Reps in Reserve, Rep Ranges and Rest Times: The Evidence-Based Guide to How Hard Every Set Should Be

For most muscle-building sets, stop 1-2 reps short of failure, pick any load you can lift for roughly 5-30 reps, and rest about 2-3 minutes on big lifts (at least 90 seconds on everything else). That is the short answer this reps in reserve RIR guide defends, study by study. Failure is optional, rep ranges are flexible, and rest matters less than the Instagram crowd thinks, but more than “60 seconds, keep the pump” suggests.

Volume tells you how many sets to do. This guide is about the other half: how hard each of those sets should be. Load, reps, proximity to failure, rest and tempo decide whether a set actually counts. (For how many sets per week, see our training volume and frequency guide. For adding weight across weeks, see the progressive overload and deloads guide.)

Here is the framework first, then the evidence behind each piece.

Variable	Practical default for muscle growth	How strong is the evidence?
Proximity to failure	1-2 RIR on most sets, 0 RIR occasionally on safe exercises	Moderate: several meta-analyses plus trials in trained lifters
Rep range	About 5-30 reps; most work in 6-20	Strong for hypertrophy across loads; heavier is better for 1RM strength
Rest between sets	2-3 min on compounds, at least 90 s on isolations	Moderate: one 2024 meta-analysis of 9 studies, plus RCTs
Tempo	Controlled lowering, intentional lift; about 1-4 s per rep	Moderate: two meta-analyses show little effect across normal speeds
How you judge effort	RIR or RPE, calibrated with occasional failure sets	Moderate: accuracy studies in trained lifters

What is reps in reserve (RIR), and how does the RPE scale work?

Reps in reserve is the number of good reps you had left when you racked the weight. Ended a set of 10 knowing you could have squeezed out two more? That was 2 RIR.

The idea got formalised for lifting in 2016, when researchers built a resistance-training RPE scale where each number maps to a rep count (Zourdos et al., 2016). RPE 10 means 0 reps left. RPE 9 means 1 left. RPE 8 means 2 left, and so on. In that study of 29 squatters, slower bar speed lined up strongly with higher RPE ratings in both experienced and novice lifters, which is what you would expect if the numbers reflected real effort. The experienced squatters also rated their true 1RM attempts higher than the novices did, a hint that judging effort is a skill that improves with time under the bar.

A companion paper laid out how coaches can use the scale to adjust loads set to set, so the weight on the bar matches how strong you are that day rather than how strong you were on test day (Helms et al., 2016).

RPE	Reps in reserve	What it feels like
10	0	Could not do another clean rep
9.5	0, maybe a little more weight	No more reps, but slightly heavier was possible
9	1	One more rep, definitely
8	2	Two more reps
7	3	Bar still moving fast-ish, three left
6 or below	4+	Warm-up territory

The RIR number matters because “I did 3 sets of 10” says almost nothing on its own. Ten reps with a weight you could lift 25 times is a warm-up. Ten reps with your 11-rep max is a hard set. If you want to turn RPE numbers into loads, our RIR and RPE calculator does the arithmetic. For a full walkthrough of the scale, see the dedicated RIR and RPE scale guide.

Training to failure vs RIR: what do the meta-analyses say?

This is where gym opinion splits hardest. One camp says a set does not count unless you fail. The other says failure just digs a recovery hole. The research sits in between.

Meta-analysis 1. A 2022 review of 15 studies in young adults found no significant difference between failure and non-failure training for hypertrophy (effect size 0.22, with a confidence interval that crossed zero) or for strength (Grgic et al., 2022). In the subgroup of resistance-trained participants, though, failure showed a small significant edge for growth (effect size 0.15). And when total volume was not matched between groups, non-failure training actually produced better strength gains.

