Estimating Repetitions in Reserve During the Bench Press Exercise: Should We Consider Sex and the Exercise Equipment?

Abstract

Background:

Estimating repetitions in reserve (RIR) has been established as a valid and practical method for monitoring set configuration concerning proximity to failure in resistance training. However, factors such as sex and exercise equipment could interfere with the estimation process. Therefore, this study aimed to determine the influence of sex (female vs male) and exercise equipment (free-weight vs Smith machine) on the RIR estimation at different relative loads during the bench press exercise.

Hypothesis:

Men would be more accurate than women in the RIR estimation, and participants would better perceive their actual effort at the Smith machine compared with the free-weight bench press exercise.

Study Design:

Repeated measures design.

Level of Evidence:

Level 3.

Methods:

In a counterbalanced order, recreationally trained participants (12 female, 14 male) performed 2 identical sessions with the free-weight bench press exercise in 1 week and 2 identical sessions with the Smith machine bench press exercise in another week (first session for familiarization and second session for validity analysis). In each session, participants performed single sets of repetitions to failure against 3 relative loads (65%, 75%, and 85% of the 1-repetition maximum [1RM]).

Results:

Results revealed a significant interaction between RIR and sex at 65% 1RM (P < 0.01; women: -1.2 [-1.7 to -0.6] repetitions; men: -0.1 [-0.6 to 0.3] repetitions), and between RIR and exercise equipment at 65% 1RM (P < 0.01; free-weight: -1.1 [-1.5 to -0.6] repetitions; Smith machine: -0.2 [-0.6 to 0.2] repetitions), but not at 75% and 85% 1RM.

Conclusions:

RIR estimation is a valid method for monitoring set configuration. However, this method tends to underestimate RIR at 65%1RM with female athletes and in the free-weight exercise for both sexes.

Clinical Relevance:

Strength and conditioning coaches should help their athletes understand effort awareness in the aforementioned context.

Precise application of various training principles, including individualization, overload, and progression, relies on the manipulation of resistance training variables such as intensity and volume. ¹² Consequently, strength and conditioning coaches require reliable and valid methods for their daily implementation in program development. ¹ In that sense, prescribing intensity as a relative percentage of the 1-repetition maximum (1RM) through velocity-based training (VBT) is a well-established practice. ²⁵ However, there is ongoing debate about the most appropriate method to prescribe volume in terms of proximity to failure. ¹⁹ For example, prescribing a fixed number of repetitions per %1RM (ie, the percentage-based method) does not ensure the same level of effort across athletes. ² Although this between individual variability could be controlled through a maximum repetition test (ie, the repetition maximum method), in practice, such maximal capacities are not maintained through the successive sets of a training session. ²³ Similar limitations are present when implementing the VBT method by measuring the lifting velocity loss up to a threshold in the set, ²¹ or by determining the repetition velocity associated with each repetition in reserve (ie, the number of repetitions remaining before reaching muscular failure; RIR) at which the set would be later stopped. ¹⁴ Thus, the subjective determination of RIR has gained interest in both scientific and training fields. ⁷ However, certain factors that could influence its accuracy still need exploration.

During a training set performed to failure, athletes experience periods of uncertainty and certainty according to the feedback received after each repetition completed. ²⁴ Through specific categorical scales designed to evaluate effort in resistance exercises,^6,26 participants can accurately interpret such information and provide precise estimations of RIR. ⁷ However, due to anatomic-physiological differences between sexes, ¹⁰ male athletes may provide more accurate RIR estimations.^4,5 This could be linked to different signaling of the muscle afferents (ie, muscle spindles, Golgi tendon organs, group III and IV) that inform about the emerging fatigue during an exhaustive task. ³ In this regard, it has been reported that female athletes are particularly affected by fatigue in proprioceptive afferents. ⁸ In contrast, female athletes present a greater proportional area of type I muscle fibers and higher perfusion, resulting in a lower metabolic disruption and discharge of the mechanical and metabolic sensitive group III and IV afferents compared with male athletes at the same relative load. ²² Despite clear anatomic-physiological differences, sex-specific information is still limited and existing literature presents conflicting results. Odgers et al ¹⁸ reported that both trained male and female athletes accurately reported RIR-4 and RIR-1 at 80% 1RM during front squat and hexagonal deadlift exercises. On the contrary, Hacket et al^4,5 found that recreationally trained men were more accurate in estimating different RIRs at 70% and 80% 1RM in chest and leg press machines. Therefore, further study is needed to draw stronger conclusions on the topic.

