Wall angels are usually sold as an exercise for rounded shoulders. They are a poor exercise and an excellent instrument, and taking them the second way is what makes them worth the minute.
The wall does not make the movement harder for its own sake. It removes the cheat. What leaves the wall first, and how far up the arms got before it did, is the reading — and the reading is more use than the repetitions.
The short answer
- Set up with the back of the head, the upper back and the hips on the wall, feet a step out, arms in a goalpost.
- Slide the arms overhead keeping the backs of the hands as close to the wall as you can, ribs down.
- What leaves the wall first is the finding: the hands, the wrists, or the lower back.
- The scored version is test 7 of Limber's seven-test battery, and it is not built — not here and not in the app.
- The movement does exist in the app's catalogue as M031 Wall Slides, a dynamic move against a wall. The move is not the test.
How to do wall angels
Five points, and the first two decide the rest.
- Stand with your back, hips and head against a wall, feet a step out from the skirting.
- Flatten the lower back toward the wall so a hand will not slide behind your waist. Ribs down.
- Bring the arms into a goalpost — elbows at shoulder height, forearms vertical, backs of the hands toward the wall.
- Slide the arms up overhead, keeping them as close to the wall as you can, moving slowly through the range.
- Lower to the start with control, keeping the ribs down. Six to twelve slow repetitions, one breath cycle each.
Ease off at any pinch in the front of the shoulder. That sensation is not a stretch, and pushing through it does not buy range.
Two habits ruin the reading. The first is letting the lower back arch, which turns a shoulder test into a lumbar one and produces a full-looking slide that measured nothing. The second is peeling the hands off the wall early and carrying on as though they were still there. Both feel like success from the inside, which is exactly the problem the wall exists to solve.
The wall is the instrument
Overhead is three links: the shoulder, the shoulder blade and the upper back. If the third does not extend, the blade stays forward, the shoulder runs out of room early, and the lower back arches to finish the movement. Standing in a room you never notice. Against a wall you cannot avoid noticing, because the spine has nowhere to go.
That is why a wall angel is unflattering and why it is informative. It is the same movement anyone can do badly, with the substitution removed.
Take it as a measurement first. Set up, slide once, slowly, and note where the contact broke. Then do the repetitions if you want them. Doing it the other way round — ten fast slides, no observation — throws away the only part of this that carries information.
What leaves the wall first
Three common results, and they point at different links.
The lower back lifts. The commonest reading, and the one that usually means the upper back is not extending. The arms borrowed the range from the lumbar spine. This is a thoracic finding wearing a shoulder costume.
The wrists or forearms lift while the hands stay near. Usually the shoulder itself, most often external rotation, running out partway through the arc. The elbows drift forward as the arms rise.
The hands leave immediately, from the goalpost. The starting position is already at the limit. That is a low reading and a useful one, because it means the first work is at shoulder height rather than overhead.
Record which one happened and roughly how far the arms got. Those two facts, written down with a date, are a better record than a repetition count.
Progressions, if the wall wins
If nothing stays in contact, reduce the demand rather than the standard.
- Bend the knees and let the pelvis tuck. It flattens the lower back and takes the lumbar spine out of the argument.
- Walk the feet further out from the wall, which does the same thing with less effort.
- Shorten the range. Slide only as far as full contact holds, then come back. Half a slide with contact is a reading; a full slide without it is not.
- Start from the floor. Lying on your back, the floor plays the wall and you can feel contact through a larger surface.
- Slow down. Speed hides the moment the contact breaks, and that moment is the whole measurement.
Progress looks like the same contact held through more of the arc, at the same tempo, three weeks later. It does not look like more repetitions.
The test that is not built
Now the part that goes against us, in full.
The wall angel is test 7 of Limber's seven-test battery, and it is not built — not on this site and not in the app. It is designed, specified and unimplemented. It has four bands in the design: hands leave the wall reads 20, wrists lift reads 45, full contact but strained reads 70, and easy full contact reads 90. Those are design values. You cannot score them today, here or in the app, and this page is not going to present them as though you can.
