What actually builds muscle?

The short answer

Muscle grows when you give it enough hard sets each week, take those sets reasonably close to failure, and keep adding work over months. Load, rep range, training split and rest periods are more flexible than gym lore suggests. The evidence for the big levers is solid and rests on several meta-analyses. The fine print, such as exactly how many sets or exactly how close to failure, is less certain.

What the evidence says

Volume: more sets, more growth, with diminishing returns

Weekly hard sets per muscle is the best-studied training variable. A 2017 meta-analysis of 15 studies found a graded dose-response: each extra weekly set was linked to roughly 0.37% more muscle gain, and the higher-volume groups within each study gained about 3.9% more than the lower-volume groups ^[1]. When the authors split volume into under 5, 5 to 9, and 10 or more weekly sets per muscle, the pattern pointed the same way but was only a trend (P = 0.074) ^[1].

A much larger analysis published in 2026 pooled 67 studies and 2,058 participants. It found that muscle growth keeps rising as weekly sets rise, but with diminishing returns: each added set buys a little less than the one before ^[2]. It also found that counting matters. Sets that train a muscle indirectly (think rows for biceps) were best counted as half a set ^[2].

Effort: close to failure, not necessarily to failure

A 2023 meta-analysis of 15 studies found only a trivial edge for training to failure over stopping short (effect size 0.19, 95% CI 0.00 to 0.37). When failure was defined strictly as being unable to complete another rep, the advantage disappeared (0.12, 95% CI -0.13 to 0.37) ^[3].

A 2024 dose-response analysis adds nuance. Muscle growth improved as sets ended closer to failure, while strength gains were similar across a wide range of effort ^[4]. The authors call their own analysis exploratory, because "reps in reserve" had to be estimated from each study's description rather than measured ^[4].

Put together: sets need to be hard. They do not all need to end in a failed rep.

Load and rep range: wide open for size, heavier for strength

A meta-analysis of 21 studies compared light loads (60% of one-rep max or less) with heavier loads, with every set taken to failure. Muscle growth was similar. Maximal strength improved more with heavy loads ^[5].

The catch is in the design. Light loads matched heavy loads when sets were pushed to failure. Twenty-five easy reps with a weight you could lift forty times is not the same stimulus.

Frequency: mostly a scheduling tool

How often you train a muscle matters less than how much total work it gets. A meta-analysis of 25 studies found no meaningful difference in growth between higher and lower frequencies when weekly volume was matched ^[6]. The 2026 analysis agrees: the effect of frequency on muscle size was compatible with negligible, although training a lift more often did help strength ^[2].

Rest between sets: at least a minute

A 2024 Bayesian meta-analysis of nine studies found a small growth benefit from resting longer than 60 seconds, probably because very short rests cut the reps you can do on later sets. It found no appreciable extra benefit beyond about 90 seconds ^[7]. The estimates overlapped heavily, so treat this as a nudge, not a rule.

Progressive overload: reps count too

Muscle adapts to what you ask of it, so the ask has to grow. It does not have to grow through load alone. In an eight-week trial of 43 trained lifters, one group added weight while keeping reps steady and the other added reps while keeping weight steady. Both built muscle to a similar degree ^[8]. The trial covered lower-body training only.

Protein and sleep: the supporting cast

A meta-analysis of 49 trials and 1,863 participants found protein supplementation added about 0.30 kg of fat-free mass on top of what training alone delivered. The benefit levelled off at a total intake of about 1.6 g of protein per kilogram of body weight per day ^[9]. That is a real effect, and a small one. Training does most of the work.

Sleep is harder to study. In a crossover study of 13 healthy young adults, one night with no sleep at all reduced muscle protein synthesis the next day by 18% ^[10]. That is a single extreme night in a laboratory, not a training study.

What is still uncertain

• Where volume tops out. Returns diminish, but the point where extra sets stop helping, or start costing recovery, is not pinned down and probably varies between people ^[2].
• Who was studied. In the largest analysis, 79.1% of participants were male and the average age was about 25 ^[2]. Findings may not transfer neatly to women, older adults or very advanced lifters.
• How close to failure is close enough. The best dose-response data rely on estimated, not measured, reps in reserve ^[4].
• The long game. Trials are short. The load meta-analysis accepted programs as brief as six weeks ^[5], and the overload trial ran for eight ^[8]. Years of training may behave differently.
• Sleep and actual muscle gain. The 18% figure comes from total sleep deprivation and a short-term lab measure ^[10]. How much habitual short sleep costs you in muscle over months is not established.

What it means in practice

1. Count hard sets per muscle per week. Around 10 or more is a reasonable target for most people who train, and fewer still builds muscle. Add sets gradually, and only while you are recovering well.
2. Make the sets hard. Finish most sets within a few reps of failure. You do not need to fail on every set.
3. Use the loads you like. Lighter and heavier weights both build muscle when the set is hard. If maximal strength matters to you, include heavy work.
4. Fit frequency to your life. Two full-body sessions or a five-day split can both work if the weekly sets are there.
5. Rest at least a minute between sets. Longer is fine, especially on big lifts.
6. Progress something. Add a rep or add load over time. Both count.
7. Eat enough protein. About 1.6 g per kilogram of body weight per day covers most people who lift.
8. Protect your sleep. The muscle-specific evidence is early, but nothing in it suggests short nights help.

This is general information for healthy adults who train. It is not personal medical advice.