Sleep and Indoor Rowing: What the Evidence Actually Says

The context

Indoor rowing is an evening sport. Most club sessions start between six and half past six in the evening, which puts the last hour of your day in the boat, and puts the hours after it under pressure whenever a session runs long or the race is on Saturday. Add a job, a commute, and a house that needs quieting, and the rower who "sleeps fine" is often the rower who has never measured anything.

There is a second reason to be careful about the literature here. Most sleep-and-performance studies recruit healthy young men and women who are not rowers, hold them in a laboratory, and restrict or remove sleep on a schedule no club ever runs. Very little of it is done on an ergometer. So the sensible starting position is this: the mechanism is well established and the dose-response curve is well measured in general populations, but the step from those findings to your own two-thousand-metre piece is an inference, and this article keeps that step visible rather than hiding it.

The core idea

The single most useful thing in this field is that you cannot feel your own sleep debt. In the best-known experiment, healthy adults spent fourteen nights in bed for either four, six or eight hours, or went three nights without sleep at all ([3] Van Dongen 2003). Six hours a night produced cognitive deficits equivalent to up to two nights of total deprivation. Yet sleepiness ratings jumped in the first day and then barely moved, so the researchers concluded that the subjects were largely unaware of how badly their performance was deteriorating. The same study found that lapses in alertness tracked cumulative wakefulness almost linearly once past about sixteen hours. That is the argument for judging your sleep by something external - a standard time, a standard test, your coach's observation of your stroke - rather than by how you feel on the morning of the session.

What a short night actually costs, in numbers, comes from the largest meta-analysis available. Sixty-nine publications, 227 outcome measures, and sleep loss defined as six hours or less inside any twenty-four-hour period: average exercise performance fell by 7.6 per cent, with a confidence interval running from about 3 to 12 per cent, and the effect was significant across every category of exercise they examined ([4] Craven 2022). What that figure is not is an ergometer figure: the categories pooled were anaerobic power, speed and power endurance, high-intensity intervals, strength, endurance, strength-endurance and skill, and no rowing ergometer appears among them. Two further details matter more than the headline. The penalty grew with time awake before the session, by roughly four tenths of a per cent for every extra hour - in the deprivation and late-restriction protocols, which were the two that carried the effect at all. And the effect was concentrated in afternoon and evening work, with morning tasks largely unaffected, which is why the authors recommend prioritising morning exercise when sleep loss is unavoidable.

One rough night is less dramatic than that sounds. In thirteen men tested for maximal and peak power after a single night without sleep, anaerobic performance was unaffected at twenty-four hours awake and only impaired by thirty-six ([5] Souissi 2003). If you lost one night before a two-thousand-metre piece, you are not finished. It is the accumulated debt that takes the performance, and it is the sleep you lose after the session that costs you the following one. One night of total deprivation cut muscle protein synthesis by eighteen per cent, raised cortisol by twenty-one per cent and lowered testosterone by twenty-four per cent ([6] Lamon 2021). That is a recovery deficit, and with back-to-back erg sessions it lands on the next session rather than disappearing overnight.

One more measurement is worth correcting, because it is the opposite of what most people assume. Elite athletes do not sleep fewer hours than matched non-athletes. In a comparison of forty-seven Olympic athletes against twenty controls, total time asleep did not differ; what differed was quality. Sleep efficiency ran 80.6 per cent in the athletes against 88.7 per cent in the controls, and the fragmentation index ran higher too, at 36.0 against 29.8 ([7] Leeder 2012). The authors noted the athletes' sleep still fell within the range of healthy sleep. So if you sleep seven hours in a noisy house, adding a tenth of an hour to a bad night's sleep does less for you than fixing the waking up at three.

On the duration number itself, be careful about who is talking. The National Sleep Foundation's expert panel, using a formal consensus method, recommended seven to nine hours for adults and eight to ten for teenagers, explicitly for healthy people without a sleep disorder ([1] Hirshkowitz 2015). The athlete consensus reaches for a different posture: it warns that a single fixed range is unlikely to be ideal for every individual and recommends working from the athlete's own perceived sleep needs ([2] Walsh 2021). Seven to nine is a sound public-health range and a reasonable floor to aim at. It is not a prescription, and treating it as one sets you up to fail a target that was never yours.

