Optimize: Sleep
Falling asleep, staying asleep, and waking rested.
The guideline first-line for chronic insomnia, ahead of pills; largest and most durable effect (~19 min faster to sleep, held at 12 months).
Reduce-lever: 400 mg within ~4-6 h of bed costs 45-51 min of sleep; cutting caffeine earlier is one of the highest-value free moves (it does not touch REM).
A cool (about 20-25 C), dark, quiet room; each 10 dB of night road noise roughly 2.5x's badly disturbed sleep.
Morning daylight anchors a delayed clock earlier (ES 0.34) and lifts sleep quality; free.
Passive body heating: a warm bath 1-2 h before bed shortens time to fall asleep, pooled across 13 studies.
Reduce-lever: a nightcap speeds onset but cuts REM and fragments the second half of the night; less before bed sleeps better.
A circadian tool: strong for jet lag and a delayed clock, modest (~7 min) for ordinary insomnia; 0.3-0.5 mg works as well as 3 mg.
Meaningful sleep-quality gain in women with sleep problems (PSQI SMD -0.54).
Reduce-lever, and a narrow one: interactive device use in bed costs sleep; blue-blocking glasses and Night Shift mode are no-effect, so do not sell them as the fix.
For restless legs with low ferritin only: cut RLS severity ~3.8 points; not a general sleep aid.
Beats a matched comparison (ES 0.33-0.54) but no better than established sleep therapies.
Older adults fell asleep ~17 min faster, on weak evidence.
Small sleep gain across 5 trials (SMD -0.59); trials small and several industry-linked.
Acupoint therapies improved sleep across 95 insomnia trials; low-cost and self-administered.
Moderately better self-rated sleep; gentle and free, tradition-supported.
Sleep improved in 5 of 8 trials, likely via calmer pre-sleep thoughts.
Cost Free · $ · $$ · $$$ Effort Easy · Moderate · Hard Results In Same-day · Weeks · Months
Levers with real evidence for sleep. Tiers are the strength of that evidence, not our endorsement. An up arrow raises the goal; a down arrow is a reduce-lever.
Evidence strength
How confidently the research supports a claim. Strength describes the evidence, not our endorsement.