RPE and RIR vs Percentages: Adjusting Training to Fatigue

Introduction: Prescriptive Models in Resistance Training

Prescribing load in resistance training traditionally relies on fixed percentages of a pre-determined one-repetition maximum (%1RM). While percentage-based resistance training (PBRT) provides structured progression, it assumes an athlete's maximal strength remains static between testing periods and does not account for daily fluctuations in readiness, fatigue, or acute recovery status.

To address these limitations, subjective autoregulation models—specifically the Rating of Perceived Exertion (RPE) based on Repetitions in Reserve (RIR)—have been widely adopted. The RIR-based RPE scale directly quantifies the number of additional repetitions an athlete could complete before reaching momentary muscular failure [17]. Subjective RIR scores correlate strongly and inversely with barbell movement velocity in foundational multi-joint exercises, such as the back squat ($r=-0.87$) and bench press ($r=-0.79$), confirming that perceived proximity to failure reflects underlying neuromuscular velocity decrements [2].

Evaluating the efficacy of RIR/RPE-based loading versus fixed percentage-based loading requires examining three key outcomes: maximal strength development, skeletal muscle hypertrophy, and the management of neuromuscular fatigue.

+-----------------------------------------------------------------------------------+
|                             LOAD PRESCRIPTION METHODS                             |
+-----------------------------------------+-----------------------------------------+
|      Percentage-Based Training (%1RM)   |      Autoregulated Training (RPE/RIR)   |
+-----------------------------------------+-----------------------------------------+
| • Fixed load relative to historical 1RM | • Dynamic load adjusted to daily state  |
| • Assumes stable day-to-day capacity    | • Targets specific proximity to failure |
| • Risk of unintended failure on off-days| • Mitigates excess neuromuscular fatigue|
| • Volume load fixed regardless of drift | • Maintains target intensity & velocity |
+-----------------------------------------+-----------------------------------------+

Maximal Strength Adaptations (1RM)

Meta-analytic and experimental evidence indicates that both percentage-based loading and RIR-based autoregulation induce significant improvements in maximal strength, with autoregulated approaches frequently demonstrating small advantages in specific contexts [1, 4, 10].

Meta-Analytic Comparisons

A meta-analysis by Hickmott et al. evaluating load autoregulation (RIR-based RPE and velocity-based training) against standardized percentage-based training identified no statistically significant difference in pooled 1RM strength (MD = 2.07 kg, 95% CI: -0.32 to 4.46 kg, $p=0.09$, SMD = 0.21) [4]. Both objective methods (e.g., velocity zones) and subjective methods (RIR-based RPE) produced comparable muscular strength adaptations relative to fixed percentage assignments [4]. Similarly, a meta-analysis by Liao et al. comparing velocity-based training (VBT) to percentage-based training found no significant differences in back squat 1RM (MD = 3.03 kg, 95% CI: -3.55 to 9.61 kg) or velocity at 60% 1RM [6].

Conversely, an earlier meta-analysis evaluating 8 studies in trained lifters reported that autoregulated resistance training significantly outperformed fixed loading for maximal strength overall (ES = 0.64, $p<0.001$), displaying distinct subgroup benefits in the back squat (ES = 4.64) and bench press (ES = 3.21) [10]. In ranking models of progression, a network meta-analysis evaluating maximal strength gains found autoregulated models to rank above PBRT, with the hierarchy: Autoregulatory Progressive Resistance Exercise (APRE) > Velocity-Based Resistance Training (VBRT) > RPE > PBRT [5]. APRE was identified as having the highest effect size among autoregulation strategies (ES = 0.78), with interventions $\le 8$ weeks yielding greater relative strength adaptations (ES = 0.87) than longer interventions of 8–10 weeks (ES = 0.32) [10].

Progression Model Hierarchy for 1RM Strength (SUCRA / Network Meta-Analysis [5]):

  [ 1. APRE ]  >  [ 2. VBRT ]  >  [ 3. RPE / RIR ]  >  [ 4. PBRT (%1RM) ]
  (Highest 1RM)                                        (Fixed Baseline)

Experimental Trials in Trained Cohorts

Direct experimental comparisons in resistance-trained lifters mirror these findings. In an 8-week daily undulating resistance training study by Helms et al., both percentage-based loading ($n=11$) and RIR-based RPE loading ($n=10$) yielded significant strength increases in the bench press (+9.64 kg vs. +10.70 kg), squat (+13.91 kg vs. +17.05 kg), and combined total (+23.55 kg vs. +27.75 kg) [1]. While conventional null-hypothesis testing showed no statistically significant between-group differences, magnitude-based inferences indicated a 79%, 57%, and 72% probability of a small effect size advantage for RPE loading in the squat (ES = 0.50), bench press (ES = 0.28), and combined total (ES = 0.48), respectively [1, 2].