It matters because of what the design gave it. The wall angel is designed to be 40 per cent of the shoulders area, with the Apley scratch reach supplying the other 60 per cent. Because it is unbuilt, the weight drops out rather than counting as a zero, and the shoulder reach currently carries the shoulders area on its own, at 100 per cent. The site prints that shortfall rather than quietly renormalising and calling the area complete. The Range Score names the test that is missing.
The same is true one station along: the Spine group is fed only by a seated thoracic rotation test that is not built either. So the segment this movement is best at exposing has no reading at all in the battery, and four of five stations is today's ceiling. The overhead mobility test sets out what each protocol can and cannot say.
The move that does exist
Separate thing, same wall. M031 Wall Slides is in the app's move catalogue: a dynamic move at level 2, 45 seconds by default, against a wall, cleared for warm-ups, easing off at any pinch in the front of the shoulder. Its cues are the ones above — keep the ribs down, hands stay near the wall, move slow through the range.
So the app can give you the movement in a session. It cannot give you a score for it. That distinction is the whole of this section: M031 is a move, not a reading, and a session that includes it has not measured anything.
What the evidence says
Thin, and worth stating before the enthusiasm.
No trial of wall angels exists. Nobody has randomised anyone to this movement and measured an outcome. Everything written about it, here included, is reasoning from what the position demands. Grade: absent for the drill itself.
What does exist is a finding that competes with it. Andersen and colleagues, in Pain in 2011, found that two minutes a day of progressive resistance training reduced neck and shoulder pain in office workers — and no stretching or mobility routine has a randomised trial with a clinical outcome like it. If the reason you are reading this is an aching shoulder at a desk, the better-supported answer is a small daily dose of resistance work, not a slide up a wall. Grade: moderate for that general strength finding, which is stronger than anything supporting the drill on this page.
And the explanation usually attached to wall angels does not hold up either. Ratcliffe and colleagues, in the British Journal of Sports Medicine in 2014, found no consistent relationship between scapular orientation and shoulder pain. So "your blades sit wrong, correct them here" is not a claim this page will make. Scapular control takes that literature apart properly.
Pain that persists, wakes you at night, radiates into a limb, comes with numbness or weakness, or followed a fall or an injury is a matter for a clinician who can examine you. Nothing here diagnoses anything and nothing here is treatment.
Where to take a real reading
Since the scored wall angel does not exist, the reading you can actually take is the other shoulder test. The shoulder mobility test is the Apley scratch: five bands, both sides, a 0–100 score off the app's own anchors, nothing stored and nothing sent.
Use the wall angel alongside it as an unscored self-check. Note which part left the wall and how far the arms got, write the date next to it, and repeat in three weeks under the same conditions. That is a record even without a number attached, and it reads the substitution the scored test cannot see.
The rest of the cluster is shoulders and upper back.
Questions
What is a wall angel meant to do?
Two jobs, and the second is the useful one. As a movement it takes the shoulder through the overhead arc under mild demand, which is a reasonable warm-up. As a check it shows you which link of the chain runs out first, because the wall stops the lower back from finishing the movement on the shoulder's behalf.
How many repetitions should I do?
Six to twelve slow ones, a single breath cycle each, is the shape the app uses for dynamic moves, and M031 Wall Slides runs 45 seconds by default. Adding repetitions beyond that mostly adds fatigue. The variable worth increasing is how much of the arc you can travel with contact held, not how many times you travel it.
My lower back keeps lifting off the wall. What now?
That is the reading rather than a mistake. Bend your knees, walk the feet further out, and shorten the slide until full contact holds, then rebuild the range from there. It usually points at an upper back that is not extending, so thoracic extension work is the sensible companion.
Does Limber score my wall angel?
No. The wall angel is the seventh test in the design and it is not built, on this site or in the app. It is designed with four bands and 40 per cent of the shoulders area, none of which you can use today, so the shoulder reach carries that area alone and the score page says so.
Are wall angels better than stretching my chest?
Unknown, and probably not the right comparison. There is no trial of either against the other for this purpose. The strongest evidence in this neighbourhood is for a small daily dose of resistance training reducing neck and shoulder pain in office workers, which is a different intervention from both.
Can I do them on the floor instead?
Yes, and for many people it is the better starting point. Lying on your back gives you a larger contact surface and clearer feedback, and gravity holds you against it without effort. Move to the wall when full contact holds through most of the arc on the floor.