Given this paradigm about fatigue and afferent mechanisms in RIR estimation, another question emerges about the potential effects of the exercise equipment (ie, free-weight vs machine-based). Stabilizing the load requires superior muscle work. ⁹ In this regard, McCaw et al ¹⁶ reported higher muscle activity in both the stabilizer (ie, medial deltoid) and primary mover muscle (ie, anterior deltoid) during the free-weight bench press compared with the Smith machine. Accordingly, this different pattern of muscle activity could interfere with RIR estimation. However, this pattern seems to be present only at low relative loads, as at higher loads, increased overall muscle activity enhances joint stiffness, minimizing the need for stabilization. ¹⁶ Likewise, the muscle afferents involved in the proprioceptive sensation of the barbell velocity loss seem to be also impaired at low relative load, ¹¹ which reinforces the need to explore its interaction with the RIR estimation.

Based on the existing knowledge gaps, this study aimed to determine the influence of sex (female vs male) and the exercise equipment (free-weight vs Smith machine) on the validity of the RIR estimation at different relative loads (65%, 75% 85% 1RM) during the bench press exercise. It is hypothesized that men would be more accurate than women in the RIR estimation,^4,5 and that participants would better perceive their actual effort at the Smith machine compared with the free-weight bench press exercise. ¹⁶

Methods

Experimental Design

A repeated measures design was used in the present study. Participants completed 4 sessions during 2 consecutive weeks with ≥72 hours between them. In a counterbalanced order, they performed 2 identical sessions with the free-weight bench press exercise in one week and 2 identical sessions with the Smith machine bench press exercise in another week. The first session was used to familiarize participants with the testing procedure, while data from the second session were used for validity analyses. Each session started determining the free-weight or Smith machine bench press 1RM through the individualized load-velocity relationship. Then, participants performed single sets of repetitions to failure against 3 relative loads (65% 1RM, 75% 1RM, and 85% 1RM) in the free-weight or Smith machine bench press exercise. This was determined in a randomized counterbalanced order. During each set, participants verbally described their estimated RIR-2, which was compared with the maximum number of repetitions performed to failure. All sessions were separated by ≥72 hours of rest. Each participant was tested at the same time of the day (±1 hour) and under similar environmental conditions (22°C and 60% humidity, approximately).

Participants

A total of 26 recreationally trained participants, 12 female (age, 26.7 ± 2.6 years [21-30]; body mass, 60.9 ± 10.8 kg; height, 164 ± 6.9 cm; free-weight bench press 1RM, 48.8 ± 5.1 kg [0.81 ± 0.09 relative to body mass]; Smith machine bench press 1RM, 50.6 ± 6.4 kg [0.84 ± 0.12 relative to body mass]) and 14 male (age, 24.1 ± 3.5 years [20-31]; body mass, 77.2 ± 6.6 kg; height, 175 ± 5.1 cm; free-weight bench press 1RM, 91.3 ± 13.3 kg [1.18 ± 0.13 relative to body mass]; Smith machine bench press 1RM, 94.1 ± 13.9 kg [1.22 ± 0.13 relative to body mass]), participated voluntarily in this study. A sample size of 24 participants was deemed adequate based on the expected RIR bias estimation reported in previous studies. ⁷ The required sample size was conducted in the G*Power (Version 3.1) given α = 0.05, effect size (ES) = 0.60 and 1 - β = 0.80. All participants were physically active and accustomed to performing the bench press exercises as part of their regular training. Using a self-administered questionnaire, no physical limitations, health problems, or musculoskeletal injuries that could impact testing were identified. All participants were informed of the procedures and signed a written informed consent form before initiating the study. The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board.

