Number of sets, reps, time under tension: what should you count to build muscle?

One gym, two discourses. On one side: "do 3 sets of 10". On the other: "forget reps, what matters is time under tension — your muscle must work 40 to 60 seconds per set". Good news: both ideas have been tested head-to-head, stopwatch in hand. And the answer looks like neither.

This story of seconds has been tested directly

If time under tension were the right unit of measurement, then two sessions with the same total number of seconds should yield the same muscle — regardless of how those seconds are sliced up. This is exactly what a 2022 study verified: 38 untrained men, 10 weeks, and sets calibrated to 36 seconds each, some in 3-second reps, others in 6-second reps.

Result: similar strength gains (1RM), similar pectoral and triceps hypertrophy. At equal time under tension, execution speed changed nothing. The stopwatch, for its part, "predicted" the result well.

Except that the stopwatch also works when you unplug it

The real test is elsewhere: what happens when we vary the time under tension instead of equalizing it? A meta-analysis gathered studies that compared different tempos, all carried to muscular failure. Verdict: hypertrophy is similar for repetition durations ranging from 0.5 to 8 seconds.

Reread this number. Between a half-second rep and an eight-second rep, there is a sixteen-fold factor on time under tension. And the muscle doesn't tell the difference. A set of 10 explosive reps (5 seconds of tension) and a set of 10 very slow reps (80 seconds of tension) produce the same growth. The authors nevertheless set an upper limit: beyond 10 seconds per rep, it would probably become less effective — but few controlled studies exist on this point.

Nor is the stopwatch a neutral measure of volume: a 2006 trial compared, in 18 trained men, what three protocols yield depending on whether work is quantified in time under tension or in load volume — the two ways of counting do not describe the same session. And at equal repetition duration (6 seconds), lengthening the concentric phase rather than the eccentric one increases muscle activation and fatigue: the same total duration, two different sets.

Important nuance: not all seconds are equal

Time under tension is not irrelevant for all that — it is simply not sufficient. Two studies show this clearly. In 2012, a study on light load (30% of 1RM) compared 6-second repetitions to 1-second repetitions for an equivalent work volume: the slow version produced a stronger rise in muscle protein synthesis, notably a delayed but robust myofibrillar response 24 to 30 hours after the training session.

But in 2015, another study compared three ways of occupying exactly the same time under tension with very different force levels. It was the protocol that produced the most force (maximal eccentric) that triggered the strongest anabolic signaling — not the one that caused the most fatigue. In other words: a second spent producing a lot of force is not worth a second spent lingering under a light load.

The unit that really works: the hard set

If it is not the stopwatch, then what? A 2021 systematic review asked the question head-on: is the total number of sets a valid way to quantify training volume? After analyzing 14 studies, the answer is yes — on one condition: that these sets are taken to failure or close to failure. And this remains valid across the entire range from 6 to over 20 repetitions.

Better yet: that unit has a measurable dose-response relationship. A 2017 meta-analysis (15 studies, 34 groups) calculated that each additional weekly set was associated with +0.37% of muscle mass gain. One more second means nothing; one more set, yes.

A 2019 study in 49 trained men tested 9, 18, or 27 weekly sets for the biceps over 6 weeks. Muscle thickness increased in all groups — 9 sets per week are already enough to make a muscle grow — even if effect size estimations suggest that 18 to 27 sets could provide greater strength.

The honest counterpoint: an analysis on the upper body finds no statistically significant advantage to doing 3 sets or more rather than fewer than 3 in beginners, and highlights the lack of well-designed studies in trained practitioners. « More sets » is therefore not a law without a ceiling — it is a trend, with real gray areas.

What about the number of repetitions per set?

Here too, the useful range is much wider than we think. A 2017 meta-analysis covering 21 studies compared light loads (≤ 60% of 1RM) and heavy loads (> 60%), with all sets taken to failure: no significant difference on hypertrophy. Maximum strength gains, on the other hand, lean clearly toward heavy loads.

The message is constant: to make a muscle grow, the zone that works ranges from about 6 to over 20 repetitions. To get strong, you need heavy weights. These are two goals, not one.

The real hidden variable: how far you go

This is what actually decides — and it is what time under tension accidentally captures when it "works". A 2022 study on 25 untrained men crossed two factors: heavy or light load, sets taken to muscular failure or stopped before. Three out of four groups progressed equally; the fourth dropped off.

Three bars on the chart are almost identical; the fourth collapses. With light loads, stopping far from failure builds almost nothing: the effort is never hard enough to recruit the muscle. With heavy loads, on the other hand, going to failure brings no additional benefit — the load is already doing the work. (On pure strength, heavy loads crush light ones: +33.8% and +33.4% on 1RM vs +17.7% and +15.8%.)

Two other trials confirm that, as long as the set is hard enough, going strictly to failure is not mandatory: over 12 weeks in 32 untrained men, muscular failure and voluntary stopping gave similar muscle cross-sectional area gains (6.1 to 7.5%); and over 10 weeks in trained lifters, failure and non-failure produced comparable gains (13.5% vs 18.1% of vastus lateralis cross-sectional area).

Two methodological caveats, to finish. A 2019 study points out that part of the gain measured early in a light-load protocol is edema, not muscle — a reason to be wary of trials that are too short. And a 2026 trial in 34 runners shows that these conclusions, established in strength training, do not mechanically transpose to other muscular actions or other populations.

What this looks like in practice

Summary

The muscle doesn't count seconds: it counts hard sets. Time under tension is a thermometer, not a lever — it goes up when you work hard, but making it go up artificially by slowing down doesn't make you grow more. Focus on three things: enough sets in the week, a load that puts you between 6 and 20 reps, and a real effort at the end of the set.

One question remains: should these sets be spread over two, three or five sessions? We answer that, studies to back it up, in how many sessions per week to build muscle.

Every figure in this article comes from a study you can open: the underlined mentions lead to the work from which the result is drawn. These summaries are written from the abstract alone. Note: several of these works focus on beginners or short durations (6 to 12 weeks), and studies comparing volumes in trained practitioners are still lacking. Informative content, which does not replace the advice of a professional.