This is why you should stop doing the reverse dumbbell fly (and what you should do instead)

The posterior deltoid is a muscle that gets much attention in the gym because it’s needed to achieve the “3D deltoids” you want. The posterior delt gets trained well when you perform compound pulling exercises such as chin-ups, pull-ups, and rows. However, this is often not enough for serious trainees that are looking to squeeze out every ounce of volume they can give the stubborn muscle to grow it.

Many trainees “isolate” the posterior delts on their shoulder day by performing the popular exercise, the dumbbell reverse fly. In this article, I’m going to explain to you why the dumbbell reverse fly is a far cry from an optimal exercise for the posterior delt and give you an exercise that does a much better job at training the posterior delt along with the synergist muscles involved.

The dumbbell reverse fly

The anatomical movement associated with the dumbbell reverse fly is transverse abduction. When your shoulder is internally rotated, as it is when you execute this exercise, the posterior delt is the primary transverse abductor. The problem, though, is this: the exercise possesses the dreaded descending resistance curve. A descending resistance curve means resistance is lowest when the muscles are in a stretched position and highest when they are at full contraction.

From the muscle length-tension relationship, your posterior delt is stronger at the bottom of the movement and much weaker at the top. The dumbbell reverse fly offers very little resistance at the bottom, but a large amount of resistance at the top, the exact opposite of what you want. Moreover, the internal moment arm (the fulcrum a muscle uses to create torque on a limb) of your posterior delt and rotator cuff muscles (which are synergists during transverse abduction) decreases over the last 30° of the exercise, making it even harder for those muscles to produce torque near the top.

If you’ve ever included this exercise in your routine for quite some time, while doing your best to adhere to progressive overload, you’ve no doubt run into issues with either form or range of motion (ROM). It becomes a royal pain in the ass to progress on this exercise because of the descending strength curve and reduced moment arm of the transverse abductors.

I’ve encountered countless people at the gym that work up to a weight in which they teeter on the edge of no longer being able to perform full ROM reps without cheating the exercise, so they end up camping out at a certain weight.

If you’re going to sit at a certain weight x reps, that will do virtually nothing to help grow the muscles involved; you need to figure out a way to break through the plateau. In this case, you need to do a better exercise.

Optimal variation: Reverse Bayesian fly

The reverse Bayesian is essentially the same anatomical movement performed with a cable tower instead of a dumbbell (You can also perform this exercise standing, bilaterally or unilaterally). A subtle key point for the exercise is to use a rope or even the bare cable such that your shoulder is internally rotated, as shown in the video. The resistance curve for this exercise is night and day compared to the dumbbell reverse fly.

As you can see, this exercise has a bell-shaped resistance curve.

A good place to position the cable height is such that the cable is perpendicular to your arm when your arm is about perpendicular to the floor. At this point, the resistance peaks where the length-tension relationship is good, which significantly reduces the sticking point and allows your muscles to perform more work on a per rep basis.

As you ascend from the bottom to the top of the exercise, resistance decreases in-line with the weakening of your transverse abductors up to full contraction. Standing further away from the tower, about three feet or so, is also desirable. The further you stand away from the tower, the less resistance there will be at the top of the exercise. You can play with your positioning a little, of course, I’ve found the setup shown in the video works well.

What’s more, with the reverse Bayesian fly, you’re loading the transverse abductors in the fully stretched position at the bottom of the exercise, which is important for inducing muscle hypertrophy not only in parallel but in series (lengthening of the muscle belly) as well [2]. The posterior delt is also activated over a much greater effective ROM (during the dumbbell fly, you’re under stimulating the posterior delt during the first 30° or so of ROM because resistance is low), which is important for inducing more overall hypertrophy [2,3].

Furthermore, as shown in the graph above, the reverse Bayesian fly is also performed with more overall ROM, which is conducive to more muscle growth [2,3].

Scapular retraction

Like the dumbbell fly, the reverse Bayesian fly also involves scapular retraction and protraction (really, they’re both compound exercises, not isolation exercises), so the mid and lower trapezius muscles are also stimulated. During the reverse Bayesian fly, the trapezius muscles also experience a bell-shaped resistance curve, similar to the transverse abductors. So, your traps will also be better stimulated. It’s a win-win.


It helps to have access to a good cable tower that allows for small incremental increases in weight or have access to magnetic plates, so you can gradually load the exercise. If you’re unable to gradually load this exercise, this will likely become a problem.

Many cable towers only allow for 10 lb jumps in weight. So, if you work your way up to doing 50 lbs for ten reps and have to jump up to 60 lbs next session, that’s a 20% increase in weight, which is a significant relative increase. Think about it like this; imagine hitting your rep target for the squat at 300 lbs and having to add 20% next session, that means you’d be squatting 360 lbs. Microloading is an underrated aspect of exercise selection; you need to be able to do this for the reverse Bayesian fly.

Leave a Comment

Your email address will not be published. Required fields are marked *