± 0.06	0.29 ± 0.05‡	0.27 ± 0.06
1RM	0.29 ± 0.05	0.27 ± 0.05	0.28 ± 0.05	0.18 ± 0.03	0.20 ± 0.05	0.19 ± 0.05
R-value	−0.97	−0.99	−0.98	−0.96	−0.97	−1.00

^*ACV = average concentric velocity (m·s⁻¹); RPE = rating of perceived exertion; 1RM = 1 repetition maximum.

^†Values are mean ± SD. Combined = weighted average of males and females.

^‡Significantly faster ACV compared with the opposite sex within the same exercise, p < 0.05.

Discussion

The main findings of this study were in support of our hypotheses and were as follows: (a) The RPEDIFF was lower at the called 9 RPE versus called 6 RPE for both the front squat and HHBD, and (b) there were very strong and inverse relationships in both sexes between ACV and RPE for both exercises tested. In addition, we observed faster ACVs for males compared with females during each of the last 4 repetitions of the front squat; however, females exhibited a faster ACV than males during the last repetition of the HHBD. Furthermore, at the called 9 RPE, the RPEDIFF for every set was <1, and the RPEDIFF at the called 6RPE was <2 for every set indicating accurate prediction of RIR. These findings support the usage of RIR-based RPE for load prescription and establish RPE/ACV relationships in men and women in the front squat and HHBD.

Previous data have reported that intraset RIR predictions are more accurate closer to failure on the back squat (⁴⁵), chest press, and leg press (¹⁷). Specifically, Zourdos et al. (⁴⁵) had well-trained males indicate when they believed they had reached RPEs of 5, 7, and 9 during a squat set to failure at 70% of 1RM and reported RPEDIFF values of 5.15 ± 2.92, 3.65 ± 2.46, and 2.05 ± 1.73, respectively. Although this study agrees with the data from Zourdos et al. (⁴⁵), in that intraset RIR predictions are more accurate closer to failure, the RPEDIFF in this study was considerably lower. For example, the highest reported RPEDIFF at the called 9 RPE in this study was 0.42 ± 0.49 in females on set 1 of the front squat (Table 4). Furthermore, at the called 9 RPEs, there were 3 instances of an RPEDIFF of 0 ± 0 (Tables 4 and 5), indicating that all subjects gauged intraset RPE perfectly. At the called 6 RPE, the findings showed more inaccurate RPEDIFF compared with the called 9 RPE, with a high RPEDIFF of 1.85 ± 1.79 on set 1 of the front squat. It seems likely that the lower RPEDIFF in this study is due to fewer repetitions performed per set than in previous research. Specifically, this study used 80% of 1RM, which resulted in fewer repetitions per set (Table 3) than the 16 ± 4 reported in Zourdos et al. (⁴⁵), who observed that more repetitions in a set were related to more inaccurate intraset RIR predictions (p = 0.003–0.011). Furthermore, Hackett et al. (¹⁸) observed trained subjects to predict RIR during a 3.0 ± 1.6 repetition set on the chest press and a 4.7 ± 2.9 repetition set on the leg press, both at 80% of 1RM, to accuracies of ±0.6 and ±1.6 repetitions. Therefore, it is not surprising that the fewer repetitions per set (∼5–10 repetitions), in this study, compared with Zourdos et al. (⁴⁵) led to more accurate intraset RPEs. In addition, since on a group level, the intraset RIR predictions in this study were in some cases perfect, it is possible that the front squat and HHBD are easier to gauge intraset RPE than previously tested exercises. However, that notion cannot be fully supported because there was no direct comparison in this study to the previously used exercises (i.e., back squat, chest press, and leg press). Overall, despite the accurate intraset RIR predictions in this study, RIR predictions significantly decrease in accuracy at lower RPEs; thus, we advise caution with using the RPE to terminate a set at ≤ 5RPE.