Testing Sessions

Height and body mass were measured at the beginning of the first familiarization session using a wall-mounted stadiometer (Seca 202, Seca Ltd) and a digital scale (Seca 899, Seca Ltd), respectively. Thereafter, participants were provided with details about how to use the estimated repetitions to failure scale, and to identify the RIR-2 through each set. ⁶ The warm-up consisted of jogging, dynamic stretching, and joint mobility exercises. Subsequently, participants completed 1 set of 6, 4, and 2 repetitions of the free-weight or Smith machine bench press exercise with the 40%, 60%, and 80% 1RM, respectively. Interset rest was fixed at 5 minutes. The individualized load-velocity relationship was then used to estimate the free-weight or Smith machine bench press 1RM as the load associated with a mean velocity of 0.17 m/s.^13,20,25 A linear velocity transducer (T-Force system; Ergotech) was used to collect the mean velocity of all repetitions. Participants then completed single sets of repetitions to momentary muscular failure (ie, the participant was unable to complete the concentric portion of a given repetition with a full range of motion) at maximal intended velocity using 3 relative loads (65%, 75%, and 85% 1RM).^19,23 The 3 relative loads were applied in a randomized order between participants in the first session, but the sequence was maintained in the subsequent sessions. A rest period of 10 minutes was implemented between successive sets. The bench press exercise was performed with a free-weight barbell (FreeWeight, Etenon Fitness S.L.) or on a Smith machine (Multipower Fitness Line, Peroga) using the standard 5-point body contact position technique (head, upper back, and buttocks placed firmly on the bench with both feet flat on the floor) and a self-selected grip.

Statistical Analysis

Descriptive data are presented as means and standard deviations. The normal distribution and homogeneity of variance were confirmed through the Shapiro-Wilk test and Levene’s tests, respectively. A mixed model analysis of variance (ANOVA) with the RIR (actual vs estimated) and exercise equipment (free-weight vs Smith machine) as within-participants factors and sex as between-participants factor was performed for each relative load (ie, 65% 75%, 85% 1RM). Pairwise comparisons were identified using Bonferroni post hoc corrections. An acceptable RIR estimation was considered individually when individual raw errors were ≤1 repetition. ²² Statistical analyses were performed using the software SPSS (IBM, Version 25.0). Statistical significance was set at an alpha level of 0.05.

Results

Sixty-Five Percent 1RM

There was a significant sex effect on the number of repetitions completed (F_(1,25) = 5.5; P = 0.03; women vs men, 2.4 [0.3-4.5] repetitions) (Table 1). No significant exercise equipment effect was observed on the number of repetitions completed (F_(1,25) = 0.9; P = 0.35).

Table 1.

Mixed model ANOVA comparing the actual and estimated number of repetitions at RIR-2 for male and female participants at free-weight and Smith bench press exercises and different relative loads

Load (% 1RM)	Exercise Mode	Sex	RIR		Individual Sets					ANOVA
			Actual	Estimated	✘↑	↑	✓	↓	✘↓	Main Effects	Interactions
65%	Free-weight	Female	18.4 ± 4.8	16.7 ± 4.6	0	0	3	3	6	RIR: F(1,25) = 13.5;P < 0.01 Sex: F(1,25) = 5.5;P = 0.03 Mode: F(1,25) = 0.9;P = 0.35	RIR × Sex: F(1,25) = 8.5;P < 0.01 RIR × Mode: F(1,25) = 16.4; P < 0.01 RIR × Mode × Sex: F(1,25) = 2.3; P = 0.14
		Male	15.0 ± 1.3	14.6 ± 1.4	0	0	9	3	3
	Smith machine	Female	16.9 ± 3.4	16.3 ± 3.3	0	2	3	5	2
		Male	14.5 ± 2.8	14.6 ± 2.5	1	4	7	2	1
75%	Free-weight	Female	13.2 ± 5.3	12.1 ± 4.2	0	2	5	4	1	RIR: F(1,25) = 5.0;P = 0.03 Sex: F(1,25) = 2.0;P = 0.16 Mode: F(1,25) = 4.3;P = 0.05	RIR × Sex: F(1,25) = 2.9;P = 0.10 RIR × Mode: F(1,25) = 0.6; P = 0.45 RIR × Mode × Sex: F(1,25) = 0.2; P = 0.69
		Male	10.7 ± 1.6	10.5 ± 1.6	0	3	7	1	4
	Smith machine	Female	11.6 ± 4.4	10.8 ± 3.6	1	1	2	5	3
		Male	10.1 ± 1.9	10.1 ± 1.9	0	2	9	1	3
85%	Free-weight	Female	6.6 ± 4.9	6.7 ± 3.7	1	4	2	3	2	RIR: F(1,25) = 1.7;P = 0.20 Sex: F(1,25) = 3.0;P = 0.10 Mode: F(1,25) = 0.2;P = 0.66	RIR × Sex: F(1,25) = 0.7;P = 0.42 RIR × Mode: F(1,25) = 0.1; P = 0.71 RIR × Mode × Sex: F(1,25) = 0.1; P = 0.71
		Male	4.6 ± 1.6	5.1 ± 1.2	0	8	3	3	1
	Smith machine	Female	6.5 ± 4.1	6.6 ± 3.6	2	2	4	3	0
		Male	4.5 ± 1.8	4.7 ± 1.6	0	5	7	3	0

