The quietest of the flexibility goals and the one with the clearest anatomy. Overhead reach — getting an arm straight up beside the ear, without the ribs flaring and the lower back arching to finish the job — is a chain of three joints, and it fails at whichever one is stiffest.
The reason it is worth naming as a goal is that the failure is invisible. A back that arches makes the arm look vertical while the shoulder has not moved, so most people who cannot get overhead believe they can.
The short answer
- Three contributors: the shoulder, the shoulder blade on the ribs, and the thoracic spine.
- The lower back substitutes for all three, silently. Testing against a wall removes the substitution.
- Full shoulder flexion in reference data is roughly 165 to 170 degrees.
- The thoracic spine is the part a desk day removes and the part nobody trains.
- We have no thoracic reading in the battery. That gap is stated rather than papered over.
What overhead reach asks for
Three things move, and they move in a fixed relationship.
The shoulder joint itself supplies most of the elevation — the arm rotating up in the socket, which requires the upper arm to turn outward as it rises. Try lifting your arm overhead with the palm facing backwards and you will find the ceiling early; the same arm with the palm turned in goes considerably further.
The shoulder blade has to rotate upward on the rib cage. The two are geared together: a blade that will not rotate limits the arm regardless of how good the joint is. Boone and Azen's reference measurements of joint motion in the Journal of Bone and Joint Surgery in 1979 put full shoulder flexion at roughly 165 to 170 degrees in their sample — a figure worth holding lightly, since it came from 109 male subjects.
The thoracic spine has to extend. The blade slides on the rib cage, and if the rib cage is fixed in a flexed position the blade has nowhere useful to go. This is the link a screen day removes and the one almost nobody works.
When any of the three runs out, the lumbar spine finishes the movement by arching. That is the substitution, it is nearly universal, and it makes the position look complete.
The wall test
The whole diagnosis takes twenty seconds and needs a wall.
Stand with your back to it. Get the back of your head, your upper back and your tailbone on the wall, then press the small of your back flat against it — hard enough that a hand cannot slide behind. Now raise both arms in front of you and take them overhead, keeping the lower back flat.
- Wrists reach the wall easily, back stays flat. Full overhead reach.
- Wrists reach only if the back leaves the wall. That is the substitution, and it is the common result. Your usable overhead range is where the arms were when the back lifted.
- Elbows bend to get there. Same finding, different cheat.
- Sharp pain at the top or the front of the shoulder. Stop, and see the referral note below.
That test is the same idea as the wall angel, which is one of the two tests in Limber's design that is not built — in the app or here. Worth saying rather than implying we measure it.
What limits it, in order
Thoracic extension, most often. The upper back is the segment a working day removes entirely, and it is the least worked of the three. A seated tall rotation and an extension over the back of a chair are the two cheapest ways to visit it; both are covered in desk and posture.
Shoulder external rotation, second. The arm has to turn outward as it rises. Chest and front-of-shoulder length — a doorway stretch, 30 seconds a side — addresses this directly.
The blade's upward rotation, third. Wall slides, done slowly with the forearms kept in contact, work it as a movement rather than a stretch.
And control, throughout. Getting an arm overhead passively and being able to hold something there are different capacities. Range you can only be pulled into is range you do not own.
What the evidence supports
That range at the shoulder responds to training: reasonably solid. The general stretching literature applies — Bandy and Irion found 30 seconds beat 15 and matched 60 over six weeks in Physical Therapy in 1994, and Thomas and colleagues found total accumulated weekly time the dominant variable in the International Journal of Sports Medicine in 2018. Both were hamstring studies, which is an honest caveat: shoulder-specific dose research is much thinner.
That overhead range predicts injury: weak. Moran and colleagues' 2017 systematic review in the British Journal of Sports Medicine found Functional Movement Screen composite scores predict injury poorly, and Bahr argued in the same journal in 2016 that screening tests generally lack the discriminative power to identify individuals who will get hurt.
That restoring it resolves shoulder pain: not established. The shoulder is the region where the confident story has fared worst. Beard and colleagues' CSAW trial in The Lancet in 2018 compared subacromial decompression surgery against placebo surgery and found no meaningful advantage. Ratcliffe and colleagues, reviewing whether shoulder pain relates to scapular orientation in the British Journal of Sports Medicine in 2014, found no consistent relationship. Improving your overhead reach is a good goal on its own terms and it is not a treatment.