Meta-analysis 2. A 2023 review, also pooling 15 studies, found a “trivial” advantage for set failure over non-failure when any definition of failure was counted (effect size 0.19, confidence interval 0.00 to 0.37) (Refalo et al., 2023). When the analysis was restricted to true momentary muscular failure versus non-failure, that advantage disappeared (effect size 0.12, not significant). The authors suggested the relationship between proximity to failure and growth may be non-linear: getting close matters, but the last grinding rep may add little.

Meta-analysis 3. A 2024 series of meta-regressions took a different approach and estimated the RIR of every training condition in the literature (Robinson et al., 2024). For hypertrophy, the slopes pointed one way: muscle growth increased as sets ended closer to failure. For strength, RIR barely mattered. The authors were careful to call this exploratory, because the RIR values had to be estimated from study descriptions rather than measured.

So do these conflict? Partly. The first two compare categories (failure versus “not failure”, which can mean anything from 1 to 6 reps left), while the third models a continuous slope. A reasonable reading that fits all three: sets that end far from failure probably leave growth on the table, sets that end 1-3 reps short capture most of it, and the final step to 0 RIR adds little or nothing on average.

The trial that tested 1-2 RIR head to head

Meta-analyses mix very different protocols, so a direct test in trained lifters is useful. In a 2024 study, 18 resistance-trained men and women trained one leg to momentary failure and the other leg to 2 RIR on the leg press and 1 RIR on the leg extension, twice a week for 8 weeks (Refalo et al., 2024).

Quadriceps thickness increased by 0.181 cm in the failure leg and 0.182 cm in the RIR leg. Practically identical. Training volume ended up similar too. The difference showed up in fatigue: the failure condition produced consistently greater losses in lifting velocity and reps from set to set.

That is one trial, in one muscle group, over 8 weeks, so it does not settle the question for every exercise or every lifter. It does line up neatly with the meta-analyses. For the full breakdown, read our deep dive on training to failure vs RIR for hypertrophy.

How close to failure should each exercise go?

Not every exercise deserves the same RIR. The cost of a failed rep on a leg extension is a weight stack that clanks down. The cost of a failed rep on a heavy back squat is a lot higher.

A practical way to set it up:

• Heavy barbell compounds (squat, deadlift, bench, overhead press): 1-3 RIR. Technique often breaks before the muscle truly fails, and the fatigue spills over into the rest of the session.
• Machine and cable compounds (leg press, chest press, rows, pulldowns): 1-2 RIR, with an occasional 0 RIR final set.
• Isolation exercises (curls, extensions, lateral raises, leg curls): 0-2 RIR. Low risk, low systemic fatigue, and you learn what failure really feels like.
• Early in a training block or after time off: 2-3 RIR, then creep closer as the weeks go on.

We expand on this split in proximity to failure on compounds vs isolations and on the safety side in failure on machines and cables vs free barbells.

There is also a central versus local fatigue question. Failure sets on big lifts leave you feeling flattened in a way that failure on a cable curl does not. The physiology there is more nuanced than “CNS fatigue”, and we cover it in CNS vs peripheral fatigue and proximity to failure.

How accurate is RIR? Can you really judge it?

Here is the uncomfortable part. Lifters often misjudge how many reps they have left, especially on long sets.

In a 2021 study, 25 trained men squatted 70% of their 1RM to failure (about 16 reps on average) and called out when they thought they were at 5, 3 and 1 RIR (Zourdos et al., 2021). The average error was about 5.2 reps at the “5 RIR” call, 3.7 reps at “3 RIR” and 2.1 reps at “1 RIR” (rounded from 5.15, 3.65 and 2.05). Being further from failure and doing more reps in the set both made predictions worse. Training age did not predict accuracy.

The picture looks better on shorter sets. In a 2024 study, 24 trained men and women benched 75% of 1RM to failure and called 1 and 3 RIR along the way (Refalo et al., 2024). Their average absolute error was just 0.65 reps, with a slight tendency to underpredict (thinking they had fewer reps left than they actually did). Accuracy was not related to sex, years of training or bench strength.

