The Military Sleep Method and Its Missing Six Weeks

In 1942 the US Navy had a problem it could not train its way out of. Pilots at the Del Monte Naval Pre-Flight School near Monterey were making fatal errors, and the pattern pointed at exhaustion rather than incompetence. Tired men fly badly, and there was no shortage of reasons to be tired.

The Navy handed the problem to an unusual person. Lloyd «Bud» Winter was a sprint coach from San Jose State who had enlisted after Pearl Harbor, and his professional expertise was in getting athletes to stop straining, on the principle that a sprinter who tenses his jaw runs slower. He built the Navy a relaxation programme on the same logic, aimed at getting pilots to sleep on demand.

Nearly forty years later, in 1981, he wrote it up in a book called Relax and Win: Championship Performance. And in that book is the sentence that now circulates on every platform on earth: after six weeks of practice, 96 per cent of the pilots could fall asleep within two minutes, even with gunfire going off nearby.

What that sentence actually is

It is a coach’s recollection, published four decades after the events, in a trade paperback about athletic performance.

There is no study behind it. No control group. No measurement protocol. No definition of «asleep» beyond whatever Winter or an observer judged at the time. Nobody has produced the underlying records, and searching for them comes up empty: «military sleep method» returns zero results in PubMed titles and abstracts, and so does «Relax and Win».

None of which makes Winter a liar. Something clearly happened at Del Monte, and a man who spent his career watching athletes relax on cue is a reasonable witness to whether pilots were falling asleep. But a remembered figure with a suspiciously round 96 in it, recalled decades later, is an anecdote wearing the clothes of a statistic.

Verdict on «96 per cent of pilots fell asleep in two minutes»: H — Untested. Unverifiable rather than false. The document that would settle it does not exist.

Everybody keeps the 96. Nobody keeps the six weeks.

Winter’s claim was never that the technique works. It was that it worked after six weeks of daily practice, every day, in a residential military training environment, with instructors, under conditions where failing to sleep had consequences rather sharper than feeling groggy at your desk.

Watch what survives the trip to your feed.

What Winter wrote	What circulates
96% of pilots	96% of people
After six weeks of daily practice	Tonight, first try
In a Navy pre-flight school, with instructors	Alone, from a 40-second video
A coach’s recollection in a 1981 book	«Backed by the US military»

The result belongs to trained people. The promise is sold to untrained ones. It is the same move as quoting a concert pianist’s performance without mentioning the twenty years of scales.

The Navy did not invent this

Strip away the aircraft carriers and look at the instructions. Relax the muscles of the face, including the tongue and jaw. Drop the shoulders, let the arms hang. Exhale, relax the chest. Relax the legs, thighs first, then calves. Then clear the mind for ten seconds, using an image such as lying in a canoe on a still lake.

The first four steps are progressive muscle relaxation, and the Navy had nothing to do with inventing them. Edmund Jacobson published the method in the 1930s, working from the observation that anxiety and muscular tension travel together, so releasing one might release the other (Jacobson, 1938). Winter’s contribution was compression. He took a long clinical procedure and made it short enough to run in a cockpit.

That matters for the evidence question. The technique is not untested. Its main ingredient has been through a great many trials, under a different name.

An effect size big enough to be suspicious

Two meta-analyses of progressive muscle relaxation landed in 2026. Read together, they are a lesson in how to hold a positive result at arm’s length.

	Donato et al.	Li et al.
Trials / people	31 RCTs, 2,277	14 studies, 957
Effect on sleep quality	SMD −1.74	Hedges g −1.24
Heterogeneity (I²)	92.1%	85.5%
Adults 55+	not reported	no significant benefit
More sessions, better sleep?	not tested	no association (p = 0.21)

Take the headline number first (Donato et al., 2026). An effect size of −1.74 is enormous. Most behavioural interventions in this field land somewhere between 0.2 and 0.5, so this is roughly the difference between a nudge and a shove. When a non-drug intervention measured by self-report produces a number three times larger than its neighbours, the correct response is suspicion rather than excitement.