Data are presented as means ± SD. 1RM, 1-repetition maximum; ANOVA, analyses of variance; ✘↑/↓, Subjects overestimated or underestimated the RIR-2 by >1 repetition; ↑/↓, Subjects overestimated or underestimated the RIR-2 by 1 repetition. ✓, Subjects correctly estimated the RIR-2; RIR, repetitions in reserve.

There was a significant RIR effect (F_(1,25) = 13.5; P < 0.01; estimated vs actual: -0.6 [-1.0 to -0.3] repetitions). A significant interaction between RIR and sex was observed (F_(1,25) = 8.5; P < 0.01; women: -1.2 [-1.7 to -0.6] repetitions; men: -0.1 [-0.6 to 0.3] repetitions), as well as between RIR and exercise equipment (F_(1,25) = 16.4; P < 0.01; free-weight: -1.1 [-1.5 to -0.6] repetitions; Smith machine: -0.2 [-0.6 to 0.2] repetitions). No significant interaction between RIR, exercise equipment, and sex was observed (F_(1,25) = 2.3; P = 0.14). Specifically, in the free-weight bench press exercise, female and male participants provided valid RIR estimates on 50% and 80% of occasions, respectively. In contrast, for the Smith machine bench press, the valid estimates were higher, with female and male participants achieving 83.3% and 86.6%, respectively (Figure 1).

Figure 1.

Individual and group values for the raw errors between the actual and estimated number of RIR for male and female participants at different exercise modes and relative loads. Data are presented as means ± 95% CI. Dashed lines represent the acceptable threshold established. 1RM, 1-repetition maximum; RIR, repetitions in reserve.

Seventy-Five Percent 1RM

There was no significant sex effect on the number of repetitions completed (F_(1,25) = 2.0; P = 0.16) (Table 1). No significant exercise equipment effect was observed on the number of repetitions completed (F_(1,25) = 4.3; P = 0.05).

There was a significant RIR effect (F_(1,25) = 5.0; P = 0.03; estimated vs actual: -0.5 [-1.0 to -0.0] repetitions). No significant interaction between RIR and sex was observed (F_(1,25) = 2.9; P = 0.10). Additionally, no significant interaction was observed between RIR and exercise equipment (F_(1,25) = 0.6; P = 0.45), as well as between RIR, exercise equipment, and sex (F_(1,25) = 2.3; P = 0.14). Specifically, in the free-weight bench press exercise, female and male participants provided valid RIR estimates on 80% and 75.3%, respectively, Meanwhile, in the Smith machine bench press, the valid estimates were 66.6% for female and 80% for male athletes (Figure 1).

Eighty-Five Percent 1RM

There was no significant sex effect on the number of repetitions completed (F_(1,25) = 3.0; P = 0.10) (Table 1). No significant exercise equipment effect was observed on the number of repetitions completed (F_(1,25) = 0.2; P = 0.66).

There was no significant RIR effect (F_(1,25) = 1.7; P = 0.20). No significant interaction between RIR and sex was observed (F_(1,25) = 0.7; P = 0.42). In addition, no significant interaction was observed between RIR and exercise equipment (F_(1,25) = 0.1; P = 0.71), as well as between RIR, exercise equipment, and sex (F_(1,25) = 0.1; P = 0.71). Specifically, in the free-weight exercise, female and male participants provided valid RIR estimates on 75% and 93.3%, respectively. Meanwhile, in the Smith machine bench press exercise, the valid estimates were 75% for female and 100% for male participants (Figure 1).

