Triceps extensions: overhead, lying, or by your side

Overhead extension, skullcrusher, pushdown: the muscle is the same, the growth you get is not.

The question always shows up as a duel between two exercises. Overhead extension or skullcrusher. Pushdown or kickback. Framed that way it has no answer, because what separates these movements is not the exercise: it is the shoulder angle while the elbow opens.

Only one of the three triceps heads crosses the shoulder. That is what makes arm position decisive, and that is the head most training sessions leave out.

The long head is the only one crossing the shoulder

The triceps brachii has three heads. The lateral head and the medial head attach to the humerus: their length depends on elbow angle alone. The long head attaches to the scapula, so it crosses two joints, the elbow and the shoulder.

The consequence is direct: when you raise your arms overhead, the long head lengthens before the elbow has moved at all. When you keep your arms by your side, it starts the movement already shortened. The lateral and medial heads see no difference between those two positions.

That is the one variable that truly separates an overhead extension from a pushdown. Everything else (barbell, dumbbell, rope, machine) changes comfort and resistance profile, not muscle length.

What Maeo et al. (2023) measured

Maeo et al. (2023) tested exactly this variable. 21 untrained people did 12 weeks of cable elbow extensions, one arm overhead and the other arm by the side, within the same person. 5 sets of 10 reps at 70% of 1RM, twice a week, with a 2-second lowering and a 2-second lifting phase. Muscle volume was measured by MRI before and after.

The long head gained +28.5% in volume on the overhead side against +19.6% on the by-the-side arm (p < 0.001). That is one and a half times the growth for the same number of sets.

The unexpected result concerns the other two heads. The authors expected a tie, since the shoulder does not change their length. The gap showed up anyway: +14.6% against +10.5% (p = 0.002). And the overhead side produced those gains with lower absolute loads.

The principle is not specific to the triceps. Pedrosa et al. (2022) found it on the quadriceps and Maeo et al. (2021) on the hamstrings: at equal volume, training a muscle in a lengthened position produces more hypertrophy than training it shortened.

The 67 triceps exercises in the Overload reference library

The Overload exercise library holds 891 movements, 72 of which have the triceps as primary muscle. Five are set aside here: three stretches and two movements whose triceps tag is a data error. That leaves 67 strength exercises, sorted by arm position during the extension, and crossed with the number of sessions members actually logged.

Arm position	Long head	Exercises	Member sessions	Share
By the side (pushdown, dips, close-grip press, kickback)	shortened	39	3,862	74.9%
Perpendicular to the torso (skullcrusher, lying and incline extensions)	intermediate	14	812	15.8%
Overhead (seated triceps press, cable rope overhead, one-arm dumbbell)	lengthened	14	479	9.3%

The ranking of the most-logged movements says the same thing. The top three are the triceps pushdown (1,201 sessions), the parallel bar dip (548) and the rope pushdown (529). The three most-used exercises in the library all train the long head in its shortened position.

A note on method, because it changes how the table reads: the sort is based on the exercise's English name, which describes arm position reliably in this library. The JM press and the incline extensions are intermediate cases that I placed in the middle row by convention. The orders of magnitude carry the information here, not the exact row of any single exercise.

What the imbalance means

Three quarters of the logged triceps work happens with the arms by the side, in the position where the long head is at its shortest. Less than a tenth happens overhead.

That is not surprising. The pushdown station is in every gym, it is comfortable, and it loads heavy. An overhead extension needs a setup, pulls on the shoulder when mobility is short, and forces the load down. The most convenient position and the most effective position are not the same one.

This imbalance is also why many lifters end up with a triceps that looks thick from the side and flat from behind. The long head sits on the back of the upper arm.

Practical recommendations

• Put an overhead extension in every training week, not in every session. One to two sessions a week is enough, and that is the protocol Maeo et al. tested.
• Expect to cut the load noticeably when you move to the overhead position. Participants got more growth with lower absolute loads; the weight on the stack is not the marker here.
• If the shoulder limits your overhead range, the seated version with a back support or the low-pulley rope reduces the strain while keeping the arm above the head.
• Do not replace dips or the close-grip press. They load the triceps heavy inside a compound movement, which no isolation extension does.
• Choosing a cable or a free weight for a given exercise does not change the growth you get, as covered in the guide on cable training. Arm position is what counts, not the equipment.

Conclusion

The duel between overhead extension and pushdown was never about the exercise. It is about shoulder angle, and only one of the three triceps heads responds to it. Overload member data says that head receives less than a tenth of the logged work. One overhead extension a week is enough to fix that.

This article presents the current scientific evidence. For a personalized program, consult a health professional.