Sleep is where adaptation happens. The mortality nadir for adults sits around seven hours. Most people skip it. This article is the protocol for fixing it.
Duration Before Architecture
Before optimizing sleep stages, sleep latency, or any wearable metric, hit seven hours of time in bed for fourteen consecutive nights.
A meta-analysis of 1.5 million participants showed both short (<7) and long (>9) sleep increase mortality risk (Cappuccio FP, et al. Sleep. 2010). Compared to seven hours, five hours produces 4% higher mortality. Nine hours produces 21% higher. Ten hours produces 37% higher.
The seven-hour floor is not arbitrary. It is the empirical nadir.
Most people who think they have a "sleep quality" problem actually have a duration problem. Fix duration first. Most quality complaints resolve at the same time.
The Foundations That Have Evidence
Separating signal from noise on sleep advice matters. Most of the supplement-industry sleep stack has thin or no human-trial evidence. The behavioral foundations are well-supported.
Cool, dark, quiet. Room temperature between 60 and 67 °F (15-19 °C) supports the natural drop in core body temperature that initiates sleep. Black-out curtains or a sleep mask block ambient light that suppresses melatonin. Earplugs or a white noise machine block intermittent disruption.
Phone out of the bedroom. Not silent. Not face-down. Not on the nightstand. In another room. The presence of the phone is a cognitive load even when not in use, and the temptation to check it during normal nocturnal awakenings shortens total sleep.
Consistent wake time. More important than consistent bedtime. The circadian system anchors to wake time and morning light exposure. A consistent wake time, even on weekends, stabilizes the entire sleep schedule.
Caffeine cutoff at least 8 hours before bed. Caffeine has a half-life of 5 to 6 hours but a quarter-life of about 12. Afternoon coffee is the saboteur most people do not connect to their poor sleep. Morning-only is the cleanest rule.
Alcohol disrupts sleep architecture. Alcohol shortens sleep latency (you fall asleep faster) but suppresses REM and fragments the second half of the night. The data is consistent across decades. One drink with dinner is a different impact than three drinks before bed; both have measurable effects.
Light
Light is the strongest signal the circadian system uses. The protocol is simple.
Morning: Get bright light into your eyes within an hour of waking. Outdoor sun is best (10-30 minutes depending on latitude and weather). On dark mornings, a 10,000 lux light therapy lamp at 30 cm for 20 minutes is a usable substitute.
Evening: Dim ambient light starting two hours before bed. Switch to warm-color bulbs in the rooms you spend evening time in. Reduce screen brightness. The goal is signaling night to the brain.
Blue blockers: The evidence on blue-blocking glasses is mixed. The clearest effect is in people who genuinely cannot avoid screens late at night. They do not replace dimming the environment; they reduce one specific input.
Meal Timing
Eating most calories earlier in the day improves overnight glucose regulation and sleep quality.
The practical rule: stop eating 2 to 3 hours before bed when possible. Late dinners disrupt sleep architecture even when the sleep duration looks fine on a wearable.
One exception: if you are training hard and genuinely under-eating during the day, going to bed actively hungry can disrupt sleep more than a small protein-forward snack. Cottage cheese, Greek yogurt, casein. Context matters. The principle is finishing food earlier when you can. The exception is not a license to eat late routinely.
Pre-Sleep Down-Regulation
The body needs a transition from active to resting. This is the "coming down" piece most protocols skip.
A 30 to 60 minute pre-sleep window with low cognitive demand:
- Light stretching or mobility work
- Slow breathing (box breathing or 4-7-8 work)
- Reading something low-stakes (fiction, not work)
- Journaling: a short list of what was done today, what is pending tomorrow
- Conversation with a partner
- Warm shower or bath (the post-shower temperature drop helps initiate sleep)
The point is not the specific activity. The point is signaling to the nervous system that the day is finished.
Magnesium Glycinate
Magnesium deficiency is common, especially in low-carb populations and heavy trainers. Glycinate is the form best-supported for sleep.
Dose: 200 to 400 mg taken 30 to 60 minutes before bed. Start at the low end.
It is not a sleep aid. It supports the systems that produce good sleep when the rest of the protocol is in place.
Citrate works similarly but has a laxative effect at higher doses. Oxide is poorly absorbed.
If you are not magnesium-deficient and your sleep is good, supplementing will not produce noticeable changes. Test it for two weeks. If you notice nothing, stop.
