Dumbbell 21s

What dumbbell 21s actually train

Dumbbell 21s are a curl protocol rather than a distinct exercise: seven reps from a straight arm to halfway, seven reps from halfway to the top, then seven full reps, performed back to back without rest.

The biceps brachii does the work, with the brachialis contributing to elbow flexion and the brachioradialis assisting, exactly as in any other curl. Nothing about the rep scheme changes which muscles are involved.

What it changes is time under tension and fatigue. Twenty one reps without a break, at a load light enough to finish them, produces a burn that people reliably interpret as evidence that something special is happening. It is worth separating what is real from what is folklore.

The real part is that this is a lot of hard reps in one set. Volume and effort drive growth, and 21s deliver both, at a load your elbows will not object to.

The folklore is that the two partial blocks somehow target different parts of the biceps. They do not. The bottom half and the top half of a curl load the same muscle at different lengths and different leverages. Schoenfeld and Grgic reviewed range of motion and muscle development and found full range work generally at least matches partial work, with the caveat that partials performed at long muscle lengths hold up better than partials at short lengths.

That caveat is the one interesting thing about the protocol. The bottom seven reps are long length partials, which is the useful kind. The middle seven are short length partials, which is the less useful kind. The final seven are ordinary full reps performed under heavy fatigue.

So 21s are not nonsense, and they are not magic. They are a structured way to accumulate a lot of biceps work in ninety seconds.

Setting up and executing

Take a pair of dumbbells at roughly half your normal curl weight. This is the part everyone gets wrong on the first attempt.

Stand tall with the arms straight, palms facing forward. Brace the abs and pin the elbows against your sides.

Curl from a straight arm to about ninety degrees, lower to full extension, and repeat for seven reps. Pick the halfway point deliberately and hold it for all seven.

Without resting, curl from ninety degrees to the top and back down to ninety, for seven reps.

Without resting, perform seven full range reps from a straight arm to the top.

Keep the torso still throughout. Put the dumbbells down and rest properly before the next set, because the set was longer than it looks.

The mistakes that cost you the most

Starting far too heavy is the standard first attempt. Most people pick their usual curl weight, complete the first seven, struggle through the second seven, and abandon the full reps entirely. Half your normal load is the honest starting point and you can add from there.

Resting between the blocks removes the only thing that distinguishes this from three ordinary sets of seven. If you need a pause, the weight is wrong.

Swinging in the last block is almost guaranteed if you have not planned for it. By the final seven the biceps are heavily fatigued and the torso and shoulders will happily contribute. A rep that starts with a hip thrust is not a curl. Stop the set instead.

Vague halves make the protocol incoherent. If the bottom block drifts up to three quarter reps, the top block has nothing left to cover. Choose a halfway point, ideally forearms parallel to the floor, and keep it.

Using 21s as your main biceps work is the biggest programming error. Marcolin and colleagues found curl variants differ in how they load the elbow flexors, but load is the variable that partial protocols sacrifice. Heavier full range curls should carry the bulk of your training and this should finish it off.

Sets, reps and rest by goal

The reps are fixed at twenty one, so the variable is sets and rest.

For muscle gain, two sets with ninety seconds rest, at the end of your biceps work. For strength, two sets with two minutes, though 21s are a poor strength tool and heavier straight sets are better. In a fat loss phase, three sets with a minute, where the higher metabolic demand suits the context.

Do not run this every session. The novelty and the discomfort make it feel productive, which encourages people to do it far more often than the stimulus justifies.

Schoenfeld and colleagues found a dose response between weekly hard sets and hypertrophy, and a set of 21s counts as one hard set, not three. Counting it as three is how people end up believing they trained arms harder than they did.

Refalo and colleagues found stopping a rep or two short of failure produces comparable growth to going all the way, which is worth remembering when the final block turns into a whole body movement.

Control the lowering phase across all three blocks. Schoenfeld and colleagues found eccentric actions produce hypertrophy at least equal to concentric actions, and the descent is the easiest thing to abandon when a set gets long.

Where it fits in your split

At the very end of a pull or arm session, as the last thing you do.

On push, pull and legs it closes pull day. On a bro split it finishes arm day after the heavier curls. On an upper and lower split it can appear occasionally in one upper session. On a full body programme it is rarely worth the time.

Frequency should be low. Once a week, or once a fortnight, keeps it as a change of stimulus rather than a habit.

The strongest case for it is practical rather than physiological. It gets a lot of biceps work done in one set with a light weight, which suits the end of a long session, a busy gym with only light dumbbells free, or a training phase where your elbows are unhappy with heavy loading.

Variations worth knowing

Barbell 21s use a bar instead, which keeps the two arms symmetrical and makes the halfway point easier to judge.

Cable 21s run the same protocol on a low pulley, which keeps tension constant across all three blocks rather than letting it fall away at the top.

Preacher 21s are performed with the arms on a preacher pad, which removes any possibility of swinging and makes the last block far more honest.

Hammer 21s use a neutral grip, shifting more work to the brachialis and brachioradialis, and are noticeably more forgiving on the wrists.

Reverse order 21s run the top halves first, then the bottom halves, then the full reps, which is considerably harder because the long length work arrives already fatigued.

And ordinary straight sets of curls are the honest comparison. They achieve the same result with more load and less ceremony, which is why they should still make up most of your curling.

What the evidence says

Marcolin and colleagues compared electromyographic activity of the biceps brachii and brachioradialis across curl variants and found activity patterns shift with arm and grip position, which supports treating variations as different loading profiles rather than as different muscles.

Schoenfeld and Grgic reviewed range of motion and muscle development and concluded full range training is generally at least as effective as partial training, which is why 21s belong at the end of a session rather than in place of full reps.

Pallarés and colleagues reached a similar conclusion in their meta-analysis, finding full range training produced superior strength and hypertrophy outcomes compared with partial range work.

Vigotsky and colleagues cautioned against interpreting muscle activity amplitudes as a direct measure of growth, which is a useful check on claims that a particular rep block targets a particular region of the biceps.

Schoenfeld and colleagues found eccentric actions produce hypertrophy at least equal to concentric actions, supporting a controlled descent through all three blocks rather than a fast drop once fatigue sets in.

Schoenfeld and colleagues on weekly volume and Refalo and colleagues on proximity to failure supply the programming, and together they explain why one or two sets of this is plenty.

Common questions

Do 21s build bigger arms than normal curls?

No. They are a way of accumulating a lot of fatiguing reps at a light load, which is useful, but full range curls with more weight remain the better foundation. Use 21s as a finisher, not as the main course.

Why seven reps per block?

Because someone chose a number that adds up to twenty one and the name stuck. There is nothing special about seven. Ten, ten and ten would work identically, and would be called 30s.

What weight should I use?

Roughly half what you curl for sets of ten. Most people start far too heavy, fail to complete the last block, and conclude the exercise is stupid. It is not stupid, it is just twenty one reps.

Does the bottom half or the top half matter more?

The bottom half, if either. Schoenfeld and Grgic found partials at long muscle lengths hold up better than partials at short lengths, and the bottom of a curl is the long length position. The top half is the least useful block in the protocol.

Is this any different from barbell 21s?

Barely. The dumbbell version lets each arm move independently and lets the wrists rotate slightly, which some people find more comfortable. The bar version is easier to keep symmetrical. Pick whichever is free and do not think about it further.