The how-to

If you know you are short, move the test to the morning. This is the most direct application of the meta-analysis: sleep loss hurt afternoon and evening performance while leaving morning performance largely intact ([4] Craven 2022). If the club has a choice between an evening time trial and a morning one, and you have had five hours, ask for the morning.

Treat caffeine as a single decision, not two. The same molecule that helps a race hurts the night. Four hundred milligrams taken at bedtime, three hours before, or six hours before all significantly disrupted sleep measured in the home environment ([9] Drake 2013). But caffeine before a two-thousand-metre erg piece is genuinely effective in rowers: six and nine milligrams per kilogram improved performance time by 0.7 and 1.3 per cent in trained oarswomen, with most of the gain in the first five hundred metres ([10] Anderson 2000). Those two facts only conflict if you treat them separately. The dose that helps a piece is the dose that will cost you the night. Count the caffeine from breakfast onward, not just the pre-race cup.

A thirty-minute nap buys you something measurable. After a night of four hours, ten men who took a half-hour nap after lunch showed improved alertness, better reaction-time accuracy, and faster two-metre and twenty-metre sprints ([11] Waterhouse 2007).

Do not read one morning heart rate reading as a verdict. Across twenty-eight thousand people and nine million measurements over five years, heart rate variability moved with training of different intensities, alcohol, the menstrual cycle and illness, and the authors' conclusion was that it is a more sensitive but not specific marker of stress ([8] Altini 2021). It is a good instrument and a poor oracle. A single low reading on a Tuesday morning is a data point; a trend across a fortnight is evidence.

Alcohol is a dose question, not a yes-or-no question. A review of alcohol and recovery in male athletes concluded that around 0.5 grams per kilogram of body mass consumed after exercise is unlikely to affect most aspects of recovery, while larger amounts may alter blood flow and protein synthesis, with rehydration and glycogen resynthesis affected to a lesser extent ([12] Barnes 2014). The same review is explicit that the answer also depends on the timing of the alcohol after the session, the recovery time left before training again, and any injury you are carrying. The Friday night after a race and the night before a hard session are not the same question, and the threshold is a fraction of your own body mass rather than a count of drinks.

The trade-offs

Sleep is a training variable, and every training variable competes with something. More time in bed costs training time, which means the honest way to buy sleep is to schedule it rather than to wish for it - by moving an evening session, by capping the length of a long Saturday, by not scheduling a hard test the night after a night out. The expert consensus is also clear that we do not yet know how to bank sleep across a busy week ([2] Walsh 2021), so do not plan a single eight-hour night to rescue a bad fortnight.

There is a trade-off in the research itself that is worth naming. Most of these studies are small, most are short, and very few include rowers. The caffeine rowing study had eight oarswomen in it ([10] Anderson 2000). The nap study had ten men ([11] Waterhouse 2007). The sleep-deprivation study of muscle protein synthesis had thirteen adults ([6] Lamon 2021). These are good studies, and they are also small ones. When the rower in front of you quotes a percentage, it is worth asking how many people were in it.

What to do tomorrow

Pick two, not five.

Write down your wake time for seven days before changing anything else. A fixed wake time, including at the weekend, is what makes the other advice work, and it tells you which nights are actually short.

On the next session where you know you slept badly, do the session but do not set a personal best in it, and if the club offers the choice, take the morning slot.

Move your caffeine cut-off to at least six hours before bed and keep it there through the training week, not just on race days. Six hours is the minimum the evidence supports, and at that interval the effect on sleep was still significant - the dose that was tested was 400 milligrams ([9] Drake 2013). And a racing dose can reach that: six milligrams per kilogram is about 386 milligrams in an oarswoman of 64.4 kilograms ([10] Anderson 2000).

If there is a hard block in your week, put a thirty-minute nap in the middle of it rather than another hour in bed at night.