Recovery Beyond Sleep
Sleep is the foundation of recovery, not the entirety of it.
Active recovery days. Light movement (walking, mobility work, easy bike) on training-rest days promotes circulation and clears metabolic waste without adding stress.
Walks. Twenty to thirty minutes outdoors, not for fitness, just for nervous system regulation. The data on walking and stress recovery is older but consistent.
Sauna. 15 to 20 minutes at 80 to 100 °C, two to four times per week, is associated with reduced cardiovascular mortality and improved cardiovascular fitness markers (Laukkanen JA, et al. JAMA Internal Medicine. 2015). It also raises growth hormone acutely.
Cold exposure. The data is mixed. Cold water immersion immediately after resistance training appears to blunt some hypertrophic adaptation. Cold on rest days or hours after training is fine. The acute mood and alertness effects are real. Long-term metabolic claims are less well-supported than the supplement industry suggests.
Wearables
Wearables (Whoop, Oura, Apple Watch, Garmin) are useful for one thing: trend data over weeks.
They are unreliable for: nightly sleep stage breakdowns, exact sleep onset, individual REM duration. The accuracy on these specific metrics, when validated against polysomnography, is roughly 60 to 80% depending on device and night.
Use them for the trend. If your seven-day rolling average shifts, that is data. A single bad score is noise.
Avoid the trap of chasing the score. The wearable is a tool. The protocol is what produces the outcome.
The Two-Week Protocol
Pick a target sleep window with at least 7.5 hours of time in bed. Honor it for fourteen nights.
Week 1: Foundation.
- Wake time consistent every day (within 30 minutes)
- Phone out of the bedroom
- No food in the 2 to 3 hours before bed
- Caffeine before noon only
- Lights dimmed in the last hour before bed
Week 2: Build.
- Add morning light exposure within an hour of waking
- Add a 30-minute pre-sleep down-regulation window
- Add magnesium glycinate at night if symptoms warrant (cramps, restless sleep, low magnesium foods in your diet)
- Track only: bedtime, wake time, subjective rating 1-10
After two weeks, assess. Most people see noticeable change at the duration level alone. The architecture work (REM, deep sleep) tends to improve as a downstream effect.
What Changes When Sleep Is Fixed
Within 1 week: Daytime alertness improves. Cravings decrease. Mood stabilizes.
Within 2 to 4 weeks: Recovery from training accelerates. Cognitive output (focus, memory, decision-making) measurably improves. Hunger and satiety hormones (ghrelin, leptin) re-regulate.
Within 8 to 12 weeks: Body composition shifts. Sleep regulates the hormones that drive visceral fat storage. Adequate sleep alone does not produce fat loss, but inadequate sleep makes fat loss substantially harder.
Across years: Cardiovascular and metabolic risk markers improve. Cognitive trajectory in midlife is one of the strongest predictors of late-life cognitive function, and sleep is upstream of most of it.
Common Questions
I wake up at 3 AM and cannot fall back asleep. What now? This is usually a stress, alcohol, or late-meal issue. Look at the prior 6 hours. Did you drink? Did you eat late? Was there a difficult conversation or work problem you carried to bed? The 3 AM wake is rarely about 3 AM.
Is melatonin worth taking? For shift work or jet lag, yes, in low doses (0.3 to 1 mg). For chronic insomnia, the evidence is weaker than the supplement industry suggests, and habituation is common. Most people taking it long-term would do better fixing the foundations.
How much sleep do I actually need? Most adults function best on 7 to 9 hours. A small minority (genuinely 1-3% of the population) function on less. If you think you are in that minority, you probably are not. Test 7.5 hours for two weeks.
Should I nap? Short naps (10 to 20 minutes) early in the afternoon are generally fine and can improve alertness. Long naps (over an hour) or late-afternoon naps can disrupt nighttime sleep. The first rule is: do not use naps to compensate for chronically inadequate nighttime sleep.
What about supplements like ashwagandha, glycine, or apigenin? Some have small effects. Most have weaker evidence than magnesium. None replace fixing duration and the behavioral foundations. Start with the behaviors. Add supplements only after the foundations are in place and only one at a time so you can tell what is working.
Sleep is upstream of recovery, body composition, mood, cognition, and longevity. Fix duration first. Add the foundations. Layer the architecture work last.
If you want the room of people running this and the rest of the practice, the community is free.