Rotator cuff mobility is a phrase with a small error inside it. The four muscles it names are not the ones that run out of length when your arm will not go where you want it. Their job is to hold the ball in the socket while larger muscles swing the arm, which makes them a control problem rather than a length problem — and control and length respond to different work.
That does not make range work at the shoulder pointless. It makes it honest about what it is doing, which is a different and better position to write from.
The short answer
- Four muscles: supraspinatus, infraspinatus, teres minor, subscapularis. Their job is to keep the ball centred in the socket.
- They are mostly stabilisers, so "stretching the cuff" is not really what any shoulder stretch does.
- What can be worked is rotation range — the arm turning in and out at the socket — and the tissue behind the joint.
- On a throwing side, total rotation matters more than the loss of internal rotation on its own.
- The strongest clinical result in this region is strength training, not stretching.
What the rotator cuff is
Four muscles run from the shoulder blade to the top of the upper arm bone, and together they form a sleeve around the joint.
The shoulder is a shallow socket carrying a large ball, which is why it has more range than any other joint in the body and why it is the one that dislocates. Nothing bony holds it in. The cuff does: as the big movers — the deltoid, the lats, the pecs — pull the arm around, the cuff pulls the head of the bone into the socket and keeps it centred so that the movers have something to work against.
Three of the four also turn the arm outward or inward. Subscapularis, on the front of the blade, rotates in. Infraspinatus and teres minor, on the back, rotate out. Supraspinatus, over the top, helps start the arm rising.
So the cuff does produce movement. It is just that its main contribution is holding a joint together during movement that other muscles are producing.
Rotator cuff mobility, honestly
Here is what follows for anyone searching for a stretch.
A muscle whose job is stability sits short and works isometrically most of the time. It is not the structure that stops your arm going overhead — that is much more often the upper back, the shoulder blade's ability to rotate on the ribs, and the capsule and tissue at the back of the joint. Stretching aimed at "the cuff" is nearly always landing somewhere else, which is fine, as long as nobody claims otherwise.
What can honestly be worked at this joint is rotation range in both directions, and the movement of the arm through the whole circle. What cannot be claimed is that lengthening anything addresses a painful cuff. The evidence for that is not there, and the section below prints the reason.
The back of the shoulder
The one genuine length finding in this area is the tissue behind the joint. A shoulder that has thrown, served or slept on the same side for years can lose internal rotation — the arm turning in — and the restriction sits at the back of the capsule.
Three moves in Limber's catalogue work that direction.
| Move | What it is | Dose |
|---|---|---|
| M033 Sleeper Stretch | Side-lying, arm rotated in | 30 s per side |
| M034 Cross-Body Shoulder Hold | Arm drawn across the chest | 30 s per side |
| M036 Standing Shoulder CARs | Slow circle at the joint | 40 s per side |
The sleeper stretch is the app's level 3 — its hardest tier — and it is the only shoulder move in the catalogue carrying a contraindication flag. It also carries an instruction to ease off at any pinching in the front or top of the shoulder, which is the sensation that means the position is doing something other than stretching the back of the joint. The cross-body hold reaches similar tissue with far less to go wrong, and is the sensible place to start.
The shoulder circles are not a stretch at all. They are the joint moved slowly through its own outline under your own control, which is the closest thing on the list to training the cuff for what it actually does.
Total rotation, not the deficit
Wilk and colleagues, in the American Journal of Sports Medicine in 2011, examined shoulder rotation in professional pitchers. The finding that has lasted is that total rotational motion — internal plus external, added together — was more informative than the loss of internal rotation taken alone.
That is worth understanding rather than repeating. A throwing shoulder trades range: it gives up internal rotation and gains external rotation, and the sum often stays close to the other side. Look only at internal rotation and you see a deficit and reach for a stretch. Look at the total and you may see a shoulder that has adapted correctly to what it does.
The general rule follows. An asymmetry with an obvious cause in what you do is a finding to be curious about rather than alarmed by. A racket arm, a throwing arm, a side you have slept on for a decade — those are explanations, and closing the gap is not automatically the goal. What this argues for is knowing which of your differences has a reason behind it.
Grade: reasonably solid for the measurement point, in throwing athletes. It has been stretched a long way past that population.
Where the evidence actually is
Now the part that goes against the standard article on this subject.
The mechanical account of shoulder pain has not held up. Beard and colleagues' CSAW trial in The Lancet in 2018 compared the decompression operation against placebo arthroscopy and against no treatment; both surgical groups improved and did not meaningfully differ from each other. That is set out properly in shoulder impingement and mobility work, and it is the reason Lewis, in Manual Therapy in 2016, renamed the whole area rotator cuff related shoulder pain — a name that describes where it hurts and declines to explain why.