In a randomized crossover trial by Barnes et al., autoregulated periodization (AR) and traditional percentage periodization (TP) produced comparable increases in squat ($p<0.001$), deadlift ($p=0.001$), and 3-lift total ($p<0.001$) [9]. AR facilitated a higher average training intensity (%1RM) during squat and deadlift sessions; however, bench press 1RM demonstrated a significant program $\times$ time interaction favoring traditional percentage-based loading ($p=0.014$, $d=1.11$) [9].

Across high-load paradigms, loads exceeding 60% 1RM remain superior for maximizing 1RM adaptations compared to low-load paradigms ($\le 60\%$ 1RM; ES difference = 0.58), an effect that persists independent of total volume-load [16].

Hypertrophic Adaptations

Skeletal muscle hypertrophy demonstrates near-complete equivalence between RPE/RIR prescription and percentage-based loading, provided sets are performed with comparable proximity to failure and matched volume [1, 4].

+-----------------------------------------------------------------------------------+
|              8-WEEK HYPERTROPHY: %1RM VS. RPE LOADING (HELMS ET AL. [1])          |
+----------------------------+-----------------------+------------------------------+
| Muscle Site                | %1RM Group (+Δ mm)    | RPE/RIR Group (+Δ mm)        |
+----------------------------+-----------------------+------------------------------+
| Pectoralis Major           | 1.59 ± 1.33           | 1.90 ± 1.91                  |
| Vastus Lateralis (50% FL)  | 2.13 ± 1.95           | 1.85 ± 1.97                  |
| Vastus Lateralis (70% FL)  | 2.40 ± 2.22           | 2.31 ± 2.27                  |
+----------------------------+-----------------------+------------------------------+
*Note: Differences between groups were not statistically significant (p > 0.05) [1, 2].

In the trial by Helms et al., ultrasound measurements of muscle thickness revealed no significant differences ($p>0.05$) between percentage-based and RPE-based groups across the pectoralis major (+1.59 mm vs. +1.90 mm), vastus lateralis at 50% femur length (+2.13 mm vs. +1.85 mm), or vastus lateralis at 70% femur length (+2.40 mm vs. +2.31 mm) [1, 2].

Similarly, when evaluating proximity to failure directly, an 8-week study on trained lifters found that momentary muscular failure and a prescribed 1- to 2-RIR produced nearly identical increases in average quadriceps thickness (0.181 cm vs. 0.182 cm) under matched volume conditions [21]. Longitudinal data over 10 weeks comparing a periodized proximity-to-failure model (varying between 1 and 4 RIR) to a constant 1-RIR protocol found identical vastus lateralis cross-sectional area gains (6.5% vs. 5.5%), squat 1RM gains (9.8% vs. 9.2%), and bench press gains (7.0% vs. 9.6%), while the periodized RIR group elicited greater triceps brachii hypertrophy (5.8% vs. 2.2%) alongside lower average session RPE [18].

Meta-analytic evaluations confirm that training to momentary muscular failure is not mandatory for maximizing muscle cross-sectional area (CSA), provided relative intensity and volume are equated [4]. When volume is autoregulated using velocity loss thresholds, higher velocity loss thresholds ($>25\%$) induce significantly greater CSA hypertrophy compared to conservative thresholds ($\le 20\%$; MD = 0.64 cm², $p=0.03$; and $\le 25\%$; MD = 0.61 cm², $p=0.03$), whereas no significant difference exists between $>25\%$ and 20–25% thresholds (MD = 0.36 cm², $p=0.28$) [4].

Prescriptive Fidelity, Fatigue, and RIR Accuracy

A critical distinction between prescription models lies in how they manage acute fatigue and maintain prescriptive accuracy under varying physiological states [12, 20].

Accuracy and Load Drift Under Fatigue

When lifters perform resistance training following fatiguing bouts, fixed percentage prescriptions lose fidelity [12]. In a crossover study by Cowley et al. comparing %1RM, RPE, RIR, and VBT at 70% 1RM, VBT demonstrated the highest accuracy, maintaining 100% of sets within 5% of target velocity [12]. In contrast, %1RM and subjective RPE resulted in substantial prescriptive drift; fixed %1RM frequently forced sets to momentary failure as fatigue mounted, whereas RIR and VBT allowed athletes to maintain targeted training volumes without unintentional failure [12].

