TL;DR: Nap duration is not a matter of preference. In Brooks and Lack’s controlled trial of 24 adults restricted to roughly five hours of night sleep, a nap of exactly 10 minutes of sleep produced immediate improvements in every outcome measured, with some benefits still detectable 155 minutes later. Twenty minutes produced improvements that only emerged 35 minutes after waking and ran out at 125 minutes. Thirty minutes produced a period of measurably impaired alertness and performance immediately on waking, followed by improvements lasting to 155 minutes. Five minutes produced few benefits at all. Convert those windows into a return on the clock time you actually surrender and the 10-minute nap returns 15.50 minutes of improved function per minute invested, against 1.64 for the 20-minute nap and between 1.85 and 0.48 for the 30-minute nap depending on the task. The reason is sleep inertia: waking out of slow-wave sleep costs a 41 percent performance reduction against pre-nap levels, while waking out of stage 2 sleep leaves you performing like someone who never slept. Ten minutes is short enough to stay out of the expensive stage. Budget roughly five to six minutes of falling-asleep time on top, which is what NASA’s flight-deck data suggests a tired person actually needs.
“How long should a nap be” is one of those questions where the folk answer, the wellness answer and the experimental answer are three different things. The folk answer is twenty minutes. The wellness answer is usually ninety, to complete a full sleep cycle. The experimental answer, from the study that tested the durations head to head under controlled conditions, is ten.
That is worth taking seriously, because the difference between these answers is not stylistic. Choose wrong and you can spend half an hour to end up temporarily worse than if you had stayed awake.
The trial that tested the durations against each other
Most nap research compares napping to not napping. Brooks and Lack’s 2006 study in the journal Sleep did something more useful: it held everything else constant and varied only the duration.
Twenty-four healthy young adults, all good sleepers and none of them regular nappers, had their night sleep restricted to approximately five hours at home. They then came into a laboratory for an afternoon nap at 3:00 pm and three hours of post-nap testing. There were five conditions: a no-nap control, and naps of precisely 5, 10, 20 and 30 minutes of actual sleep. Outcomes included sleep latency, subjective sleepiness, fatigue, vigour and cognitive performance.
Table 1. Nap duration against benefit onset and benefit duration (verified data). All results as reported in Brooks A and Lack L, “A brief afternoon nap following nocturnal sleep restriction: which nap duration is most recuperative?”, Sleep, volume 29, issue 6, 2006, pages 831 to 840. Durations refer to minutes of actual sleep, not time in bed.
| Nap length | When benefits appear | How long benefits last | Immediate after-effect |
|---|---|---|---|
| No nap (control) | not applicable | not applicable | none |
| 5 minutes | few benefits reported versus control | not established | none reported |
| 10 minutes | immediately, across all outcome measures | some benefits to 155 minutes | none reported |
| 20 minutes | 35 minutes after the nap | to 125 minutes after the nap | none reported |
| 30 minutes | after an initial impaired period | to 155 minutes after the nap | impaired alertness and performance, indicative of sleep inertia |
The authors’ own conclusion was that the 10-minute nap was overall the most effective duration tested, and they noted that something physiologically significant appears to happen within the first ten minutes of sleep that accounts for the benefit.
What each nap actually costs you
The table above is the published finding. What it does not do is price the durations against each other, because “benefits last to 125 minutes” is not comparable to “benefits last to 155 minutes” when one of them does not start for the first 35.
So here is the arithmetic, which is worth doing because it changes the ranking.
Table 2. Alertness return per minute of clock time surrendered (CEOtudent editorial framework). Onset and end-of-benefit figures are as published by Brooks and Lack, 2006. The benefit window, clock cost and return columns are computed by CEOtudent from those published figures and are not published statistics. Clock cost is defined as nap duration plus any delay before benefits begin, that being the total period during which you are not better off. The 30-minute row is split because the published onset for delayed benefit is task-dependent, reported at 35 minutes for some tasks and up to 95 minutes for others.