Shoulder pain that wakes you at night, arm weakness, numbness or pins and needles down the arm, a shoulder that suddenly lost range, or pain after a fall needs a clinician who can examine you. Nothing here diagnoses anything and nothing here is treatment.
The dose, and the schedule
Most days, briefly, and mix movement with holds.
- Thoracic extension, over a rolled towel or the back of a chair, 60 seconds.
- Doorway pec stretch, 30 seconds each side.
- Wall slides, six slow repetitions with the forearms in contact.
- Overhead reach against the wall, the test position itself, held 20 seconds with the back flat. Training the shape is part of owning it.
Four minutes. Expect the wall test to move over three to six weeks if the limit is soft tissue, and to move much less if the limit is the shape of your upper back — some people carry a fixed thoracic curve that does not respond, which is worth knowing before deciding you have failed.
Measure the chain, and mind the gap
The shoulder mobility test is the built reading: the Apley scratch, one arm over the top and one underneath, measured in hand-widths, both sides kept apart. More than two hand-widths reads 10, one to two reads 30, under one reads 55, fingertips touching reads 75, overlapping fingers read 95.
Two honest limits. It measures three joints at once and cannot say which one is short — a stiff upper back closes the gap between the hands as surely as a stiff shoulder does, which is the same chain problem this page is about. And the smallest gap it can show between your sides is 20 to 25 points, while Limber calls 12 points an imbalance, so any one-band difference between sides is flagged. On a banded test that means re-measure, not finding.
Now the gap on our own side. In the app, the shoulders area is designed as shoulder reach at 60% and a wall angel at 40%. The wall angel is not built, so its weight drops out rather than counting as a zero and the reach carries that area on its own — the Range Score prints the shortfall in those terms, Shoulder reach 60% (now 100%); Wall angel 40% — not read, rather than renormalising in silence and calling the area complete. The thoracic spine, which this page has just named as the usual limit, is fed only by a seated rotation test that is not built either, so the battery has no reading at all for the part of the chain that most often stops you. Four of five stations is the ceiling, and every page that composes a score says so.
Take the shoulder reading three weeks apart under matched conditions; Limber's default re-test interval is 21 days, adjustable to 14 or 28. How to track flexibility progress sets out the method, realistic flexibility goals and timelines covers what to expect, and the cluster is goals.
Questions
Why can I not lift my arms straight overhead?
Usually one of three links: the shoulder joint, the shoulder blade's rotation on the ribs, or thoracic extension. The last is the one a desk day removes and the one nobody trains. Test it against a wall with the lower back pressed flat — most people find their arms stop well short once the back can no longer arch.
How do I test my overhead reach at home?
Stand with your back to a wall, get the head, upper back and tailbone touching, press the small of your back flat, then raise both arms overhead with the elbows straight. If the wrists only reach the wall by letting the back arch, that is your answer, and the point where the back lifted is your usable range.
Is arching my back to get overhead bad?
It is not dangerous in itself, and it hides the problem, which is the real cost. The lower back substituting makes the position look finished while the shoulder and upper back have stopped moving, so you keep training a range you already have. Under load — pressing something overhead — the substitution also puts the effort somewhere it was not intended.
How long does it take to improve overhead reach?
If soft tissue is the limit, expect measurable change over three to six weeks, which is where the general stretching trials find range changes. If a fixed thoracic curve is the limit, considerably less will change, and that is a fact about your spine rather than about your effort.
Will better overhead mobility stop my shoulder hurting?
There is no good evidence that it does. Shoulder pain has been an area where confident mechanical stories fared badly — a trial comparing subacromial decompression against placebo surgery found no meaningful advantage, and reviews find no consistent relationship between shoulder pain and scapular position. Improving the range is worth doing for its own sake; pain that wakes you at night or comes with arm weakness needs a clinician who can examine you.
Does Limber measure overhead reach?
Not directly. The battery's shoulder test is the Apley scratch, which reads the chain rather than pure overhead elevation. The wall angel test that would read the overhead position is not built, so the reach carries the whole shoulders area instead of the 60% it was designed for, and the thoracic spine has no reading at all. The site prints those gaps rather than averaging around them.