Page and colleagues reviewed exercise for rotator cuff disease for Cochrane in 2016 and found low-quality evidence, small benefits, and no programme clearly better than another. That is the ceiling on what anybody can claim, including us.
Grade: moderate for exercise in general, weak for any specific mobility protocol, absent for stretching as a treatment of cuff symptoms.
Note the third one. There is no trial showing that a stretching programme addresses symptoms at this joint, so nothing on this site says so.
The strongest result is strength
The best clinical result in this region is not a stretch, and printing it here is the point of the section.
Andersen and colleagues, in Pain in 2011, gave office workers two minutes a day of progressive resistance training for the neck and shoulder muscles and found a real reduction in neck and shoulder pain over ten weeks. Two minutes. With a randomised design and a clinical outcome.
No stretching routine has a trial like that behind it. Not one. If you have a shoulder that aches through a working day and you want the intervention with the best record, the honest answer is a small daily dose of resistance work, and that answer is inconvenient for a mobility app to give.
Grade: reasonably solid, and it points away from stretching.
Range work still earns its place. It changes what a joint can do, which is measurable and which is our subject. It just is not the thing with the clinical trial behind it, and you should know that before you spend six weeks on it.
What to measure, and when to stop
Take a reading before you decide anything. The shoulder mobility test is the Apley scratch reach — one arm over the shoulder, the other up behind the back — recorded on both sides in five bands: 10, 30, 55, 75, 95. It reads shoulder flexion, extension and rotation together with the blade's travel on the ribs, so a poor reach tells you the chain is short somewhere without telling you where.
One caution about the sides, because it is our own instrument's limit. The smallest gap that test can express is 20 points, and Limber calls 12 points an imbalance — so any difference at all between your arms is flagged, and a one-band gap means measure again rather than measure a problem. The Range Score folds the reading into the shoulders group and leaves the rest of the ground blank.
And the line that comes before all of it. Pain that persists, wakes you at night, radiates into a limb, comes with numbness or weakness, or followed a fall or a wrench is a matter for a clinician who can examine you. An arm you cannot lift at all, or shoulder pain with fever, belongs in the same category. Nothing here diagnoses anything and nothing here is treatment.
The rest of the region is in shoulders and upper back, the range work itself in shoulder mobility exercises, and the thinnest evidence in the cluster in sleeping position and shoulders.
Questions
Can you stretch the rotator cuff?
Not in the way the phrase suggests. The four muscles are stabilisers that spend most of their working life holding a joint together rather than lengthening, and the tissue that actually restricts shoulder rotation is usually the capsule behind the joint. A cross-body hold or a sleeper stretch reaches that tissue; calling it a cuff stretch is loose language rather than anatomy.
Is the sleeper stretch safe?
It is the hardest tier in Limber's catalogue and the only shoulder move carrying a contraindication flag, which should tell you how it is regarded. Ease off at any pinching in the front or the top of the shoulder — that sensation means the position has stopped stretching the back of the joint. The cross-body hold reaches similar tissue with much less to go wrong.
Why is my throwing arm tighter?
Because it has adapted. A throwing or racket shoulder reliably loses internal rotation and gains external rotation, and research in professional pitchers found the total of the two more informative than the loss on its own. A difference with an obvious cause in what you do is worth curiosity rather than alarm, and closing it is not automatically the goal.
Do cuff exercises help shoulder pain?
Exercise-based care is what guidelines recommend, and the Cochrane review behind that found low-quality evidence, small benefits and no programme better than another. Separately, two minutes a day of progressive resistance training reduced neck and shoulder pain in office workers in a randomised trial. That result is about strength, not stretching, and it is the strongest thing in this area.
How often should I work shoulder rotation?
Short and frequent beats long and rare, because the limit is usually time spent in one shape rather than a structure that needs breaking down. Daily is reasonable for the gentle work — a cross-body hold and a slow circle at the joint take under three minutes for both sides. Re-measure on a slow clock, three weeks apart, in the same conditions.
Can a range test tell me if my cuff is torn?
No. A reach test is a screen, it reads three joints at once, and it says nothing about the state of any tissue. Tears and tendon changes also appear routinely in shoulders that do not hurt, which is why imaging alone does not settle the matter either. Weakness, an arm you cannot lift, or pain after an injury needs a clinician who can examine you.