| Nap length | Benefit window | Clock cost | Return per minute | Arithmetic |
|---|---|---|---|---|
| 5 minutes | no measurable window reported | 5 min | approximately 0 | not computable |
| 10 minutes | 155 min (0 to 155) | 10 min | 15.50 | 155 / 10 |
| 20 minutes | 90 min (35 to 125) | 55 min | 1.64 | 90 / 55 |
| 30 minutes, faster tasks | 120 min (35 to 155) | 65 min | 1.85 | 120 / 65 |
| 30 minutes, slower tasks | 60 min (95 to 155) | 125 min | 0.48 | 60 / 125 |
The 10-minute nap does not win narrowly. It wins by roughly an order of magnitude, and it wins for a reason that is structural rather than lucky: it is the only duration in the set that has no delay before the benefit starts. Every minute of delay is charged twice, once as time you have spent and once as time you are not yet better.
Note also the bottom row. On the slower-recovering tasks, a 30-minute nap returns less than a minute of improved function for each minute of clock time it consumes. That is a losing trade on any given afternoon, and it is the specific scenario people are describing when they say naps make them feel worse.
Why the extra twenty minutes costs so much
The mechanism is sleep inertia, and the current review literature is unusually blunt about its size.
Hilditch and McHill’s 2019 review in Nature and Science of Sleep collects the comparisons. Performance on an addition test immediately after waking was significantly more impaired than after an entire night of sleep deprivation. In another comparison, participants waking from sleep performed equivalently to or worse than they did after 40 hours of continuous wakefulness. This is not grogginess in the colloquial sense. It is a measurable cognitive deficit that happens to be brief.
How brief depends on what you were doing when you woke up. Performance typically returns to baseline within 30 minutes of awakening, with some studies showing recovery as soon as 15 minutes. But the asymptotic pattern of dissipation suggests full recovery takes at least an hour, and one rigorous study found addition-task performance took up to 3.5 hours to fully dissipate.
The single most decision-relevant finding in the review is about sleep stage. Participants waking from slow-wave sleep showed a 41 percent reduction in performance compared with their own pre-nap level. Participants waking from stage 2 sleep performed similarly to people who had been awake the whole time.
That is the entire case for the short nap in one sentence. The cost is not the sleeping. The cost is the stage you have to climb out of.
Two other findings from the same review are worth carrying. Sleep inertia is worst during the biological night, near the circadian low in core body temperature, and appeared attenuated around 15:00, which is when the Brooks and Lack naps were taken and roughly when most people would take an afternoon nap anyway. And under chronic sleep restriction of 5.6 hours per 24-hour day, participants showed a 10 percent worsening of performance immediately on awakening and had still not reached baseline 70 minutes later, against 8-hour controls. Sleep debt does not just make you want the nap more. It makes the exit from it more expensive, which is one reason structuring recovery deliberately rather than reactively matters more the more depleted you are.
What happens when you plan a nap in a real operation
Laboratory durations are clean. Real napping has a problem the lab controls away: you cannot decide to sleep for exactly ten minutes, because falling asleep takes time you do not control.
The best public data on this comes from an unusual source. In 1994 NASA published Technical Memorandum 108839, reporting a study run with the Federal Aviation Administration on planned cockpit rest in long-haul flight operations. Twelve crew members in a Rest Group were given a planned 40-minute rest period during the low-workload cruise portion of the flight. Nine crew members in a No-Rest Group had a 40-minute control period in which they continued usual flight activities. Both groups wore continuous brain-wave and eye-movement recorders and performed a reaction-time vigilance task.
Table 3. NASA planned cockpit rest, measured outcomes (verified data). All figures as reported in Rosekind MR, Graeber RC, Dinges DF, Connell LJ, Rountree MS, Spinweber CL and Gillen KA, “Crew Factors in Flight Operations IX: Effects of Planned Cockpit Rest on Crew Performance and Alertness in Long-Haul Operations”, NASA Technical Memorandum 108839, September 1994.
| Measure | Rest Group (n=12) | No-Rest Group (n=9) |
|---|---|---|
| Slept during the opportunity | on 93% of opportunities | not applicable |
| Time taken to fall asleep | 5.6 minutes average | not applicable |
| Actual sleep obtained | 25.8 minutes, about 64% of the 40-minute window | not applicable |
| Sleep composition | 30% stage 1, 62% stage 2, 8% slow-wave, 0% REM | not applicable |
| Median reaction time | reference | 10% to 16% slower |
| In-flight vigilance lapses | 81 total | 124 total |
| Physiological sleepiness events (transformed average) | 2.90 | 6.37 (p = 0.02) |
| Sleepiness events during descent and landing | none after top of descent | 22 |
| Uncommanded sleep episodes during the period | not applicable | 4 of 9 crew (44%), 5 episodes, from a couple of minutes to over 12 minutes |
Three things in that table deserve attention.
