Nobody skips their long run to catch up on emails. Plenty of runners skip an hour of sleep to fit the long run in.

That trade gets made most weeks, usually without much thought, and it's almost always the wrong way round. Training is the stimulus. Sleep is when your body actually does something with it. Skip the sleep and you've paid for a session you won't fully absorb.

Here's what the research on sleep and running performance actually shows for endurance athletes, what a single short night costs, and how to work out whether your own sleep is the thing holding you back.

The study that used endurance athletes

Most sleep-and-sport headlines come from team sports. The one worth knowing about is Roberts and colleagues, published in Medicine and Science in Sports and Exercise in 2019, and it's the study that put endurance athletes in the lab.

Nine competitive cyclists and triathletes went through a counterbalanced crossover trial across three conditions, restricted, normal and extended sleep, seven days and nights each. The performance test was a fixed-work time trial, individualised so it took each athlete around an hour.

After three consecutive nights with an extra 90 minutes of sleep, the athletes completed those time trials on average around two minutes faster. Over an hour of racing, that's roughly 3%.

The mechanism the authors point to is the interesting part. Cumulative sleep changes the perceived exertion of a given intensity. More sleep meant a given effort felt more manageable, so a higher intensity could be tolerated. Their conclusion was blunt: endurance athletes should be sleeping more than eight hours a night.

Three nights. Ninety extra minutes each. Two minutes over an hour of racing. That's a better return than most things runners spend money on.

What one short night costs a runner

The deprivation research is less quotable and more directly relevant.

A 2023 systematic review and meta-analysis in the European Journal of Sport Science found sleep deprivation has a moderate negative effect on endurance performance, and that the effect worsens for efforts lasting more than 30 minutes. The longer your event, the more sleep was doing for you.

In one study using a 12-minute self-paced run, partial sleep deprivation reduced distance covered by 6%. Same runners, same test, six percent less ground.

Again the mechanism is perceived exertion. Sleep deprivation reliably raises RPE, so the same pace feels harder with no change in what your body is actually doing. You aren't imagining it and you aren't being soft. Your brain is reporting a higher cost for identical work, and since perceived effort governs how long you'll hold a pace, that matters over a long run or a race.

This is also why a poor night shows in your data before it shows in your legs. Suppressed HRV and elevated resting heart rate after bad sleep is among the clearest signals a wearable produces, and it's the basis of the approach in our guide to HRV-guided training.

How does poor sleep affect race-day decisions?

Perceived exertion is a brain output, and it isn't the only one that degrades.

Sleep loss reliably slows reaction time, narrows attention, and impairs the executive function you rely on to make judgements while tired. For a runner that surfaces in the decisions that decide a race rather than in any single physical measure. Whether you go with the group that surges at 15km. Whether you take the gel you promised yourself at 40 minutes when your stomach is arguing. Whether you back off when your heart rate has been telling you to for the last two kilometres.

The sport-specific evidence here is thinner than the laboratory evidence, but it points the same way. When Mah and colleagues extended eleven basketball players' sleep by an extra 111 minutes a night, sprint times improved by about 4%, while the larger gains came in shooting accuracy and reaction time, which are precision skills performed tired. This is the study usually quoted as proof that sleep makes athletes 9% better. The 9% is free throw accuracy, in basketball players, so treat it as evidence about skill and judgement rather than endurance.

That distinction is worth keeping, because judgement is where most amateur races are actually lost. A runner who blows up at 30km having ignored three separate warning signs is not always a runner who was unfit.

What sleep does while you're in it

Sleep isn't uniform. The stages that matter most for runners are unevenly spread across the night.

Slow-wave sleep is where repair happens. Growth hormone release is synchronised with deep sleep, and that pulse drives tissue repair, protein synthesis and muscle glycogen resynthesis. Slow-wave sleep is concentrated in the first half of the night, so a late bedtime with a fixed alarm costs more than the missing hours suggest.

REM consolidates motor learning. The pacing judgement and movement economy you built in a session get filed during REM, which dominates the second half of the night. Cutting sleep short at the front end takes REM disproportionately.