An additional novel aspect of this study is the RPE/ACV profiles. Recently, longitudinal studies have demonstrated that using both RPE (^16,23) and ACV (¹⁰) for load prescription can augment strength on compound movements. However, each of these autoregulatory tools have limitations. Specifically, although RPE can gauge proximity to failure, the rating is subjective. Although this study observed that subjects predicted RIR with precision, these ratings were not always perfect and previous data (²⁴) have demonstrated a larger error associated with intraset RPE (RPEDIFF ∼ 2–5) on the back squat exercise. Furthermore, Hackett et al. (¹⁷) observed that intraset RIR predictions were not perfectly accurate on the leg press and chest press exercises, while Steele et al. (³⁶) found that predicting the number of repetitions performed before a set was not perfectly accurate. In addition, the accuracy of predicting RIR has been shown to be improved when fewer repetitions in a set are performed and the prediction is made closer to muscular failure (⁴⁵) and with greater training experience (³⁶). Therefore, if solely using RPE to program load in a team setting, it is likely that some athletes will rate RPE inappropriately and potentially fail on a prescribed set because of a myriad of reasons. Moreover, since training to failure can elongate the recovery period from resistance training by 24–48 hours compared with volume-equated nonfailure training (³⁰), using RPE to prescribe training load may not be advisable for all athletes. Conversely, velocity is an objective load prescription tool in which individual velocity profiles can be created (⁹); however, unlike RPE velocity is not inherently individualized. Therefore, since creating individual velocity profiles can be time-consuming, prescribing a velocity range for all repetitions in a set (i.e., 0.35–0.60 m·s⁻¹) or prescribing a specific velocity loss percentage (i.e., 40%) may also result in different proximities to failure per set between athletes. For example, Rodriquez-Rosell et al. (³³) observed that at a 40% velocity loss during a set to failure at 80% of 1RM on the Smith machine squat, subjects completed on average ∼78% of their total repetitions. However, there was an SD of 10.3% indicating a large interindividual variation in the number of repetitions performed. Furthermore, Pareja-Blanco et al. (³¹) reported that subjects failed on 56.3% of sets during an 8-week training study with 40% velocity loss on the squat, while Weakley et al. (⁴⁰) observed a wide range of repetitions performed (20–55) in trained men over 5 sets at a 30% velocity loss when using a load in which the first repetition ACV corresponded to 0.70 m·s⁻¹. Therefore, the present RPE/ACV relationships (Table 6) provide a framework to mitigate the possibility of training to an undesired proximity to failure. To illustrate, in women, an ACV of 0.43 m·s⁻¹ on the front squat was associated with a 7 RPE (3 RIR). In this case, sets with 80% of 1RM could be prescribed on the front squat, in which each set should be terminated at 0.43 m·s⁻¹. To verify the desired proximity to failure and individualize the velocity profile, the athlete could also record an RPE after 0.43 m·s⁻¹ is reached. If the RPE is not 7 at 0.43 m·s⁻¹, then the target velocity could be adjusted to associate with the appropriate RPE for this athlete. Therefore, these RPE/ACV relationships provide normative values for both males and females on the front squat and HHBD that can be used across a group setting, and the usage of RPE can aid in quickly individualizing the relationship between ACV and RIR for the current training session to individualize the training stimulus.

Most previous data related to RIR-based RPE/velocity relationships have been conducted with mostly male subjects (^22,28,30,44). Therefore, a significant novelty of this study is the inclusion of well-trained females and direct comparison of ACV values between the sexes. The ACV on the final 4 repetitions (RPEs 7–10) of the front squat was faster in males than in females, whereas the ACV on the final repetition (RPE 10) of the HHBD was faster in females compared with males. However, there were no differences between the sexes for ACV values at a 1RM in either exercise. These findings highlight that RPE/ACV relationships may be sex-dependent, but it is not always the case that 1 specific sex has a faster ACV than the other. Furthermore, although some sex differences were present, there were no differences in ACV at 1RM on either exercise. Torrejón et al. (³⁹) reported a slower velocity at 1RM in males (0.17 ± 0.04 m·s⁻¹) vs. females (0.21 ± 0.05 m·s⁻¹) in the Smith machine bench press and indeed attributed that finding to sex. However, Torrejon et al. also reported higher training age and relative strength in males (training age = 7 ± 2 years; relative bench press = 1.17 ± 0.19 times body mass) versus females (training age = 1 ± 2 years; relative bench press = 0.66 ± 0.13 times body mass). Furthermore, previous studies have reported significantly slower ACV at 1RM in both the squat (⁴⁴) and bench press (³⁰) in experienced lifters versus novice lifters, citing enhanced neuromuscular efficiency in the experienced lifters as the reason. Therefore, the difference in ACV at 1RM between the sexes in Torrejón et al. (³⁹) was potentially due to training status and not inherent sex differences. This assertion is supported by this study, which reports no difference in relative strength (Table 2) or ACV at 1RM between males and females (Table 6). In addition, the lower ACVs in females in the front squat are potentially due to the fact that 9 females had a strong training background in CrossFit and Olympic lifting versus 3 males. It may be that the potentially higher training status for females on the front squat led to the slower front ACV values.