First, the yield. A 40-minute window produced 25.8 minutes of sleep, or 64.5 percent. If you want ten minutes of sleep, ten minutes in a chair will not reliably give it to you.
Second, the architecture. Only 8 percent of the sleep obtained was slow-wave, and there was no REM at all. This is exactly the profile that keeps sleep inertia cheap, and it happened naturally because the window was short.
Third, and most striking, the control group. Four of the nine pilots who were not given a rest period fell asleep anyway, involuntarily, during their control period, with one episode running over 12 minutes. These were people who knew they were in a fatigue study, were wired to continuous monitors, and had two NASA researchers on the flight deck. The report notes this was, to the authors’ knowledge, the first physiological documentation of unplanned sleep during long-haul flight operations.
The choice, in other words, was never between napping and not napping. It was between a nap you scheduled and a nap that happened to you.
A widely repeated NASA statistic that is not in the NASA report
If you have read anything about napping, you have probably encountered the claim that the NASA study found a 34 percent improvement in performance and a 54 percent improvement in alertness. Those two numbers appear in an enormous number of articles, usually attributed directly to NASA.
They do not appear in NASA Technical Memorandum 108839. A full-text search of the report returns the strings “34%” and “54%” only in unrelated contexts: a breakdown of microevent types, a sleep-stage percentage, and a duration distribution. The report’s own headline performance findings are the ones in Table 3 above: median reaction times 10 to 16 percent slower in the No-Rest Group, more than twice the physiological sleepiness events, and 22 sleepiness events during descent and landing against none in the Rest Group.
This matters beyond pedantry. The 34 and 54 percent figures are more dramatic than the documented ones and are almost always quoted without the study design attached, which is how a specific finding about sleep-deprived long-haul pilots on night flights became a general promise about afternoon naps at a desk. The documented findings are strong enough. They just describe a narrower thing. The same substitution shows up across performance advice, which is why checking what the research actually shows before restructuring a day around it is usually the higher-yield move.
How to actually run a nap
The evidence converges on a fairly narrow protocol.
Target ten minutes of sleep, and budget separately for falling asleep. The 10-minute condition is the only one in the trial that pays out immediately, and the NASA data suggests a tired person takes around 5.6 minutes to drop off. A 15 to 16 minute window is therefore the practical equivalent of the winning laboratory condition. Setting a 10-minute alarm and lying down at minute zero is targeting roughly a 4-minute nap, which the trial suggests is close to the duration that did almost nothing.
Take it in the early-to-mid afternoon. The trial’s naps were at 3:00 pm, and the inertia review reports that sleep inertia effects appeared attenuated around 15:00 and are worst near the circadian low during the biological night. Afternoon is not a folk preference here; it is where the measured effects were obtained. If you do not know where your own trough sits, a chronotype self-test will place it more usefully than guessing.
If you have to nap longer, plan for the exit, not just the entry. A 30-minute nap is not wrong in every context. It is wrong when you need to perform within the next half hour. If your next commitment is two hours away, the 30-minute nap’s benefit window running to 155 minutes is genuinely useful. Match the duration to when you need to be sharp, not to how tired you feel.
Consider caffeine before, not after. The review reports that caffeine taken before a short nap of around 20 minutes alleviates the symptoms of sleep inertia following it. Caffeine taken after waking truncates the duration of inertia but leaves the initial, most severe period unaffected. The ordering is the entire effect.