Discussion

This study aimed to determine the influence of sex and exercise equipment on the RIR estimation at different relative loads during the bench press exercise. The current findings revealed that female athletes tend to underestimate the RIR at 65% 1RM, as do both sexes in free-weight exercise. No significant interactions were observed at superior loads.

First, the validity of the RIR estimation method requires analysis in a practical context. In the present study, the acceptable bias established (≤1 repetition for RIR-2) well supports the results reported by Refalo et al. ²² It was observed that for 6 sets of free-weight bench press at 75% 1RM ending at failure, RIR-1, and RIR-3, participants experienced significantly different degrees of fatigue, illustrated by the change in lifting velocity from pre- to postexercise (ie, -25, -13, and -8%, respectively) at a specific load (ie, 85% of 75% 1RM). These effects were diminished to -3%, -3%, and +2% at 24 hours postexercise and became negligible at 48 hours postexercise. Taking this bias into account, our results revealed that the risk of reaching muscle failure unintentionally was inappreciable. However, a high rate of significant underestimates (ie, >RIR-3) was observed at 65% 1RM, especially in female compared with male participants, which could impact the expected allostatic load of the training session. Therefore, it is of paramount importance to explore this limitation of the RIR estimation method at low relative loads.

During a set performed to failure, there is a progressive recruitment of superior motor units, reaching a similar threshold between relative loads for a given RIR. ¹⁷ Accordingly, it could be expected that a similar motor command is being integrated by the sensory areas of the cerebral cortex, resulting in a similar RIR perception. ¹⁵ However, at low relative loads, the sense of discomfort is accentuated and erroneously summed to the sense of effort, potentially causing the observed underestimation. ⁷ It was observed that sex played a crucial role in the notable underestimation observed at the lowest relative load (65% 1RM), especially evident in female athletes, despite their well-documented higher resistance to fatigue.^10,22 It was initially hypothesized that, due to their greater proportional area of type I muscle fibers and higher perfusion, female athletes would present a lower motor command and metabolic perturbation, resulting in a more favorable context to perceive their actual effort. Notwithstanding these claims, reflected in higher repetitions performed to failure, a greater perception of effort was observed in female participants, which could be justified by their superior sensitivity to fatigue in their muscle afferents. ⁸ Han et al ⁸ discussed that fatigue impairs signal transmission speed, resulting in a decoupling between the sensory information and interpretation. Although these differences between sexes have been observed in previous studies,^4,5 others have reported accurate RIR estimations for both sexes. ¹⁸ Thus, it is essential to explore further settings such as participant level, muscle groups involved, and different proximities to failure.

With regards to exercise equipment, the disparities observed at the lowest relative load (65% 1RM) could be justified by the superior muscle work required at the free-weight bench press exercise. ⁹ As aforementioned, McCaw et al ¹⁶ observed that the stabilizer muscle (ie, medial deltoid) displays a superior muscle activity in the free-weight compared with the Smith machine bench press exercise at low relative loads. Whereas the primary movers may trigger a similar motor command, the increased demand for stabilization may contribute to a heightened sense of discomfort, thus influencing an overall perception of effort. Likewise, the nonsignificant interactions at superior loads may have been related to the increased joint stiffness as a result of the overall greater muscle activity. ¹⁶

Limitations

Some limitations to this study should be recognized. First, the influence of sex and the exercise equipment on the validity of the RIR estimation was tested using bench press exercises. Other exercises, such as the back squat, deadlift, or weightlifting derivatives will require further investigation. Second, the recreationally-trained participants confirmed that they were familiarized with the different stages of the RIR scale after sets performed to failure during the familiarization session. Therefore, practitioners should be aware of the experience of their athletes with the scale and with reaching failure, and, if necessary, dedicate a specific session to this aim.

Conclusion

The RIR method results in a valid and practical tool for controlling proximity to failure during the bench press exercise. Importantly, this method requires no equipment, facilitating its implementation into daily practice, particularly when working with multiple athletes simultaneously. However, strength and conditioning coaches should consider that (1) athletes tend to underestimate their actual RIR at low relative loads during the bench press exercise, (2) female athletes display a greater underestimation than male athletes at low relative loads, and (3) both sexes display a greater underestimation in the free-weight compared with the Smith machine bench press at low relative loads. Therefore, the focus should be on accurately gauging the number of possible remaining repetitions, disregarding any sense of discomfort for this purpose, especially in the contexts mentioned above.