Two studies, two exercises, slightly different methods. They agree on the direction, though: you judge best when you are close to failure and when the set is not marathon-length. That has a practical consequence for rep ranges, which we come to next. To sharpen your own calls, use the drills in calibrating true RIR in the gym and the research summary on how accurately lifters gauge RIR.

What is the best rep range for hypertrophy? The 5-30 rep spectrum

The old textbook “repetition continuum” goes like this: 1-5 reps for strength, 8-12 for size, 15+ for endurance. Tidy. Mostly wrong for muscle growth.

A 2017 meta-analysis pooled 21 studies comparing light loads (60% of 1RM or less) with heavier loads, with every set taken to failure (Schoenfeld et al., 2017). Hypertrophy was similar. 1RM strength gains were clearly better with heavy loads.

A 2021 network meta-analysis went further, with 28 studies and 747 healthy adults, splitting loads into high (8 RM or less), moderate (9-15 RM) and low (over 15 RM), all to failure (Lopez et al., 2021). No significant hypertrophy differences between any of them. Strength was another story: both high and moderate loads beat low loads, with standardised mean differences of roughly 0.6 and 0.35.

Those findings led the same research group to argue for a new model in which muscle can be built, and in some cases maximised, across a wide range of loading zones (Schoenfeld et al., 2021).

So why not do all your training with 30-rep sets? A few practical reasons:

1. Effort is harder to judge. As we just saw, RIR estimates drift badly on long sets.
2. Discomfort becomes the limiter. Burning lungs and lactic misery often end a 30-rep set before the target muscle is truly close to failure.
3. Time. A set of 25 takes far longer than a set of 8.
4. Strength still matters. Heavier work builds 1RM strength that lighter work does not.
5. Joints and technique. Very heavy triples on every exercise beat up some lifters more than moderate loads do (we look at the evidence in rep ranges and joint wear).
6. Low-load sets on big compounds are miserable. A 30-rep squat set is a cardio event.

A sensible split for most natural lifters: 5-10 reps on heavy compounds, 8-15 on machine compounds, 10-20 (occasionally up to 30) on isolations. The detailed version is in our hypertrophy rep spectrum guide, and the “why does light weight work” mechanism (motor unit recruitment as fatigue builds) is explained in low-load vs high-load training and motor unit recruitment.

If maximal strength is your main goal, the rules shift toward heavier loads and lower reps, and our maximal strength and powerlifting guide covers that end of the spectrum.

Optimal rest intervals for lifting: 1 minute vs 3 minutes

Short rest got popular because it “keeps the pump” and supposedly raises hormones. The longer-term muscle data tell a different story.

In a 2016 trial, 21 resistance-trained young men did the same full-body program (3 sessions a week, 3 sets of 8-12 RM on 7 exercises) for 8 weeks, with either 1-minute or 3-minute rests (Schoenfeld et al., 2016). The 3-minute group gained more squat and bench strength and more front-thigh muscle thickness, with a trend toward more triceps growth too.

A 2024 Bayesian meta-analysis pulled together 9 randomised studies on rest and muscle growth (Singer et al., 2024). Results across rest categories overlapped substantially. The authors concluded there is likely a small benefit to resting longer than 60 seconds, perhaps because short rest cuts the load you can lift on later sets, but they did not find appreciable differences once rest exceeded 90 seconds. Whether people trained to failure or stopped short did not meaningfully change that pattern.

So 1 minute versus 3 minutes is not really “short is useless, long is magic”. It is closer to: very short rest makes your next sets worse, and past about 90 seconds you have most of the benefit. On a heavy squat or deadlift, 2-3 minutes (sometimes more) usually lets you repeat your reps. On a cable lateral raise, 90 seconds is often plenty.

There are legitimate uses for short rest: time-limited sessions, supersets of non-competing muscles, or isolation finishers. The trade-offs are covered in short rest intervals: penalties and uses, and the main recommendation in more depth in rest intervals for hypertrophy: 2-3 minutes vs short. If you want shorter sessions without shorter rest, set structures like rest-pause and myo-reps are an option (see rest-pause, myo-reps and cluster sets).