Do not use naps to paper over a sleep debt. Under chronic restriction of 5.6 hours a night, waking performance was 10 percent worse immediately and had not returned to baseline at 70 minutes. Napping into a structural deficit gets progressively less efficient, which is the argument for treating sleep as infrastructure rather than as a variable to be squeezed. If you are tracking this with a device, be careful about which numbers deserve the weight you give them, a question covered in the guide to which wearable metrics actually mean anything.
Treat the nap as a scheduled block, not a reward. The NASA finding that 44 percent of unrested pilots fell asleep involuntarily is the operational version of the point. Attention does not negotiate. A planned ten minutes is cheaper than an unplanned twelve at the wrong moment, and it slots naturally into a day already organised around the body’s ninety-minute energy cycles.
This is the CEO half and the student half of the same decision. The CEO half is scheduling recovery as a capacity input with a known return rather than taking it when depleted. The student half is being willing to discover that the duration you have used for years is the one the controlled trial ranked below a nap a third of its length, and changing it.
Frequently asked questions
What is the ideal nap length?
In the only trial that compared the durations head to head under controlled conditions, ten minutes of actual sleep was the most effective. It produced immediate improvements across every outcome measured, with some benefits still present 155 minutes later, and it was the only duration tested that produced no delay before benefits began.
Why do I feel worse after a longer nap?
Sleep inertia. Waking from slow-wave sleep is associated with a 41 percent reduction in performance against your own pre-nap level, whereas waking from lighter stage 2 sleep leaves performance similar to having stayed awake. Longer naps make slow-wave sleep more likely, so they are more likely to cost you on the way out.
How long does that grogginess last?
Performance typically returns to baseline within 30 minutes of waking, and some studies show recovery as soon as 15 minutes. The pattern of dissipation, however, suggests full recovery takes at least an hour, and one rigorous study found one cognitive task took up to 3.5 hours.
Is a 20-minute nap better than a 10-minute nap?
Not in this trial. The 20-minute nap’s improvements only emerged 35 minutes after the nap and ended at 125 minutes, against immediate onset and a 155-minute tail for the 10-minute nap. On the derived return calculation in Table 2, the 10-minute nap returns 15.50 minutes of improved function per minute of clock time surrendered, against 1.64 for the 20-minute nap.
How much time should I actually block out?
Around 15 to 16 minutes. NASA’s flight-deck data recorded an average of 5.6 minutes to fall asleep among tired crew, and pilots given a 40-minute window obtained 25.8 minutes of sleep, a yield of about 64 percent. A 10-minute alarm from the moment you lie down is targeting a much shorter nap than you intend.
Did NASA really find that napping improves performance by 34 percent?
Those figures do not appear in NASA Technical Memorandum 108839, the report they are usually attributed to. What that report documents is that the group denied a rest period had median reaction times 10 to 16 percent slower, more than twice the physiological sleepiness events, and 22 such events during descent and landing against none in the rested group.
Does a five-minute nap do anything?
In this trial, very little. The 5-minute condition produced few benefits in comparison with the no-nap control, which is part of why the authors concluded that something meaningful happens in the first ten minutes of sleep specifically.
Sources
Brooks A and Lack L, A brief afternoon nap following nocturnal sleep restriction: which nap duration is most recuperative?, Sleep, volume 29, issue 6, 2006, pages 831 to 840.
Rosekind MR, Graeber RC, Dinges DF, Connell LJ, Rountree MS, Spinweber CL and Gillen KA, Crew Factors in Flight Operations IX: Effects of Planned Cockpit Rest on Crew Performance and Alertness in Long-Haul Operations, NASA Technical Memorandum 108839, National Aeronautics and Space Administration Ames Research Center, September 1994.
Hilditch CJ and McHill AW, Sleep inertia: current insights, Nature and Science of Sleep, volume 11, 2019, pages 155 to 165.
Lovato N and Lack L, The effects of napping on cognitive functioning, Progress in Brain Research, volume 185, 2010, pages 155 to 166.
Milner CE and Cote KA, Benefits of napping in healthy adults: impact of nap length, time of day, age, and experience with napping, Journal of Sleep Research, volume 18, issue 2, June 2009, pages 272 to 281.
This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.
This post is also available in:
