Sleep is the part of the week everything else depends on. The routine starts at sunset: dim light, an early last meal, no screens in the last hour, then reading and a slow downshift in bed. Mornings start with sunlight in the eyes before any screen.
This is my log, not medical advice.
Everything from sunset to the first hour of the day, on one line. The window never moves: lights out 23:30, up 08:00, weekends included.
Watch the sunset, outsidearound sunset
Dim, warm lightsafter sunset
Last meal2–3 h before bed
Blue + green blocking glasses
No liquidslast hour
No screenslast hour, first hour
Read, then NSDR in bed
Sleep23:30 – 08:00
Sunlight in the eyes15 min
Two short routines bracket the night. Times are for a 23:30 bedtime; sunset moves with the seasons.
Wind-down
Wake-up
The last 14 nights from Whoop: when I fell asleep and woke up, in Lisbon time. What matters is how little it moves from night to night, weekends included.
Asleep
± 38min
Spread of the time I fall asleep (SD)
Awake
± 11min
Spread of wake time (SD)
Weekend shift
+ 21min
Sat–Sun vs weekdays, middle of the night (4 weekend nights)
Dark and not too warm, the same window every night, and water early in the day.
Temperature
19–23°C
Cool-ish; I don't measure it closely
Light
0lux
Blackout blinds, no standby LEDs
Window
8:30h
23:30 – 08:00 in bed
Water
2.5–3L
Spread over the day, none in the last hour
Where each habit stands. Some rest on strong lab work, some on large cohorts, and a few are habits of mine with little behind them, labelled as such. Bars show strength of evidence.
Ordinary room light before bed delayed melatonin onset in 99% of people; 25 lux halved it on average, with a tenfold spread between people.
Lab studies; sensitivity varies a lot from person to person.
Sources: Gooley 2011 · Phillips 2019 · Brown 2022
In 60,977 people with wrist sensors, the most regular sleepers had 20–48% lower all-cause mortality than the least regular; regularity beat duration.
Observational, mean age 63, one week of data. If the week is short on sleep, weekend catch-up may help.
Sources: Windred 2024 · Sletten 2023 · Phillips 2017
Sleep-medicine consensus: adults need 7 hours or more on a regular basis. My window gives 8.5 h in bed.
Consensus, not a trial; the exact need varies.
Sources: Watson 2015
One night at 100 lux raised night-time heart rate and next-morning insulin resistance; in 88,905 people, the brightest nights were linked with higher mortality.
A one-night lab study and an observational cohort.
Sources: Mason 2022 · Windred 2024b
Heat during sleep fragments it and cuts deep and REM sleep; older adults slept most efficiently at 20–25 °C at home.
No proven optimum. My room sits around 19–23 °C; I don't monitor it.
Sources: Okamoto-Mizuno 2012 · Baniassadi 2023
In people with insomnia, amber lenses for 2 h before bed improved sleep; a meta-analysis found small, mixed benefits.
Unproven in healthy sleepers; dimming the lights may do the same job.
Sources: Shechter 2018 · Shechter 2020
Reading on a bright e-reader delayed melatonin and sleep compared with a printed book.
12 people, 4 h of maximum-brightness reading. The consensus panel found no agreement on screens for adults.
Sources: Chang 2015 · Hartstein 2024
5 minutes of cyclic sighing a day improved mood more than meditation; mindfulness improved sleep quality across 18 trials.
Sleep wasn't the main outcome; yoga nidra evidence is very low certainty. A habit that works for me.
Sources: Balban 2023 · Rusch 2019 · Singh 2026
Morning light advances the clock (strong lab evidence) and time outdoors is linked with easier waking. Eating within 3 h of bed is linked with reflux. Men need about 2.5 L of water a day, food included.
The 15-minute dose, the sunset, no screens after waking and no liquids in the last hour are my habits: no trial tests them.
Sources: Khalsa 2003 · Burns 2021 · Fujiwara 2005 · EFSA 2010
Every source links to PubMed or the original report. This is a personal log, not medical advice.
Want this in your own AI? Open it in Claude or ChatGPT, or copy the prompt below and send it to any agent. It connects to this protocol and can plan with it.
Prompt for your agent
Connect to Simone Staffa's training, nutrition and sleep protocols. Add the MCP server https://simonestaffa.com/mcp (read-only, no sign-in needed). If you can't add servers, read https://simonestaffa.com/training.md, https://simonestaffa.com/nutrition.md, https://simonestaffa.com/sleep.md instead. Then help me use them: start by planning today's evening and morning.
Or start with