Both halves are doing different work, and there's no version of six hours that gets you both.

The next two sections are the ones coaches actually watch for. Neither is about how fast you run on the day. Both are about whether you get to run at all.

Does poor sleep increase injury risk?

A 2021 review in Current Sports Medicine Reports put a usable threshold on this. Seven hours or less, sustained for at least a fortnight, is associated with roughly 1.7 times the risk of musculoskeletal injury.

The word doing the work there is sustained. This is not a penalty for one bad night. It's what happens when short sleep becomes the normal state for a fortnight, which for most runners means a busy month at work rather than anything they would describe as sleep deprivation.

The same review makes a point about causation that most runners have backwards. In studies of both children and adults, poor sleep predicts next-day pain more consistently than pain predicts poor sleep. The assumption is usually that the niggle came first and costs you sleep. More often it's the other way around.

The mechanisms are unglamorous and they stack. Tissue repair depends on a growth hormone pulse you're getting less of. Movement control degrades, and running is thousands of repetitions of a task that only stays cheap while the control is good. The session costs more than it should. Then you run it anyway, because it was in the plan.

Why do runners on low sleep keep getting colds?

Ask a coach what they actually see when an athlete's sleep slips and the answer usually isn't a poor time trial. It's the third cold of the winter.

The evidence here is unusually strong for a question like this, because somebody ran the experiment properly. A 2015 study in Sleep measured 164 adults' sleep for a week using wrist actigraphy, then quarantined them in a hotel, administered rhinovirus as nasal drops, and tracked who actually developed a cold.

Under six hours a night in the week beforehand meant 4.2 times the odds of getting ill, compared with more than seven hours. Under five hours meant 4.5 times. Short sleep predicted infection more strongly than age, stress, income, education or smoking did.

For a runner that lands on top of something else. Hard training is transiently immunosuppressive in its own right, particularly around long sessions and races, so a heavy block on short sleep is two pressures on one system.

And the cost isn't really the cold. It's the ten days of training the cold takes with it, usually at the point in a block where you can least afford to lose them. Nobody plans a taper around a head cold in week nine.

The part nobody connects: sleep and fuelling

This one surprised me, and it's the link most runners never make.

Two nights of sleep restriction measurably reduces peripheral insulin sensitivity in healthy young adults. Your muscles become worse at taking up glucose from the blood.

Follow that through. Impaired glucose handling means less carbohydrate available to working muscle and slower glycogen resynthesis after training. The window in the first couple of hours after a hard session, when glycogen synthase activity and insulin sensitivity are both elevated, is exactly the window sleep loss degrades. So the runner who trains hard, fuels properly and then sleeps five hours has undermined the recovery nutrition they just paid attention to. You can do everything right with your race day nutrition and short sleep will still blunt what your body does with it.

There's a second fuelling problem alongside it, and this one is about the signals rather than the machinery. Spiegel and colleagues gave twelve young men two short nights and then two long ones, holding calories and activity constant throughout. Leptin, the hormone that tells you you've had enough, fell 18%. Ghrelin, which tells you to eat, rose 28%. Self-reported hunger climbed 24%, and the sharpest rise was in appetite for calorie-dense, high-carbohydrate food. Two nights again, which is the same short timescale as the insulin finding.

Read that as a runner rather than as a dieter. In precisely the weeks your sleep is short, appetite stops being a reliable instrument for working out what you need. So fuel to a plan rather than to hunger when you know you're down on sleep, and treat unusual hunger as information about your sleep rather than about your training.

It also cuts the other way, which is worth knowing: large meals within about two hours of bed disrupt sleep architecture, so late evening fuelling after a late session is a genuine trade-off rather than a free choice.

The uncomfortable part

Runners are among the worst sleepers in sport. Survey data consistently finds shorter sleep duration in runners than in team-sport athletes, and depending on the survey, between a fifth and two fifths of athletes report under seven hours a night.

The reason is obvious. Training around a job and a family leaves two windows, early morning and late evening, and both come out of sleep.