Previous data have reported load-velocity relationships in the front squat (^27,35); however, no study has examined ACVs on each repetition during a set to failure on the front squat and HHBD; thus, this study is the first to create RPE/ACV profiles on the front squat and HHBD. The fact that the ACV at 1RM during the front squat (0.28 ± 0.05 m·s⁻¹) was faster compared with 1RM ACV in the HHBD (0.19 ± 0.05 m·s⁻¹) is not surprising. Helms et al. (²²) reported significantly faster 1RM ACVs in the back squat (0.23 ± 0.05 m·s⁻¹) versus the deadlift (0.14 ± 0.05 m·s⁻¹). This discrepancy between squat and deadlift variations is possibly due to a diminished eccentric component on the deadlift or potentially different ranges of motion between the exercises. Nonetheless, deadlift ACV is slower than squat ACV at a similar proximity to failure. Although the stretch-shortening cycle was likely not eliminated between repetitions as subjects were not required to wait 4 seconds (the time it takes for the benefit of the stretch-shortening cycle to completely dissipate (⁴¹)), the barbell was required to come to a complete stop on the ground between deadlift repetitions. In addition, the ACV at 1RM (0.28 ± 0.05 m·s⁻¹) in the front squat was 0.09 m·s⁻¹ slower than the ACV during the last repetition (0.37 ± 0.07 m·s⁻¹) of the sets to failure in the front squat, despite both a 1RM and final repetition of a moderate repetition set both being a 10 RPE. Therefore, it is important to note that in the front squat, a 1RM ACV should not be used interchangeably with the ACV at a 10 RPE during a set to failure at a moderate intensity. On the HHBD, however, the 1RM ACV and ACV on the last repetition of the sets to failure was only different by 0.01 m·s⁻¹ for the combined cohort. There is a paucity of previous literature on the agreement between the ACV at a 10 RPE during a set to failure and the ACV during a 1RM test. García-Ramos et al. (¹³) recently reported similar ACVs at a 10 RPE during sets to failure at ∼75–90% of 1RM compared with ACV during a 1RM test on the Smith machine bench press. Therefore, the potential similarity between ACV during the final repetition of a set to failure and the ACV at 1RM could be exercise-specific.

Several limitations existed in this study. First, the RPE accuracy and RPE/ACV relationships presented in this study should not be extrapolated to other exercises. Indeed, it has been established that load-velocity profiles are exercise-dependent (¹¹). Caution should also be used when extrapolating the presently reported RPE/ACV relationships to other percentages of 1RM, as these relationships could vary. However, Morán-Navarro et al. (²⁸) reported RIR/ACV relationships to be statistically similar during sets to failure at 65, 75, and 85% of 1RM on the back squat and bench press, although it cannot be known whether that principle would hold true on the front squat and HHBD. Furthermore, it seems likely that the accuracy of the intraset RPEs would have been diminished if a lower relative intensity (i.e., higher repetition sets) was used. It has previously been reported that novice lifters record less accurate RPEs and different ACVs than trained lifters (^29,44); thus, the present results should only be applied to strength-trained athletes. In addition, although this study did have appropriate statistical power, when males and females were analyzed separately, it is likely that individual sex analyses were underpowered. Finally, only 48 hours of rest was allotted between the experimental sessions, despite recent data demonstrating that sets to failure at 80% of 1RM on the squat, bench press, and deadlift could cause fatigue for up to 72 hours (⁴). However, although not reported in the results, we analyzed number of repetitions performed in both exercises between individuals who performed the exercise in the first experimental sessions, and those who performed the exercise in the second experimental session, and observed no significant difference in repetitions performed. Therefore, we do not believe fatigue from session-to-session was an issue.

In summary, this study confirms that intraset RIR-based RPE ratings are more accurate closer to failure in males and females. Moreover, athletes in this study perfectly predicted RIR on various occasions, and although individualized RPE/ACV are preferable to group profiles, these findings provide normative RPE/ACV profiles on the front squat and HHBD. Importantly, this study is the first to demonstrate the accuracy of intraset RIR-based RPE ratings in women and in the front squat and HHBD exercises.

Practical Applications

These finding reveal that coaches and athletes can effectively program the front squat and HHBD using RIR-based RPE across a team setting to inherently individualize proximity to failure, which cannot be achieved with percentage-based loading charts. For example, training could be prescribed as 3 sets at 80% of 1RM with repetitions performed until a 9 RPE is reached. This method, previously termed RPE Stop (²⁵), is similar to terminating a set when an absolute velocity threshold has been reached and could be used with near-perfect accuracy according to the present results. Furthermore, this study provides guidance on individualizing velocity profiles on the front squat and HHBD. Specifically, athletes could perform a set to failure at 80% of 1RM and retroactively assign an RPE to each velocity over the final 4 repetitions just as in this study. In this way, an individual athlete would now have the exact velocity, which corresponds to an exact RPE, and an athlete could then perform repetitions until a specific ACV is met and then terminate the set to precisely control for proximity to failure. The specific intensity (i.e., 80% of 1RM) could be determined from a previous 1RM test or by using linear regression to predict 1RM using submaximal velocity and then calculating 80% of the prediction; however, previous results are equivocal regarding the accuracy of 1RM predictions in free-weight barbell exercises (^2,5,26,34). After that baseline assessment, coaches and athletes would know how many RIR an athlete had at each velocity. This information could be used to individualize velocity profiles and objectively quantify RIR to equate for session stimulus between athletes. It should be emphasized that the subjects in this study were experienced lifters and largely proficient in both the front squat and HHBD. Therefore, we encourage coaches who wish to implement RPE-based training in a team setting, to do so only with athletes who are experienced in the exercises being used. Importantly, in a group of athletes training as a team, RPEs could be influenced because of the competitive nature of a group training session, which could lead to inappropriate load selection and adjustment, and potentially poor technical execution, especially on free-weight barbell exercises.