How much should I train after a bad night of sleep?

It is six in the morning, you slept five hours, and there is a threshold session in the plan. The honest answer is not "skip it" and not "push through". It depends on which of your systems the short night actually taxed, and that is knowable.

What short sleep actually costs you

Sleep loss does not degrade every capacity equally. Maximal strength holds up surprisingly well for one night; a single session of sleep restriction barely moves a one-rep max. What degrades first is everything that depends on the central nervous system staying sharp: reaction time, time to exhaustion, perceived effort, technical skill and decision-making.

That is why the Stanford sleep-extension work found free-throw accuracy up 9% and sprint times down about 5% when players slept more, with no change to their training. The pattern runs in both directions: the skill and the speed move, the raw force does not move much.

The practical consequence is that after one short night, heavy lifting at a familiar weight is usually fine. A session that depends on holding good form while very tired — intervals, technical lifts, anything near failure — is where the risk is.

One night is noise. Three is a trend.

A single short night is within the normal variation of an adult life and your body treats it that way. The signal to act on is accumulation: consecutive short nights, or short nights landing on top of a training block that is already climbing.

This is also where the injury evidence sits. Milewski's cohort of 112 adolescent athletes found those sleeping under eight hours were 1.7 times more likely to be injured — and the exposure measured was habitual sleep, not one bad Tuesday.

A decision rule you can use tomorrow

Check three things before you decide, in this order.

How far below your own normal was last night? Not below eight hours — below your average. An hour down is noise. Two or more, or under about six hours in absolute terms, counts as a short night.

Is this the first one? If yesterday and the day before were normal, train roughly as planned. If this is the second or third, the session changes.

What was the session going to cost you? Easy volume is nearly always fine and often helps. Heavy technical work and hard intervals are what you move.

The default adjustment is not cancellation. It is swapping the quality session for easy volume of the same duration, and moving the quality work to the first day after a normal night. You keep the week's volume, you lose only the intensity you could not have executed well anyway.

What PeakU does with this

PeakU reads last night's sleep from whatever wrote it to Health Connect — Samsung Health, Garmin, Oura, Whoop, Polar — scores it against your own rolling baseline rather than a population target, and weighs it alongside HRV and your acute-to-chronic load. That becomes the morning call: push, moderate, easy or rest.

Crucially it knows whether this is the first short night or the third, because it holds the history, and it knows what today's session was going to be, because it holds the plan. So the answer is not a generic warning. It is keep the run, drop it to easy, and move the intervals to Thursday — and the plan reflows so Thursday still works.