So the honest question: is the run you're getting up early for worth more than the hour of sleep it costs?

Sometimes yes. A key session, on reasonable sleep, usually earns the early alarm. What changes the answer is what the session is and how the last few nights have gone.

The thing to protect is intensity. Hard sessions are what sleep loss taxes most and what return least when you run them wrecked, so when sleep is short, intensity is the first thing to move. Easy running is a different case and shouldn't simply be deleted. Keep it in, make it shorter, and take it for what it is on the day.

The honest version of the timescale: one bad night won't derail anything. Three to five in a row will, and that's the point where the week needs to change around you, with missed morning sessions, minimal intensity, and a general reframing of what this block is going to be.

A better way to plan the week. Most runners look at the week ahead and ask what training they can fit in. The more useful question is whether there's time to genuinely recover from the sessions already written down, given everything else competing for the same hours. How much slack is actually in the schedule? If the answer is none, the plan needs adjusting rather than hoping, and that's a decision worth making on a Sunday rather than discovering on a Thursday.

The runners who improve fastest are rarely the ones fitting in the most sessions. They're the ones whose training and recovery are actually balanced, which is the argument in our piece on training load management.

The sleep audit

Run this once. Ten minutes, and it will tell you more than another month of guessing.

Duration and consistency

1. Over the last fortnight, what was your average time asleep, not time in bed? Under seven hours is a problem worth solving before you change anything else in training.

2. Does your bedtime vary by more than an hour across the week? Irregularity costs nearly as much as short duration.

3. Do you use weekends to catch up? Needing to is the signal, not the solution.

The hour before bed

4. Do you have any kind of wind-down routine, or do you go from a laptop straight to a pillow? Half an hour of something dull and low-lit beats most of what gets sold as a sleep aid.

5. Are you finishing hard sessions within three hours of bedtime? Core temperature and adrenaline both stay elevated and delay sleep onset.

6. Are you eating your main meal late? Large meals inside two hours of bed disrupt sleep architecture even when you fall asleep fine.

7. Are you eating enough across the day, and enough carbohydrate at dinner after a big session? Under-fuelling is a common and badly underrated cause of waking in the night or waking early. Fuel for the work you've done.

8. Alcohol in the evening? It shortens time to fall asleep, then fragments the second half of the night, taking REM specifically.

9. Screens in bed? The behavioural stimulation matters more than the blue light, whatever your phone settings claim.

The environment

10. Is the room genuinely dark and cool? Around 18°C suits most people.

11. Is there light from a charging device, a router or a street lamp? Small amounts matter more than people expect.

The signals

12. Do you wake before your alarm most days? The strongest single indicator you're getting enough.

13. Is your resting heart rate elevated and HRV suppressed the morning after short nights? If it's visible in your data, it's real.

14. Do you need caffeine to function before 10am? A tell, not a solution.

Scoring it. Three or more problem answers and sleep is likely your biggest available performance gain, ahead of any change to your training. One or two, fix those specifically. None, and sleep isn't what's limiting you, so look elsewhere.

What about your sleep score?

Worth saying plainly, from a company that builds software which reads your sleep: the number is an input, not a grade.

Consumer wearables are genuinely good at the parts that matter most, duration and consistency, and considerably weaker at splitting the night into stages, which is the part of the display people study hardest. Treat your hours and your bedtime regularity as a signal, and the deep sleep percentage as an estimate.

A 2017 paper in the Journal of Clinical Sleep Medicine described patients turning up at sleep clinics with problems their trackers had helped create, a perfectionist pursuit of ideal numbers that was itself making their sleep worse. The authors named it orthosomnia, after orthorexia. If the first thing you do each day is check a score, and your mood for the next hour depends on what it says, the device has stopped helping. You’re better off asking yourself how you feel first then check the score. If it's an ongoing source of anxiety for you, quit tracking or using that score.

The test is whether the number changes a decision. A run of short nights that tells you to move Thursday's intervals is data doing its job. A 62 that ruins a morning you'd otherwise have felt fine about is not. Use the weekly trend and ignore the nightly verdict.

