Why Ultramarathon Runners Should Bank Sleep

For 100-mile and longer ultramarathons, the best sleep tip is not a race-morning trick. It is to arrive with more sleep already behind you. The most defensible version is simple: extend sleep for the last 2 to 3 nights before the start, then protect that advantage with careful caffeine timing and short naps only when the race duration makes them necessary.

That matters because an ultramarathon is not only long enough to drain glycogen, damage muscle, and make feet uncooperative. It is long enough for sleep loss to become a navigation problem, a footing problem, and a judgment problem. In a survey of 636 ultramarathon runners, 55% reported intentionally extending sleep before a race; the same study found that only 16.4% slept during one-night races, compared with 94.6% during races lasting more than two nights.[1] In a later survey of 1,154 trail runners, 80% reported sleep-deprivation symptoms, and pre-race sleep extension was associated with a lower prevalence of falls.[2]

What “Banking Sleep” Should Mean Before an Ultra

“Sleep banking” is useful language only if it stays modest. It should not mean that sleep is literally stored like water in a bladder or miles in a training log. Professor Russell Foster of Oxford has argued that “you can’t really bank sleep,” because sleep is self-regulated by the body rather than deposited into a reserve account.[3] That objection is worth taking seriously, especially in a sport where a phrase can become a ritual faster than the evidence can catch up.

A better interpretation is that sleep extension gives a runner more margin before deprivation takes over. You are not making sleep optional. You are reducing the chance that the first overnight section begins on top of an already-short week.

The real-world evidence points in that direction. In the Ultra-Trail du Mont-Blanc setting, runners who used sleep extension before the race finished faster than those who did not.[4] The survey data are not perfect: much of this work relies on self-report, which can be affected by recall bias, and a small actigraphy study in this area included only 19 participants.[5] Still, the pattern is hard to ignore because it lines up with controlled sleep-deprivation research. In one experiment, 6 nights of sleep extension helped prevent reaction-time degradation and reduced microsleep episodes during later sleep deprivation.[6] Another study found that sleep extension before deprivation enabled faster recovery afterward.[7]

For an ultrarunner, reaction time and microsleeps are not laboratory curiosities. They are the difference between lifting a toe over a root and clipping it, reading a course marker and drifting past it, or making a reasonable decision at an aid station instead of solving the wrong problem with more caffeine.

The Practical Pre-Race Protocol

The most race-relevant protocol is not complicated: aim for 10 hours in bed for each of the 3 nights before the start. Ultrarunning coach Jason Koop has recommended that approach as a practical pre-race sleep-extension strategy, paired with a bias toward pushing through the first night and saving planned naps for the second night when needed.[8]

Ten hours in bed does not guarantee 10 hours asleep, and it does not need to. The point is to make sleep possible. If a runner normally sleeps 7 hours and spends the final week cutting sleep for packing, travel, work, and race anxiety, the start line becomes the first sleep-deprivation checkpoint. The 3-night approach moves the decision earlier, when the runner still has control over lights, alarms, caffeine, and obligations.

This is also why the final week is a poor time to treat caffeine as part of a personality. Morning coffee is one thing. Late-day caffeine during the sleep-extension window is different because it can erase the opportunity the protocol depends on. A runner who sacrifices the second pre-race night and then asks caffeine to rescue the first overnight descent has changed the plan without admitting it.

One-Night Races and Multi-Night Races Need Different Sleep Choices

A 100-mile race that runs through one night is not the same sleep problem as a 200-mile race, a multi-day mountain event, or a back-of-pack 100 with a second night in play. The Martin survey makes that split visible: sleeping during the race was uncommon in one-night events but nearly universal in races lasting more than two nights.[1]

For many one-night 100-mile attempts, the reasonable plan is to extend sleep before the race, start rested, and avoid unnecessary sleeping during the first night unless safety deteriorates. That does not mean everyone should refuse sleep. It means that a planned nap at the first sign of darkness is not automatically the best use of time. Terrain, weather, cutoffs, crew access, and the runner’s history with sleepiness all matter.