The reasons for suspicion are printed in the paper. Heterogeneity came in at 92.1 per cent. That statistic asks how much the individual trials disagree with each other, and 92 per cent means they disagree almost totally: the pooled average is the mean height of a group containing both toddlers and basketball players. Technically correct, descriptively useless.

Then there is who was doing the measuring. Nearly every outcome was self-reported by people who knew perfectly well whether they had spent twenty minutes relaxing their thighs. You cannot blind someone to that.

The second review is smaller and more careful (Li et al., 2026), and it found two things the larger one did not report. The benefit was significant in adults under 55 and not significant in adults 55 and over. And a meta-regression found no relationship at all between how many relaxation sessions people did and how much their sleep improved.

That last one sits awkwardly against the whole six-weeks story. If practice were doing the work, more sessions should mean more benefit. Here, they did not.

Verdict on «the relaxation part works»: P — Partial. A large body of randomised evidence, a very large apparent effect, and heterogeneity and self-report severe enough that the size of the number should not be taken at face value.

The one step that fights you

The final instruction is the problem: clear your mind completely for ten seconds, and push away any thought that appears.

Trying not to think is among the most reliably counterproductive instructions in psychology, and somebody has tested it in exactly this setting. Normal sleepers were told either to fall asleep as quickly as they could or to fall asleep whenever they liked, while listening to either soothing music or John Philip Sousa marches (Ansfield, Wegner & Bowser, 1996).

With gentle music, the people trying hard fell asleep faster. With the marches going, so with the mind under load, the people trying hard to fall asleep quickly took longer than the people who were not trying at all.

Nobody looks up a sleep technique on an easy night. They look it up at one in the morning with a busy head, which is the high-load condition, which is where urgency backfires.

A broader review of pre-sleep mental strategies reaches a compatible verdict, listing suppression among techniques whose effect is «either beneficial, negligible, or detrimental», depending on which one and who is using it (Lemyre et al., 2020).

Verdict on «clear your mind for ten seconds»: R — Refuted as reliable advice. Effortful mental control of sleep has been shown to slow sleep onset under load.

Why «fighter pilots» beats «31 trials»

Part of this method’s reach has nothing to do with sleep. It is the setting.

A technique attributed to combat pilots arrives pre-loaded with credibility that ordinary evidence cannot match. The military is the one institution most people assume has tested things properly, under pressure, on people who could not afford failure. «Navy pilots used this» sounds sturdier than «31 randomised trials found this», even though only one of those sentences describes a measurement.

There is a second effect, quieter and more persuasive. The wartime frame implies that your own trouble sleeping is trivial by comparison, so the technique must be more than enough for you. If it worked with artillery in the background, surely it works in a quiet bedroom. The reasoning feels airtight and does no work at all, because the pilots had six weeks of drill and you have a video.

Winter himself appears to have been straightforward about all of it. The claim sits in a book about coaching, stated as his experience, not dressed up as research. Everything that makes it sound like research was added later, by people who needed it to sound that way.

Two minutes against nineteen

Here is the ceiling. The best-supported insomnia treatment in existence, delivered over multiple sessions by a trained clinician, shortens time to fall asleep by about 19 minutes compared with inactive controls, across 20 randomised trials and 1,162 people (Trauer et al., 2015).

Nineteen minutes. That is what decades of clinical sleep research, applied by a professional over weeks, buys on this specific outcome.

«Two minutes, tonight» is a bid to beat that on the first attempt, unaided, using a claim about people who had drilled the technique daily for a month and a half.

So should you do it?

Probably, and the reasons are worth saying out loud rather than lost in the grading.

• Its core is a real method with a long trial record, not an invention of the internet.
• It is free, brief, and physically harmless.
• The historical origin is genuine, even though the famous number cannot be checked.
• Unlike most sleep advice, it asks you to do something specific with your body rather than to feel differently about your day.

The realistic version: an abbreviated form of a technique with decent evidence, whose famous statistic is unverifiable, whose stated training period is six weeks, whose final step is the one part likely to backfire, sold with a promise that outbids clinical treatment.