Does rep tempo matter for muscle growth?

Probably less than your gym’s tempo-obsessed coach claims.

A 2015 meta-analysis of 8 studies found similar hypertrophy with repetition durations ranging from 0.5 to 8 seconds, all with sets to failure (Schoenfeld et al., 2015). It also suggested that very slow reps (over 10 seconds) were worse, though the authors noted there were too few controlled studies for firm conclusions.

A 2025 meta-analysis of 14 studies revisited the question (Enes et al., 2025). Comparing faster reps (averaging about 1 second) with slower ones (averaging about 3.5 seconds), the difference in growth was trivial and, if anything, slightly favoured the faster tempo. The authors concluded tempo has minimal overall effect on hypertrophy, with possible differences under specific conditions.

That does not mean tempo is irrelevant. It means the useful version of tempo is simple: lower the weight under control so you are not bouncing out of the bottom, and lift with intent. Our guides on tempo training and repetition duration and intent to move explosively cover the details.

Putting it together: a set-by-set blueprint

Here is how the evidence translates to a normal training day.

1. Warm up with ramp sets at 4+ RIR so the work sets start fresh (see warm-up protocols and ramping sets).
2. First exercise, heavy compound: 5-8 reps, 2-3 RIR on the first set, 1-2 RIR on the last, 2-3 minutes rest.
3. Second exercise, machine or dumbbell compound: 8-12 reps, 1-2 RIR, about 2 minutes rest.
4. Isolation work: 10-20 reps, 0-2 RIR, 90 seconds to 2 minutes rest. Last set to failure is fine here.
5. Tempo: controlled lowering (roughly 1-3 seconds), no bouncing, lift with intent.
6. Log weight, reps and RIR. Next session, add a rep or a little weight at the same RIR.
7. Calibrate with a failure set on a safe exercise every few weeks.

That last progression step is where this guide hands over to the progressive overload guide. How many of these sets you do per muscle per week belongs to the volume and frequency guide. If you want all of this already laid out, our free push, pull and legs plans apply these RIR and rest targets, and the 1RM calculator helps you pick starting loads.

Special cases worth knowing

Cutting. Effort is the signal that tells your body to keep muscle while you eat less. Keep sets close to failure, but fatigue builds faster on fewer calories, so fewer failure sets is a reasonable adjustment. Direct research is thin; our proximity to failure while cutting guide walks through the options.

Beginners. Novices gain from almost any reasonable stimulus, and their RIR calls are rougher. Keep big lifts at 2-3 RIR while technique settles, and learn failure on machines.

Advanced techniques. Lengthened partials past failure and blood flow restriction training are ways to extend sets or make light loads effective. They are tools, not foundations.

Rotating rep ranges. Some programs cycle heavy, moderate and light days within a week or across blocks. Whether that beats a steady mix for hypertrophy is covered in rep range periodisation: DUP vs block.

Health first. If a set to failure brings on chest pain, dizziness, unusual breathlessness or sharp joint pain, stop and get checked by a doctor before training hard again. People with high blood pressure, heart conditions or recent surgery should clear high-effort lifting with their doctor, since straining sets cause big temporary blood pressure spikes.

The bottom line

Most of the arguments about intensity come down to a few research-backed defaults. Stop most sets 1-2 reps short of failure; failure itself is optional and best saved for low-risk exercises. Use whatever rep range you can push hard in, roughly 5-30, and lift heavier if a bigger 1RM matters. Rest long enough to repeat your performance, which for most people means at least 90 seconds and 2-3 minutes on the big lifts. Keep tempo controlled but not slow-motion.

None of this replaces the boring work of adding a rep or a kilo over time, eating enough protein and sleeping. If you are new to the bigger picture, start with our muscle hypertrophy guide for natural lifters, and use the training volume calculator to check that your hard sets add up.