+-----------------------------------------------------------------------------------+
|              ACUTE FATIGUE MARKERS BY PROXIMITY TO FAILURE (BENCH PRESS [20])     |
+--------------------------+---------------------+----------------------------------+
| Condition                | 4-min Post-Loss (%) | 24-h Neuromuscular Deficit (%)   |
+--------------------------+---------------------+----------------------------------+
| Momentary Failure (FAIL) | -25%                | -3%                              |
| 1-RIR                    | -13%                | -3%                              |
| 3-RIR                    | -8%                 | +2% (recovered)                  |
+--------------------------+---------------------+----------------------------------+

Neuromuscular Fatigue Dynamics

True muscular failure imposes a disproportionate fatigue cost compared to non-failure repetitions [17, 20]. Following 6 sets of bench press at 75% 1RM, acute velocity loss measured 4 minutes post-exercise was -25% for sets taken to failure, compared to -13% for 1-RIR and -8% for 3-RIR [20]. Furthermore, set-to-set velocity loss from the first to the final set was significantly higher for failure sets (-22%) than for 1-RIR (-9%) or 3-RIR (-6%) [20]. Although 24-hour neuromuscular decrements were present in both failure (-3%) and 1-RIR (-3%) conditions compared to 3-RIR (+2%), all fatigue markers fully resolved by 48 hours [20]. Fixed velocity loss thresholds do not completely eliminate failure; performing squats to a 40% velocity loss threshold causes lifters to reach true failure in approximately 56% of sets [20].

For athletic performance metrics, network meta-analyses show that low velocity loss thresholds ($\le 20\%$ loss) yield higher P-scores ($\ge 0.59$) for strength, jump, and sprint performance compared to high velocity loss thresholds ($>20\%$, P-score $\ge 0.50$), traditional percentage-based training (P-score $\le 0.46$), and RIR alone (P-score $\le 0.46$) [8]. In volume autoregulation, velocity loss thresholds $\le 25\%$ produce significantly greater 1RM strength gains than thresholds $>25\%$ (MD = 2.32 kg, $p=0.02$) when supplementary exercises are incorporated [4].

Accuracy of Subjective RIR Estimation

A common concern with RPE/RIR prescription is subjective error. However, resistance-trained individuals accurately predict proximity to failure within approximately 1 repetition when working within 0 to 3 RIR [1, 21]. Experimental trials show prediction errors of $0.40\pm 0.68$ repetitions at a 1-RIR target and $0.90\pm 0.81$ repetitions at a 3-RIR target, with accuracy improving over baseline across longitudinal training blocks [18, 21].

Practical Application and Programming Guidelines

+-----------------------------------------------------------------------------------+
|                     APPLICATION MATRIX: %1RM VS. RIR/RPE                          |
+-------------------+--------------------------------+------------------------------+
| Objective         | Recommended Prescriptive Model | Target Proximity / Threshold |
+-------------------+--------------------------------+------------------------------+
| Maximal Strength  | Autoregulated (RIR/RPE / APRE) | 1–3 RIR (≥60% 1RM) [10, 16]  |
| Hypertrophy       | RIR or %1RM (Matched Volume)   | 1–2 RIR / >25% VL [4, 21]    |
| Power / Sprinting | Low-Velocity Loss VBT          | ≤20% Velocity Loss [8]       |
| Fatigue Control   | Periodized RIR                 | 1–4 RIR (Avoid Failure) [18] |
+-------------------+--------------------------------+------------------------------+

1. For 1RM Strength Development: Autoregulation through RIR, APRE, or VBT provides slight advantages over rigid percentage-based systems by capitalizing on days of high physical readiness and preventing premature failure on fatigued days [1, 5, 10]. Loads should remain $>60\%$ 1RM to optimize specific neural and mechanical adaptations [16].
2. For Skeletal Muscle Hypertrophy: Both RIR-based and percentage-based loading are equally viable provided sets terminate within 1 to 3 repetitions from failure (1–3 RIR) [1, 21]. Terminating sets at 1–2 RIR avoids the disproportionate neuromuscular fatigue associated with complete failure while providing an equivalent hypertrophic stimulus [20, 21].
3. Prescriptive Fidelity: In environments where training occurs under acute residual fatigue, RIR and VBT should be prioritized over static %1RM to prevent unintended failure and excessive intrasession velocity loss [12, 20].