The shift: treat it as a skill with a training period rather than a switch. The original claim was six weeks of daily practice, and everyone drops that half.

First move: keep the muscle work, which is the part with 31 randomised trials behind it, and drop the deadline. Nothing in the evidence supports two minutes.

Second move: replace «clear your mind» with «hold one dull image». Effortful suppression under mental load has been shown to slow sleep onset; occupying attention with something uninteresting has not.

Third move: if this is a months-long problem rather than a bad week, the treatment with 1,162 people behind it exists and buys about 19 minutes.

The boring bottom line

A sprint coach taught wartime pilots to relax, and wrote down four decades later that 96 per cent of them could fall asleep in two minutes after six weeks of daily practice. There is no study, and two database searches for it return nothing. The internet kept the 96 per cent and the two minutes and deleted the six weeks.

What remains underneath is a shortened version of progressive muscle relaxation, which has 31 randomised trials behind it, an implausibly large pooled effect, and heterogeneity high enough to make that effect hard to trust. Plus one final instruction, to empty your mind on command, that has been experimentally shown to make things worse in exactly the state you would be using it.

Relax the jaw. Keep the canoe. Skip the stopwatch.

Sources

• Winter, L. C. (Bud) (1981). Relax and Win: Championship Performance. Book (not a peer-reviewed study). The sole source of the '96% of pilots in two minutes after six weeks' claim, written roughly four decades after the wartime programme it describes. 'Military sleep method' and 'Relax and Win' each return zero hits in PubMed titles and abstracts. Checked August 2026.
• Donato, A., Falcao, L., Nishizima, R., et al. (2026). Progressive muscle relaxation technique improves sleep quality and mental health: A systematic review and meta-analysis of randomized controlled trials. Journal of Psychosomatic Research. 31 RCTs, 2,277 patients. Sleep quality SMD -1.74 (95% CI -2.14 to -1.34) with I2 = 92.1%; PSQI MD -3.79. The effect size is unusually large and the heterogeneity is close to total. doi:10.1016/j.jpsychores.2026.112563
• Li, X., Chen, Y., Liu, J., et al. (2026). The effects of progressive muscle relaxation on sleep quality in adults: A systematic review and meta-analysis of randomized controlled trials. Frontiers in Public Health. 14 studies, 957 participants. Hedges g = -1.24 (95% CI -1.71 to -0.77), I2 = 85.5%. No significant benefit in adults 55 and over, and meta-regression found no association between number of sessions and sleep quality (p = 0.21). doi:10.3389/fpubh.2026.1906525
• Jacobson, E (1938). Progressive relaxation. The American Journal of the Medical Sciences. The origin of the muscle-relaxation procedure that forms the first four steps of the method. doi:10.1097/00000441-193811000-00037
• Ansfield, M. E., Wegner, D. M., & Bowser, R (1996). Ironic effects of sleep urgency. Behaviour Research and Therapy. 34(7), 523-531. Under low mental load, trying to fall asleep quickly worked. Under high mental load, sleep onset latency was greater for those trying than for those not trying. doi:10.1016/0005-7967(96)00031-9
• Lemyre, A., Belzile, F., Landry, M., Bastien, C. H., & Beaudoin, L. P (2020). Pre-sleep cognitive activity in adults: A systematic review. Sleep Medicine Reviews. Reviews mental imagery, hypnosis, paradoxical intention, suppression and distraction. Effects are 'either beneficial, negligible, or detrimental'. doi:10.1016/j.smrv.2019.101253
• Trauer, J. M., Qian, M. Y., Doyle, J. S., Rajaratnam, S. M. W., & Cunnington, D (2015). Cognitive behavioral therapy for chronic insomnia: A systematic review and meta-analysis. Annals of Internal Medicine. 20 RCTs, 1,162 participants. Sleep onset latency improved by 19.03 minutes: the benchmark any two-minute promise is bidding against. doi:10.7326/m14-2841