Enforce a hard caffeine cutoff time

With a 5–7 hour half-life, any caffeine consumed after noon or 1pm is still meaningfully active at 10pm in most people.

Why it works

Caffeine blocks adenosine receptors without clearing the adenosine that has accumulated. As the caffeine is metabolized by CYP1A2 in the liver, the blocked adenosine floods the receptors — producing the crash. But meanwhile, sleep-architecture effects persist: residual caffeine at bedtime reduces slow-wave sleep (the most restorative phase) even when sleep onset is not delayed. This is why people who claim afternoon caffeine does not affect their sleep are often wrong — sleep onset is intact but sleep depth and quality are not.

How to do it

1. Calculate: your target sleep time minus 8–10 hours = latest caffeine intake for an average metabolizer.
2. For a 10pm bedtime target: last caffeine by 12–2pm.
3. If you are a known slow caffeine metabolizer (genetic CYP1A2 variant), move to 10–11am.
4. Track sleep quality (including depth and dream recall, not just onset) for 2 weeks after changing your cutoff.

Evidence

An RCT found that caffeine consumed 6 hours before bedtime significantly reduced total sleep time compared to placebo, even though subjects did not report sleep difficulty — demonstrating the architecture-impairment effect without subjective awareness. (rct)

Individual CYP1A2 variation is substantial — genetic fast metabolizers may have a much shorter effective cutoff (3–4 hours rather than 6+).

Sources

• Drake et al. (2013), caffeine effects on sleep taken 0, 3, or 6 hours before going to bed, Journal of Clinical Sleep Medicine

Common mistake

Judging caffeine’s sleep impact only by how hard it is to fall asleep, rather than by sleep depth and energy the following morning — the architecture disruption is invisible without measurement.