Move

Training on Empty: How to Adjust Your Workout When You Have Barely Slept

Cameron Usher · Founder, PX

13 August 2026 · 7 min read

12 years in competitive sport; trained alongside the Australian Olympic squad; coached 100+ people one-on-one.

The alarm goes off and you already know. Four, maybe five hours, the broken kind. Today has a session on the plan, and the two voices start immediately: one says push through, sleep is for the weak. The other says skip it, training tired is pointless. They are both wrong, and they are both expensive.

Working out on poor sleep is not a yes-or-no question. It is a sizing question. Sleep is when the adaptation you train for actually gets built, when tissue repairs and the nervous system resets, so a short night genuinely changes what today can absorb. But it almost never changes it to zero. The skill is knowing what one rough night actually costs, adjusting the day to match, and recognising the point where a pattern of rough nights turns this from a training question into a recovery one.

What a short night actually changes

Be precise about the cost, because it is narrower than it feels.

After one bad night, your capacity for high-intensity and highly technical work drops the most. Coordination, reaction time, top-end power, appetite for hard efforts, willingness to grind the last reps: these all take the hit, and with them, the injury risk of complex movement under fatigue quietly rises. Your capacity for easy aerobic work barely changes. A steady Zone 2 session after a five-hour night is entirely defensible. Maximal intervals are not the same conversation.

Depth matters as much as duration, and this is where your wearable earns its place. Two seven-hour nights can leave you in different states depending on how much deep sleep they contained, because deep sleep is where the physical repair concentrates. A night that looked acceptable in hours but was cut badly short on depth, a late night after alcohol is the classic, deserves the same caution as a plainly short one. If you only read one number off your sleep data, read duration. If you read two, read depth.

And one bad night is not a trend. Your baseline matters more than any single reading, which is the whole premise of wearable workout data tracking: every number is a comparison against your own last fourteen days, not a verdict in isolation. One outlier night against a solid baseline is a paper cut. The same night as the fourth in a row is a different animal, and we will get to it.

The adjustment, by how bad it was

The quick version, by hours slept. Anchor it to your own normal, not to eight hours as a universal law.

The evidence behind taking this seriously is not soft: a systematic review and meta-analysis of acute sleep loss found physical performance measurably impaired after short sleep, with the impairment growing the longer you have been awake. Now the three tiers, properly.

Tier 1: an hour or so short, or noticeably broken. Train, with the top trimmed. Keep the planned session type, pull intensity back a step and trim the volume, roughly the final set or last hard rep of each block. Stretch the warm-up and let how it feels cast the deciding vote: some days the body shows up anyway and you finish nearly full; some days it confirms the trim was right. Either way the day counts.

Tier 2: five to six hours, or deep sleep badly cut. Switch the session type. This is the tier people get wrong, forcing planned intervals through a nervous system that did not reset. Today's high-intensity or heavy session becomes easy aerobic work or technique and mobility, at a conversational effort, thirty to forty-five minutes. The hard session is not deleted, it moves two or three days out. You lose nothing real by moving it; you lose plenty by doing it badly today and paying for it tomorrow as well.

Tier 3: under five hours, or the third-plus bad night running. The honest options are gentle movement or deliberate rest. A walk, easy mobility, breathwork, or nothing, chosen on purpose. Under real sleep deprivation, hard training adds stress to a system that is already failing to clear the stress it has. If you feel the guilt rising, that is the moment to reread how to decide whether you should train today, because guilt is the single worst input to this decision.

One overriding rule across all three tiers: no new territory on a short night. No first attempt at a heavier lift, no unfamiliar complex movement, no leap in volume. Fatigue plus novelty is where avoidable injuries live.

When it stops being a training question

A pattern of short nights changes the frame. If most of your week is Tier 2 or worse, the highest-leverage training decision available to you is not inside the training at all. It is the bedtime, the caffeine cut-off, the screen habits, the late sessions that push your wind-down back. Training load itself can be part of the cause: hard evening sessions too close to bed disturb the very sleep that pays for them, one more way the pieces of your day are one system wearing app-shaped disguises. Fixing the week beats perfecting the workaround, and a run of genuinely easy days while sleep rebuilds, the deliberate version described in how to tell if you need a rest day, is often the fastest route back to real training.

Getting through the tired day well

Whichever tier you land in, a few practical moves make the adjusted day go better.

Train earlier than you think, not later. The intuition says push the session late and recover first. The evidence says the opposite: after a short night, impairment grows with hours awake, so the late-evening session lands on the worst version of you, and it then threatens tonight's sleep as well. If the morning slot is gone, aim for late morning or midday, and treat the evening as protected wind-down rather than a second chance.

Front-load the caffeine, then stop. Caffeine genuinely helps a tired session, and it also has a long half-life. On a bad-sleep day the single most self-defeating move is the 4pm coffee that props up the afternoon and then quietly takes a bite out of the night you were counting on to fix everything. Have it early, have enough, and cut it by early afternoon.

Do not under-eat the tired day. Short sleep reliably drags appetite signals toward quick carbohydrates and makes disciplined eating harder, while the day itself needs normal fuelling to support even the adjusted session. Eat properly and deliberately today; a fuel shortfall stacked on a sleep shortfall is how one bad night becomes a bad week.

Protect tonight like it is training. Because it is. The adjusted session, the early caffeine cut, a proper wind-down: today's real workout is the one that happens after dark. One good night usually resets the whole picture, and everything above goes back in the drawer.

What this looks like decided for you

Everything above is a method you can run by hand at 5:50am, against every incentive to just do whatever the plan already said. Or it is simply how your plan behaves.

Run automatically, it looks like this: your wearable's sleep data, duration and depth both, lands against your baseline overnight. Before you are up, today has already been resized: the interval session swapped to forty easy minutes, the intensity capped, the hard day moved to Friday. And the why sits in plain language on the day itself: deep sleep came in well under your normal last night, so today protects the week instead of spending it.

The reason is the point. Anyone can shuffle a calendar. Seeing why, morning after morning, is how you learn your own sleep-to-performance pattern, which nights genuinely cost you and which ones your body shrugs off. That knowledge outlasts any app.

PX resizes the day while you are still asleep. It reads duration and depth against your own baseline, adjusts the session, the intensity and the timing, and tells you why, across training, mind and food as one plan. Start your free trial and wake up to the adjusted day.