How long to rest between sets

The best single test of the idea is Longo et al. (2022). Twenty-eight people trained one leg at a time on the knee extension at 80% of 1RM for ten weeks. Each leg followed one of four plans: 3 minutes or 1 minute of rest, with or without its volume load matched to the other. Quadriceps cross-sectional area grew 13.1% on 3-minute rest and 12.9% on 1-minute rest when the short-rest leg was given extra work to match the volume load. It grew only 6.8% and 6.6% in the two legs where volume load fell. Rest length did not decide the result. Volume load did.

Singer et al. (2024) pooled nine randomised trials and 19 measurements. Longer rest (over 60 seconds) against 60 seconds or less gave a standardised mean difference of 0.13 for the arms (95% credible interval −0.27 to 0.51) and 0.17 for the thigh (−0.13 to 0.43). The whole-body measure leaned the other way at −0.08 (−0.45 to 0.29). The full text puts the chance that longer rest is better at 74% for the arms and 88% for the thigh. That is a lean, not a verdict. The authors concluded there is a small benefit from rest over 60 seconds, perhaps through less lost volume load, and that they found no appreciable difference beyond 90 seconds.

Do not read 90 seconds as a ceiling. In the same paper's network analysis of the thigh, the very long category (180 seconds or more) had the largest estimate against short rest (0.32), ahead of 61 to 119 seconds (0.13) and 120 to 179 seconds (0.01). All three had wide credible intervals that included zero.

The trained-lifter case is thin. Six of the nine trials used untrained people and three used resistance-trained people, and the authors say there was insufficient data to analyse trained lifters separately. The one trial in trained men that gets quoted most, Schoenfeld et al. (2016), found larger 1RM squat and bench press gains and a bigger gain in front-of-thigh muscle thickness after eight weeks of 3-minute rest than 1-minute rest, in 21 young men. A 2025 preprint on trained males alone, not yet peer reviewed, found only a trivial hypertrophy difference (SMD 0.08) and reported longer rest improving strength. For trained lifters who want muscle, then, the question is open.

Longer rest earns its keep for strength. Grgic et al. (2018) reviewed 23 studies and 491 participants. They concluded that resistance-trained people need rest of more than two minutes to maximise strength gains, while 60 to 120 seconds is enough for untrained people.

Two limits apply. The same review says robust strength gains still happen with rest under 60 seconds, so the gap is one of degree and a rushed session is not wasted. And of its 491 participants, 413 were male, so the two-minute rule rests on a male-dominated base.

The 2026 ACSM position stand (Currier et al.) replaced the 2009 one in April 2026. It pooled 137 systematic reviews, and its summary of the strength findings says strength was not affected by short (under 1 minute) against long (over 1 minute) rest between sets. For hypertrophy it says there were insufficient data to judge inter-set rest. That coarse split does not contradict a 2-minute against 3-minute difference in trained lifters. It does mean no major guideline now hands you a rest number to follow.

This table is a reading of the sources above, not a prescription from one trial.

The most repeated finding is also the plainest. Short rest costs reps later in the session. Willardson and Burkett (2006) had 15 men with training experience do five sets at a 15RM load. Reps fell from set one to set five at every rest length. Two minutes beat 30 seconds in the squat, and beat both 30 seconds and one minute in the bench press. Senna et al. (2011) saw the same pattern over five sets at 10RM in 15 resistance-trained men.

Here is what that does to a session. These numbers are invented to show the arithmetic, not taken from a study. Take 5 sets of 5 at 100 kg. With enough rest you get 5, 5, 5, 5, 5: 25 reps and 2,500 kg of volume load. With too little you might get 5, 5, 4, 4, 3: 21 reps and 2,100 kg. Same plan, 16% less work.

One exercise, one population and ten weeks make Longo's trial a hint, not a law. But it explains why rest is a means of protecting the hard sets you planned, not a growth input of its own. Our methodology page counts completed hard sets for the same reason.

The old advice was 30 to 60 seconds for muscle and 3 to 5 minutes for strength. De Salles et al. (2009) recommended short rest for hypertrophy on growth-hormone grounds. The case for that has weakened. Schoenfeld (2013) reviewed the hormone hypothesis and noted that some researchers think the post-exercise spikes mobilise fuel, not build tissue. Henselmans and Schoenfeld (2014) confirmed that rest under a minute does raise acute growth hormone and lowers the testosterone-to-cortisol ratio. They called the link from those changes to long-term growth highly contentious and apparently weak. Of the longitudinal trials they reviewed, none found more growth with shorter rest and one found the opposite.

