why "just relax" fails and this doesn’t
You can’t relax on command — which is the whole reason "calm down" is useless advice. You go numb to tension that never lets up: a muscle can be tight for an hour and you’ve stopped registering it, until there’s something to compare it with.
Here the comparison is manufactured. A few seconds of obvious tension give the body a reference point, and against that, letting go becomes a noticeable event rather than an abstraction. Jacobson, the American physician who described the method in 1929, insisted on exactly this: first you learn to notice tension, and only then does it start to go.
Which leads somewhere worth knowing in advance. The first few times, the main result isn’t "I feel calm" — it’s "so my jaw has been clenched this whole time". That isn’t the practice failing. That is the practice.
how to do it
The order and the number of muscle groups differ between guides — some go upwards, some down, some use four groups and some sixteen. There’s no single canon, and that’s fine: the principle is one thing — tense it, then let it go — and the order can be whatever you like.
Here’s a workable version, going up.
- Lie or sit so your body isn’t holding anything up. Lying down is easier, though before bed you may simply fall asleep — if sleep isn’t the aim, sit.
- Tense your feet. Curl the toes, hold for five to seven seconds. Not to the point of pain — to the point where you clearly feel it. Keep breathing while you hold; there’s no need to hold your breath along with the muscle.
- Drop it. Not "release slowly" — just stop holding, all at once. Breathe out as you do. Stay there about ten seconds and notice the difference.
- Then upwards: calves, thighs, buttocks, stomach, hands into fists, forearms, shoulders to the ears, face — screw your eyes shut and scrunch it up.
- Go through the whole body, then lie still for a minute doing nothing.
The whole thing takes ten to fifteen minutes. The short version uses four groups — arms, face, torso, legs. It fits into five minutes and does the job on a night when that’s all you have in you.
Two details that change the result noticeably:
- Five to seven seconds, no longer. Thirty seconds of tension buys you cramp and nothing else.
- The letting go matters more than the tensing. Tension is only the run-up; the work happens in the ten seconds when you’re doing nothing.
thirty-one trials, and one large caveat
For a method of its age, this comes out surprisingly well.
A systematic review with meta-analysis published in 2026 pulled together thirty-one randomised trials and 2,277 participants. Sleep quality improved markedly, anxiety came down, quality of life went up. The effect on sleep was larger than the effect on anxiety — worth holding on to when you decide what you’re doing this for.
And now the caveat without which those numbers can’t be read. The heterogeneity across the trials is enormous: between 83 and 94 per cent depending on the outcome. In plain terms, the studies disagree with each other badly, and with spread like that an averaged effect size promises nothing precise. Re-run without the weakest trials, the results kept their direction — not their size.
So the honest version is this: the method works, and how much depends on who is doing it and how. On this data you cannot promise "so many points off your anxiety", and we are not going to.
who should be careful
Contraindications in the usual sense barely exist here — that’s the method’s strength. But there are three notes.
Fresh injuries and pain. There’s no point tensing something that hurts: skip that group. It makes almost no difference to the result.
A tendency to panic attacks is worth knowing about first. The practice asks you to watch your body closely, and when you’re prone to panic it’s exactly that watching that drives the anxiety up — the mechanics are in panic attack. This isn’t a ban: plenty of people find it helps. But if halfway through you’re more anxious than when you started, it isn’t your tool today, and going back to breathing is the sensible move.
When you need more than this. If anxiety has held for months, or sleep has been coming apart for longer than a month, relaxation stays a support. What does answer it, and when it’s time to see someone, is in anxiety that won’t let go and can’t sleep.
audio: do you need it
There are a lot of recordings of this, and people look for them by the same words every time — free, no ads, guided, twenty minutes. Worth deciding about in advance.
A voice in your headphones is useful for the first few sessions: you don’t have to hold the order in your head or watch the clock. After that it develops a drawback — the practice gets attached to the recording, and without it you don’t start. And the point of the method is being able to take tension off wherever you are: in a queue, before a conversation, in an aircraft seat.
Easiest way through it: the first few times with a recording, then from memory.
when to reach for which
The two tools sit next to each other and reach different things.
Breathing is faster. A long exhale brings the pulse down in minutes, and of the two it’s the one that works in an acute moment — the catalogue is in breathing techniques.
Relaxation reaches further into the body. Shoulders and a jaw stay set like stone long after the pulse has settled. Breathing doesn’t get down there; tensing and dropping does.
Which gives a simple rule. If something has just hit you, breathe. A heavy evening, a body like wood, too early for bed — this.
They stack, too: run the sequence through the body with a long-exhale rhythm playing underneath. Then every letting-go lands on an exhale.
short answers
How many times a week?
No reliable answer: the review looked at how long the course ran, not how often people practised, and it worked across the range of course lengths. Practically: the short version daily gives more than the full one at weekends.
Does it help with a diagnosed anxiety disorder?
Anxiety did come down in that review, but it wasn’t measured in every trial — the trials were picked for measuring sleep, and anxiety came along for the ride. The ground under the sleep numbers is wider, and no precise figure for anxiety holds. As a daily support, yes. Once there’s a diagnosis, the treating is done by therapy and medication, and relaxation attaches to those.
How soon does anything happen?
The sense of a muscle letting go turns up in the first session for most people. For some it doesn’t. The skill of catching tension before it piles up takes weeks.
Can I do it before bed?
Yes — that’s its best use: the evidence on sleep is stronger than the evidence on anxiety. Lying down, with no ambition of reaching the end; falling asleep halfway is a normal outcome here.
What if I can’t feel the difference?
Usually the tension is too gentle or the release too gradual. Try tensing more distinctly and dropping it more sharply — the whole thing runs on contrast.
Is this the same as autogenic training?
No. Autogenic training works by suggestion — you repeat that an arm is heavy and warm until it feels that way. Here nothing is suggested: you make real tension and then stop making it. The two get mixed up because both end in a heavy, warm feeling in the limbs. You may also see this called JPMR — Jacobson’s progressive muscle relaxation, same thing.
Is this the same as the "military method" for falling asleep?
Essentially yes: the same relaxation sequence, with a countdown added and promises nobody can support. Taken apart in how to fall asleep fast.
Does it help with panic attacks?
Not during an attack — there you want breathing with a long exhale. As a daily practice between attacks it suits, as long as watching your body doesn’t make the anxiety worse.