Every piece of nap advice ever written assumes you have two free hours.

You do not have two free hours. You have twenty minutes at 0130 before the tones drop, and maybe forty at 0400 if the city is merciful, and a supervisor who thinks a recliner is a moral failing. The advice was written for someone with a door that closes and a shift that behaves. So you skip the nap entirely, because a nap you cannot take in one piece has always been sold to you as no nap at all.

That premise is wrong, and there is now a trial that says so.

Two pieces, same result

Researchers ran an experimental reanalysis across a simulated sixteen-hour night shift, comparing a single 120-minute nap against splitting the same two hours into a 90-minute block and a 30-minute block placed later. Published in Scientific Reports in 2023, the finding is the one no shift worker expects: the split nap sustained alertness and performance as well as or better than the single long one.

As well as or better than. Not "an acceptable compromise when circumstances demand." Not "some benefit was retained." The divided nap held the line, and on some measures held it further into the shift.

Which makes a certain amount of sense once you stop thinking of sleep as a tank that fills. The 90-minute block is long enough to carry you through a full architecture — light sleep down into slow-wave, and back up — so it does the heavy repayment on adenosine, the molecule stacking up in your synapses since you woke. The 30-minute block later is not trying to repay anything. It is a top-up placed near the circadian trough, the 0300-to-0500 stretch where core body temperature bottoms out and human error rates peak, and it buys you alertness exactly where the night is trying to take it.

One long nap front-loads all your protection into hour six of a sixteen-hour night. The split spends it where the night is actually dangerous.

Where the 90 and the 30 go

The placement is most of the value, so be deliberate about it.

The 90 goes before or early. On a night shift, the strongest version is a prophylactic nap taken before you ever clock in — you arrive at the truck already having paid down some of the pressure rather than borrowing against a night you have not worked yet. If a pre-shift nap is impossible, the 90 goes as early in the shift as your service will allow. Early beats late for the long block, always.

The 30 goes into the trough. Somewhere in the 0200-to-0500 window, as close to your personal low point as the call volume permits. This is the nap that keeps your hands steady on the fourth IV of the night and keeps you on the road on the drive home.

On a 24, the same shape applies with more slack: the 90 in the evening when things typically quiet, the 30 held in reserve for the small hours. Do not stack both blocks together to reconstruct the 120. You would be undoing the thing that made it work.

Twenty to thirty, or a full ninety. Nothing between.

The short nap has a hard ceiling and it is not arbitrary.

Sleep inertia is the grogginess on waking, and its severity depends almost entirely on which stage you were in when the alarm went. Wake from light sleep and you are functional in a minute. Wake out of the middle of slow-wave sleep and you get twenty to thirty minutes of genuine cognitive impairment — slower reaction time, worse decisions, a fog that arrives precisely when you are being handed a patient.

Most people enter slow-wave sleep somewhere around the 30-minute mark. So the short nap is twenty to thirty minutes, timer set, and the long one is a full ninety, which carries you through the cycle and lets you surface out of light sleep on the other side. A 45-minute nap is the worst nap available to you. It is long enough to reach the deep water and too short to get you back out.

Set an actual alarm. Do not trust yourself to wake up on feel — the whole reason you are napping is that your judgment is degrading.

Give the short nap five minutes of grace on the other end, too. Wake, splash water, get bright light on your face, then take the handoff. Do not go straight from horizontal to a decision.

Why "I'll just push through" keeps failing

Because your instrument for measuring how impaired you are is the thing that is impaired.

Sleep-deprived people reliably rate their own performance as far better than it tests. The subjective sense of fatigue plateaus after a couple of days of restriction while the objective decline keeps going, straight down. This is not a character flaw. It is a specific, well-documented failure of self-assessment, and it means the person most confident they can push through is precisely the person you should not believe — including when that person is you at hour nineteen, feeling fine.

Meanwhile the countermeasures people actually reach for tend to be the ones with the least behind them. A 2024 systematic review of countermeasures for night-shift doctors and nurses found that napping and light exposure carry the best evidence, while most self-selected strategies — music being the named example — are not evidence-based at all. Everyone has a ritual. Most rituals are doing nothing but making the hour pass.

And in May 2025 the American Academy of Sleep Medicine published its first clinical practice guideline for Shift Work Disorder. For sleepiness on shift, it suggests a nap before the shift plus caffeine before or during it. That is the guideline-backed default now — not a hack, not a preference. A pre-shift nap and caffeine, in writing, from the body that sets the standard.

Which is worth sitting with for a second, because most of us were trained in the opposite direction. Napping before a shift felt like admitting weakness before the work even started. It turns out to be the single most defensible thing on the list.

The politics of the recliner

None of this is available to you if your service does not allow it, and plenty do not. The nap you cannot take because a supervisor believes sleeping on duty looks bad is not a personal failure of planning. It is a policy problem, and it is a safety problem wearing a policy costume, and the literature is now unambiguous enough to bring to a meeting. NASA settled this on flight decks decades ago. Ground services have been slow.

So bring the paper. The AASM guideline is a citation, not an opinion. It is harder to argue with a clinical practice guideline than with a tired medic.

And if you cannot win that fight this year, take what the shift gives you. Thirty in the recliner at 0300 is not nothing — it was half of a protocol that matched two full hours, and it is the half aimed at the most dangerous part of your night.

If you want the blocks placed against your actual rotation rather than a hypothetical one, the free plan builder puts a nap window into your schedule alongside your sleep anchor, caffeine cutoff, and light timing. And the Field Manual covers the rest of the on-shift picture — including the drive home, which is where all of this either pays off or does not.

Sources

This is not medical advice. Talk to your provider.

You were never going to get the two clean hours. You were always going to get ninety minutes and then, later, thirty. That was never the lesser version of the plan.

It was the plan.