So the rest data show no trade-off between training for size and training for strength. Longer rest is at worst neutral for growth and is the better bet for strength. Singer et al. also say the 30 to 90 second guidance for hypertrophy deserves a second look.

The convention says yes. ACSM's 2009 position stand, now replaced, recommended 3 to 5 minutes between sets for heavy loading in the 1 to 6RM range, 1 to 2 minutes in the 6 to 12RM zone for hypertrophy, and under 90 seconds for muscular endurance.

The best head-to-head does not support a clean split. In Senna et al. (2011), the chest fly, leg extension and leg press all kept improving in reps from one to three to five minutes. Only the bench press plateaued after three. The divide there is one lift against three, not compound against isolation. On that evidence, single-joint work is not the work that needs less rest.

A reasonable rule, which we label as convention and not as a finding: rest longer on heavy compounds, because a lost rep costs the most there. Do not assume curls recover faster. They may not.

Pair exercises instead of rushing them. In a narrative review, Iversen et al. (2021) report that supersets, drop sets and rest-pause training roughly halve training time against traditional sets while keeping volume. The most recent meta-analysis, Zhang et al. (2025, 19 studies, 313 participants), found supersets gave a similar number of total reps (SMD −0.03) and volume load (SMD 0.05) in a shorter session. Chronic strength (0.10) and muscle size (−0.05) did not differ either. Its training-efficiency estimate (SMD 1.74, 95% CI 0.46 to 3.01) came from only two studies.

What you pair decides the result. Agonist and antagonist pairs allowed more reps than traditional sets (SMD 0.68, 95% CI 0.20 to 1.17). Pairing two similar movements for the same muscle cut volume load (SMD −1.08, 95% CI −1.72 to −0.44). Rows with presses, or curls with triceps extensions, are fine. Two chest movements back to back is just short rest with extra steps.

The cost is felt, not adaptive. Effort ratings were higher with supersets (SMD 0.77) and blood lactate was higher both during and after. Creatine kinase and acute swelling did not differ. Most participants were male (232 of the 313 whose sex was reported, and 66 female), and the long-term results rest on only three studies.

Probably not. Henselmans and Schoenfeld (2014) concluded the literature does not support the view that fixed rest beats self-chosen rest for muscle growth. Alonso-Aubin et al. (2024) tested it on the back squat in 13 trained lifters, five sets of four at 80% of 1RM. Lifters who chose their own rest took about 97 seconds before set two where the fixed group took 120. Bar speed, velocity loss and muscle oxygenation did not differ. That is thirteen people in one session, and nobody has tested the long-term question. It argues for treating a timer as a default you may override, not a rule.

Shojin's rest timer follows that ranking. Rest is set per exercise. Whatever you set on an exercise in a template counts down, and otherwise the timer uses your profile default (2:00 for a new account; older accounts started on 90 seconds). It starts only when you complete a working set. Warm-ups and drop sets never start it, and editing a set you already logged does not restart it. It also does not start after the last set of the workout. If you link two exercises into a superset, the timer stays quiet between the pair and starts when the round is done. You can add or take 15 seconds during the countdown, or skip it.

Rest also feeds your next-load suggestion, because reps at the same weight are what the suggestion reads. If you hurry through a session and lose reps, Shojin's next-load suggestion can read that as a stall. At RPE 9 or 10, fewer reps than last session at the same top weight backs the suggestion off by about 5%. If a session went short because you cut your rest, say so in the workout note and do not take the back-off as a verdict on the lift.

It cannot give you an exact number. It supports a floor of more than 60 seconds, and two minutes or more for heavy work if you are trained. It does not support a precise best rest. The strongest synthesis reports effect sizes of 0.13 to 0.17 with intervals that cross zero, and its authors say they had too few trained lifters to analyse them. Much of what is repeated about rest comes from acute studies that count reps inside one session, which is a different question from how muscle changes over ten weeks. Grgic and colleagues flag that gap themselves. The 2024 meta-analysis found no study of torso muscles, and its trials ran only five to ten weeks. The trained-lifter samples across this literature are small and mostly male, so we do not know that the two-minute strength advice holds for women or older adults. Nobody has tested it in them at scale.

Shojin is one subscription with no cap on sets. Shojin Coach is part of it: it writes your program from your own lifts and adjusts it after every session you log.

New subscribers get one month free, then A$4.99 a month or A$34.99 a year, billed by the App Store in your local currency. The subscription renews until you cancel. Cancel any time in Settings. You subscribe in the iPhone app. If you stop, your history stays, and Coach stops adjusting your program.

Every number in this article traces to one of these. Where the evidence is contested or thin, the article says so rather than picking a side.