Naps, and how to use them properly

Napping is a tool rather than an admission of failure, but duration determines what you get.

20 to 30 minutes keeps you in light sleep. You wake alert, reaction time improves, no grogginess. This is the default and what most runners should use.

Around 90 minutes takes you through a full cycle including slow-wave sleep, so you get some hormonal repair benefit and wake at the end of a cycle. Use this when genuinely repaying a debt, such as after a bad night before a hard session.

The middle is the trap. Waking out of deep sleep at 45 to 60 minutes produces the worst grogginess and is the least useful nap length available.

On timing: nap in the early afternoon, roughly between 1pm and 3pm, when the circadian dip makes falling asleep easier. Later than about 3pm and you start eating into the sleep pressure you need that night.

And if you can't nap, lie down anyway. Twenty to thirty minutes flat in a dark, quiet room with your eyes shut does more than people expect, even if you never drop off. You won't get what actual sleep gives you, but you'll take the edge off and give yourself a reset. Bonus is it's usually available on a day when a real nap isn't.

A nap doesn't replace a night's sleep. It takes the edge off a deficit, which is worth having, but a fortnight of naps compensating for a fortnight of short nights is a losing arrangement.

What to do when you've slept badly

You'll have bad nights. What matters is what happens next.

The instinct is to carry on as planned. Better is to adjust the session rather than the week. One poor night before an easy run is nothing. One poor night before intervals means the session will cost more and return less than it should, so it's usually worth moving it, or adjusting it down and taking what you can get from it.

Three nights in a row is different. That's an accumulating deficit, and training hard through it is how runners arrive at unexplained plateaus, stubborn niggles, the third cold of the season, and the sense of working harder for less.

The difficulty is that these judgements are genuinely hard to make about yourself at 6am. This is where adaptive coaching earns its place: a system reading your sleep, HRV and recent load together can tell you today's session should move, before you've talked yourself into it.

Key Takeaways

• Endurance athletes given an extra 90 minutes of sleep for three nights completed a one-hour time trial around two minutes faster, roughly 3%. The mechanism is perceived exertion: more sleep makes a given intensity feel more manageable.

• Sleep deprivation has a moderate negative effect on endurance, and it worsens beyond 30 minutes of effort. Partial deprivation cut distance covered in a 12-minute run by 6%.

• Sleep loss degrades judgement as well as legs. Reaction time, attention and decision-making all suffer, so it shows up in pacing and fuelling choices before it shows up in your splits.

• Under six hours a night meant 4.2 times the odds of catching a cold after controlled exposure to a virus, outpredicting age, stress and smoking. The cost isn't the cold, it's the training block it interrupts.

• Seven hours or less, sustained for a fortnight, is associated with around 1.7 times the musculoskeletal injury risk. And poor sleep predicts next-day pain more reliably than pain predicts poor sleep, so the niggle often follows the short nights rather than causing them.

• Sleep loss disrupts both the machinery and the signals for fuelling. Two nights of restriction cuts insulin sensitivity, drops leptin 18% and raises ghrelin 28%, so appetite stops being a reliable guide. Fuel to a plan in weeks when sleep is short.

• Both halves of the night matter. Deep sleep and its growth hormone pulse come early, REM and motor consolidation come late. Six hours gets you neither properly.

• Nap 20 to 30 minutes for alertness or around 90 to repay a debt. Avoid the 45 to 60 minute window, keep it before 3pm, and if you can't sleep, lie down in a dark room anyway.

• Your sleep score is an input, not a grade. Trust duration and consistency, treat the stage breakdown as an estimate, use the weekly trend, and ignore the nightly verdict.

• Run the audit. Three or more problem answers means sleep is your largest available gain, ahead of anything you'd change in training.

Most runners hunting their next improvement go looking for a new session, a different plan, or a piece of kit. For a good number of them the answer is duller and considerably cheaper: go to bed an hour earlier for a month and see what happens.

Zepho is adaptive coaching for serious runners. It's live and available now on the App Store.