How tired is too tired to drive?

Roughly 17 to 19 hours awake produces impairment comparable to a blood alcohol concentration around 0.05%, and 24 hours awake is comparable to about 0.10%. You cannot judge this reliably from the inside, so use behavioural signs instead.

What is a microsleep?

A brief intrusion of sleep into wakefulness, typically a few seconds and sometimes with the eyes open, during which you are not processing the road at all. People frequently have no awareness one occurred. Four seconds at 70mph is over 100 metres travelled blind.

What should I do if I get sleepy while driving?

Pull over somewhere safe and nap 15 to 20 minutes, ideally with 150 to 200mg of caffeine taken immediately before. Opening a window or turning up the music does essentially nothing. If you are genuinely impaired, stop driving altogether.

Drowsy Driving and Microsleeps

Q: Why can't I tell how tired I am?

Sleep deprivation impairs the faculty you would use to assess it. Studies consistently find subjective sleepiness ratings plateau while objective performance keeps declining, so people stop noticing while their reaction times keep worsening.

Most sleep advice is about optimisation. This one is about a risk that kills people, and the reason it is worth its own article is a specific, well-replicated finding: you cannot accurately judge your own impairment.

The comparison that makes it concrete

Sleep deprivation degrades performance in ways comparable to alcohol. The commonly cited findings: roughly 17 to 19 hours awake produces impairment similar to a blood alcohol concentration around 0.05%, and around 24 hours awake is comparable to roughly 0.10% — over the legal limit nearly everywhere.

Reaction time, sustained attention, decision-making and hazard perception all decline. Getting up at 7am and driving at 2am puts you in that territory, and nobody would consider the equivalent number of drinks.

Microsleeps

The mechanism that makes drowsy driving different from ordinary tiredness.

A microsleep is a brief intrusion of sleep into wakefulness — typically a few seconds, sometimes with the eyes open. During one you are not processing the road at all.

Two things make them dangerous:

You often do not know it happened. People frequently have no awareness of a microsleep, or reconstruct it afterwards as a lapse in concentration.

Four seconds at 70mph is over 100 metres travelled blind. That is the whole distance in which anything could have happened.

Microsleeps increase sharply with sleep debt and cluster in the circadian troughs — roughly 2am to 6am and again in the early afternoon.

Why self-assessment fails

This is the part to internalise, because it undermines the plan most people have.

Sleep deprivation impairs the metacognitive faculty you would use to notice it. Studies of restricted sleep consistently find that subjective sleepiness ratings plateau while objective performance keeps declining. People adapt to feeling tired and stop noticing it, while their reaction times continue to worsen.

So "I'll stop if I feel too tired" is not a plan. It relies on a gauge that is broken in exactly the conditions where you need it.

The warning signs you can actually use

Behavioural, external, and more reliable than how you feel:

Drifting in the lane, or hitting a rumble strip
Not remembering the last few miles
Repeated yawning, or eyes that keep closing
Missing a turn or a sign
Tailgating without meaning to, or variable speed
Head nodding

Any one of these means stop. Not "open a window." Not "turn the music up." Those do essentially nothing and their main function is to let you keep driving.

What actually helps, briefly

Pull over and nap 15 to 20 minutes. The only genuinely effective roadside intervention. Somewhere safe — services, not the hard shoulder.

Caffeine, with the nap. 150 to 200mg immediately before a 20-minute nap works well, because caffeine takes about that long to take effect and you wake as it arrives. Effective and temporary.

Then reassess honestly. These buy you perhaps an hour or two. They do not restore you.

If you are genuinely impaired, do not drive at all. Stop for the night, or get someone else to drive.

Who is at elevated risk

Anyone with untreated obstructive sleep apnea — a well-documented and substantial increase in crash risk, and one of the strongest reasons to get assessed rather than self-manage (Titan's write-up of the underlying research.) (how to decide)
Shift workers, particularly on the commute home after a night shift, which is one of the highest-risk drives there is (the full protocol)
Anyone on sedating medication — including sedating antihistamines, which are in most over-the-counter sleep aids and night-time cold remedies
Anyone running chronic sleep restriction, where the debt is invisible because it accumulated slowly (whether you can repay it)
Anyone who has had even one drink on limited sleep — the two combine worse than either alone

The disclosure obligation

Worth knowing: in many jurisdictions, a diagnosed sleep disorder that causes excessive daytime sleepiness must be declared to the licensing authority, and driving may be restricted until it is treated. Treated apnea generally restores eligibility. Untreated apnea plus a crash is a legal problem as well as a medical one.

If you snore heavily, gasp awake, have witnessed pauses in breathing, or fall asleep unintentionally during the day, get assessed. (Whether snoring always means apnea.)

The bottom line

Seventeen to nineteen hours awake is comparable to drink-driving at 0.05%, and microsleeps take you off the road for seconds at a time without your knowledge.

You cannot judge this from the inside, because the impairment degrades the judgement. Use the behavioural signs instead — drifting, lost miles, missed turns — and treat any of them as a full stop rather than a prompt to open the window.