A great deal gets written about sleep, and the advice contradicts itself more often than you'd expect. Some is well studied; some has simply established itself through repetition.

An attempt to sort it out.

What's reasonably well supported

A consistent wake time. The single best-supported factor and the most frequently ignored. The body clock anchors strongly to the moment of waking and the light that follows. Getting up at the same time every day, weekends included, stabilises the rhythm considerably.

Interestingly, holding the wake time constant is more effective than holding bedtime constant. Evening sleepiness follows from it, not the other way round.

Morning light. Bright light shortly after waking anchors the rhythm. Outdoor daylight is orders of magnitude brighter than indoor lighting, even under cloud. Twenty minutes outside does considerably more than any lamp.

A cool bedroom. Core body temperature has to fall for sleep onset. An overheated room prevents that. Recommendations usually sit between 16 and 19 degrees, which strikes most people as cold at first.

Cognitive behavioural therapy for insomnia. For persistent sleep problems this is the treatment with the strongest evidence base, better than medication over the long run. It includes sleep restriction and stimulus control. It's recommended far too rarely, partly because it takes more effort than a prescription.

Caffeine timing. The half-life of caffeine sits around five hours for most people, with considerable individual variation. A coffee at four in the afternoon is still measurably present at midnight. Many people sleep badly without realising the afternoon coffee is involved.

What's less clear than claimed

Blue light in the evening. The most overstated piece of advice. That light influences melatonin is not in dispute. The question is dose. A phone screen at normal brightness delivers considerably less light than a ceiling fixture, and several studies of blue-light filters find only small effects on sleep quality.

More likely the problem isn't the light but the content — what you're doing on the device, how activating it is, and that you stay up longer than intended.

Eight hours for everyone. Sleep need is individual and normally distributed. Some people need seven, some nine. The widespread fixation on a single number generates anxiety in many people, which then worsens sleep further.

Alcohol as a sleep aid. It helps with falling asleep and clearly degrades the second half of the night — less REM sleep, more waking. This is well established and still routinely recommended as a home remedy.

The tracker effect

A phenomenon that now has a name: orthosomnia. People who sleep worse because of engaging with their sleep data.

The mechanism is obvious. Seeing a poor score in the morning generates worry, and worry about sleep is among the most reliable triggers of sleep problems.

On top of that, the accuracy of consumer devices at identifying sleep stages is limited. Total sleep duration is usually estimated reasonably; the breakdown into deep and REM sleep considerably less reliably.

So anyone unsettled by a number should know that the number carries substantial uncertainty.

When you're lying awake

The most practically important advice and the most counterintuitive: don't stay in bed.

The recommendation from behavioural therapy is to get up after about twenty minutes, go to another room, do something quiet under dim light, and return only when sleepy.

The reason is conditioning. Regularly lying awake in bed teaches an association between bed and wakefulness. Undoing that association is laborious, and avoiding it is the entire purpose of the rule.

Looking at the clock makes it worse. The calculation "four hours until the alarm" produces exactly the tension that prevents sleep.

What I'd leave out

A word on naps, since the advice on them is unusually contradictory.

Short naps, roughly ten to twenty minutes, generally improve alertness without much cost, and the research on this is reasonably consistent. Longer naps risk waking from deep sleep, which produces the groggy, disoriented state that makes people conclude naps do not work for them.

The more important consideration is timing. A nap in the early afternoon interacts far less with night-time sleep than one at five or six in the evening, which eats into the sleep pressure you need in order to fall asleep later.

And for anyone actively struggling with insomnia, the standard advice is to avoid napping altogether during treatment, precisely because sleep pressure is the tool being used. That is one of the few places where an intervention that is fine for most people is specifically counterproductive for the group most concerned about sleep — which is a decent illustration of why generic advice in this area is worth so little.

The elaborate evening routines circulating on social media. Ten steps before bed, supplements, special devices.

The problem isn't so much that they cause harm, but that they turn sleep into a project. And sleep is one of the few activities where effort is counterproductive.

Consistent wake time, get outside in the morning, cool room, no caffeine after early afternoon. That's the list. It's short, boring, and covers most of what's actually supported.