How To Do the Barbell Curl

Start here

A curl is a curl. You bend your elbow against resistance.

The arguments are all about details: which grip, how far to go, and whether starting with your arm stretched behind you builds more muscle.

That last one is everywhere right now. Behind-the-body cable curls, incline curls, anything that puts the biceps on stretch at the bottom. The claim is that training at long muscle lengths grows more muscle.

Two studies tested it properly, and I am going to lead with them because the answer is not what the internet says.

Test one: stretched shoulder versus neutral, ten weeks

Thirty untrained men did cable curls with one arm at their individual peak shoulder extension angle, meaning arm stretched back behind them, and the other arm at a neutral position. Same person, both conditions, which removes almost all the noise between people.

The researchers matched the resistance profile and the elbow range of motion between conditions, so the only real difference was shoulder position. Six to eight sets per week, to momentary failure, supervised, for ten weeks. Twenty-four finished. Muscle thickness was measured by ultrasound at two points along the arm, and the analysis was Bayesian.

Both conditions grew the muscle by roughly 7 to 9 percent.

The estimated difference between them sat essentially at zero. The proximal measurement came out at minus 0.40 mm with a credible interval from minus 1.06 to 0.26 mm, which straddles zero comfortably.

So, arm stretched back or arm neutral: same growth.

Test two: partial reps at long length versus full range, eight weeks

Thirteen trained people, mostly men, did unilateral preacher curls for eight weeks with one arm doing only the stretched half of the range, 0 to 70 degrees, and the other arm doing the full range, 0 to 140 degrees. Again the same person for both.

Muscle thickness came out similar between conditions at the 50 percent measurement site. At the 70 percent site, the partial-range-at-long-length arm showed what the authors called trivial to small superiority, standardised mean difference 0.10 with a Bayes factor of 4.87.

A standardised difference of 0.10 is tiny. The Bayes factor says the evidence leans real, but the size of the thing it is detecting is very small.

Put both studies together and the honest summary is this. Training at long muscle lengths is not worse. It might be very slightly better in one specific spot. It is nowhere near the difference the content about it implies.

Do the version you will actually do consistently.

Grip, and the phase that actually matters more

Twelve resistance-trained men performed barbell curls with three grips: fully supinated, semi-supinated and neutral. EMG on the biceps brachii and brachioradialis, normalised to maximal voluntary contraction, analysed in a Bayesian framework, split by contraction phase.

The phase findings were stronger than the grip findings.

Concentric contractions, the lifting half, produced greater activation than the eccentric half in both muscles. Repetitions combining both produced the highest activation of all, with what the authors graded as strong to extreme evidence.

Between muscles, the brachioradialis was more active than the biceps during the eccentric phase, while the lifting and combined phases were similar between the two.

The practical read: do full repetitions. Lower it as well as lift it. The combined action is where the activation is, and the forearm is quietly doing a lot of the lowering.

Twelve men, EMG in a single session. This is about muscle excitation, not growth.

So what is a curl actually for

The unglamorous answer: adding elbow flexion volume that rows and pulldowns do not fully provide.

All three studies here grew or activated the elbow flexors with fairly ordinary work. The 10-week study used six to eight sets a week to failure and got 7 to 9 percent thicker muscle in untrained men.

That is the useful number. Consistent sets, taken close to failure, over weeks. Not the angle you picked.

How to actually do it

Grip roughly shoulder width, palms up. The tested grips did not separate much, so pick what your wrists and elbows tolerate.

Keep your elbows near your sides. If they swing forward, your front delts start lifting the bar.

Lift with control, and lower with control. The lowering is not a rest phase, and the combined rep produced the highest activation in the tested conditions.

Do not heave with your lower back. If your torso is rocking, the weight is choosing your technique.

Go through a range your elbow tolerates. Full range and long-length partials both worked; neither is a magic setting.

End the set when your elbows start drifting forward or your body starts swinging.