Rep RoomForm Before Failure
Form Before Failure

How To Do the Incline Dumbbell Press

Everyone tells you to go incline for upper chest. Five different bench angles were tested to find where it actually peaks, and steeper is not better.

Start here

The incline press is a press on a tilted bench, aimed at the upper portion of your chest.

The question everyone gets wrong is the angle. Most gym benches sit at 45 degrees and most people just use them.

The short answer: set the bench around 30 degrees. Here is the study that gives you that number, and then the study that tells you not to get too excited about it.

Five angles, thirty lifters, one clear answer

Researchers tested the bench press at 0, 15, 30, 45 and 60 degrees with 30 trained adults, recording EMG at 60 percent of their max. They split the pectoralis major into three portions, upper, middle and lower, and also measured anterior deltoid and triceps.

The results are unusually clean.

Upper chest activity peaked at 30 degrees.

Middle and lower chest were highest at 0 degrees, which is flat.

The anterior deltoid, the front of your shoulder, was highest at 60 degrees.

The triceps showed similar activity at every angle.

So the steeper you tilt the bench, the more it becomes a shoulder press. At 60 degrees your front delt is the star. That is the trade nobody mentions when they set the bench to the steepest notch.

The authors also noted the flat bench produced similar activity for the pec major and anterior deltoid overall. Thirty degrees is where the upper chest specifically gets its peak.

Now the part that should keep you humble

Forty-seven untrained men were randomly assigned to horizontal bench press, incline bench press, or both combined. Eight weeks, once per week, with the number of sets equalised between groups. Muscle thickness, isometric strength and EMG were measured before and after.

Strength gains showed no difference between groups. Everyone gained about 10 kg on horizontal isometric strength and about 11 kg on incline isometric strength, regardless of which they trained.

Muscle thickness differed at only one of three measured sites: at the second intercostal space, high on the chest, the incline group gained most.

Read those two studies together and you get the honest picture. The EMG says 30 degrees excites the upper chest more in the moment. The training study says that after eight weeks, the difference showed up in exactly one measurement location and nowhere else, and strength was the same either way.

So angle is a fine-tuning dial for where you build, not a switch between working and not working. Untrained men, once a week for eight weeks, is also a fairly gentle test.

You can get better at using your chest

Sixteen experienced strength trainees were split into two groups for 10 weeks. One did targeted pectoralis training alongside the bench press, the other just benched.

The targeted group significantly increased pec major activity during the bench press afterwards, p equals 0.0002, effect size 2.6. The control group did not change. Anterior deltoid and triceps activity stayed roughly stable in both.

That effect size is large.

What it means practically: if you feel your shoulders and triceps doing everything on pressing, that is trainable. Direct chest work appears to change how much your chest contributes to the press itself.

Sixteen people, split into two small groups. Take the direction, not the precision.

Why dumbbells rather than a barbell

No study here compared dumbbells against a barbell on an incline, so I will not pretend there is evidence for the choice.

What I can say is what dumbbells structurally allow. Each arm moves independently, so a stronger side cannot carry a weaker one. Your wrists and shoulders can rotate to a position that suits you instead of being fixed by a bar. And the weight you can safely bail on is smaller.

That is reasoning about the tool, not a result. Use the version your shoulders prefer.

How to actually do it

Set the bench to about 30 degrees. If the bench only has fixed notches, take the shallower one, not the steeper one.

Sit back so your whole spine is supported and your feet are flat on the floor.

Set your shoulder blades down and back before the dumbbells move. Do not let them roll forward at the bottom.

Lower under control until your upper arms are roughly level with your torso. Going deeper because you can is how shoulders get unhappy.

Press up and slightly inward. Do not clash the dumbbells at the top.

End the set when your shoulders start rolling forward or the dumbbells start drifting apart. That is the position going, and on an incline the position is the whole point.

The takeaway

Set the bench to about 30 degrees, which is where upper chest activation peaked across five tested angles. Steeper is not more chest: at 60 degrees the front of your shoulder takes over, and the flat bench is where the middle and lower chest work hardest. Keep expectations calibrated though, because over eight weeks of training, incline versus flat produced the same strength gains and a thickness difference at only one of three measured sites. Shoulder blades down and back, control the bottom, and end the set when your shoulders start rolling forward.

Limits of this evidence

  • The angle study measured EMG at a single load (60 percent 1RM) during a session. Muscle excitation in the moment is not the same as muscle growth over months.
  • The eight-week training study used untrained men training only once per week, which is a low frequency, so it may underestimate differences that appear with more volume or in trained lifters.
  • That same study found a thickness difference at only one of three measurement sites, so the incline advantage it detected was narrow and localised.
  • The neuromuscular repatterning study had 16 participants split into two groups, so each group was very small despite the large reported effect size.
  • All three studies used the barbell bench press, not dumbbells, so applying them to an incline dumbbell press is inference.
  • The case for dumbbells over a barbell here is mechanical reasoning, not a finding; no source compared them.
  • Participants were male in the studies reported, limiting generalisation.
  • Each claim rests on a single source with no independent replication checked. Treat the piece as provisional.
This piece rates evidence on a conservative ladder that tops out at Moderate. We never publish certainty we do not have. See how we rate