Once the race becomes multi-night, the calculation changes. Sleep pressure is no longer a brief inconvenience to outlast. It becomes a recurring operational problem. In Kishi and colleagues’ survey, 82% of mid-race naps were shorter than 30 minutes.[2] That detail deserves more attention than the folklore about athletes who barely sleep, because it describes a tool ordinary runners can actually use: brief, planned sleep that restores enough alertness to keep moving without turning an aid-station chair into the finish line.

A short nap is most useful when it has a job. That job might be to get through a technical descent before dawn, reset after navigation errors, or stop a string of near-falls. It is less useful when it is vague: “I’m tired, so I’ll lie down and see what happens.” In a race environment, vague sleep tends to expand.

• Before the race: extend sleep for 2 to 3 nights, with about 10 hours in bed as a practical target.
• During a one-night race: protect forward progress unless sleepiness becomes a safety issue.
• During a second night or later: plan short naps before decision-making and footing collapse.
• At aid stations: decide the nap length before lying down, and make waking someone’s explicit job.

Caffeine Helps Most When It Does Not Sabotage Sleep

Caffeine is not the enemy here. Poor timing is. Sports nutrition guidance commonly discusses caffeine dosing in the range of 3 to 9 mg/kg for performance, but an ultramarathon sleep plan should decide when caffeine is worth spending.[9] Used too early, it may make the first night feel cleaner while damaging the next sleep opportunity. Used late in the pre-race evening, it can interfere with the very sleep extension the runner is trying to build.

The cleaner rule is to keep caffeine boring during the sleep-extension nights, use it conservatively on race day, and save more deliberate caffeine use for the period when sleep pressure is actually threatening performance or safety. For many runners, that means the second night or the hours before dawn, not the afternoon before the start.

There is a useful parallel in extreme sleep-deprivation settings outside running: solo sailors also have to combine short naps, stimulant discipline, and decision-making under fatigue. Restful Ground’s guide to solo sailor sleep deprivation survival is not evidence for ultramarathons, but it is a useful comparison because the problem is similar: a tired person still has to make accurate decisions in an unforgiving environment.

What Not to Borrow From Elite Outliers

Elite no-sleep stories are interesting for about as long as they remain stories. They are not protocols. A runner who can cover extraordinary distances on minimal sleep may have unusual tolerance, exceptional pacing skill, years of adaptation, unusual support, or all of those at once. None of that tells an amateur runner how to descend a rocky trail at 3 a.m. after a full workweek.

The better lesson from the available evidence is less glamorous. Many ultrarunners already try to extend sleep before racing. Pre-race sleep extension has been associated with faster finishing and fewer falls. Controlled sleep studies make the mechanism plausible through reaction time, microsleeps, and recovery. The evidence is not strong enough to pretend that sleep is literally stored, and it is not weak enough to ignore.

So the usable answer is bounded: bank sleep in the 2 to 3 nights before the race, understand “bank” as added margin rather than stored immunity, protect the sleep-extension window from late caffeine, and use short strategic naps when the race becomes multi-night. That is not a guarantee of a clean race. It is a way to avoid making preventable impairment part of the plan.

References

1. Sleep habits and strategies of ultramarathon runners, PubMed Central, 2018, link
2. Sleep strategies and fall prevalence in trail running, PubMed Central, 2024, link
3. Run Ultra interview with Professor Russell Foster, Run Ultra, 2024, link
4. Sleep management strategy and performance in an extreme mountain ultra-marathon, Research in Sports Medicine, 2015, link
5. Sleep and ultramarathon running: actigraphy field data, 2022
6. Sleep extension before sleep loss: effects on performance and neuromuscular function, 2016
7. Banking sleep: realization of benefits during subsequent sleep restriction and recovery, Sleep, 2009
8. Sleep Strategies for Ultrarunners, CTS Ultrarunning Coach
9. International society of sports nutrition position stand: caffeine and performance, Journal of the International Society of Sports Nutrition, 2010