New: Skyward Dream — explore Aviation Careers after 12thOpen portal →
Skip to content
Fatigue

Legal, but sleepy: are India’s pilot duty rules keeping pilots alert?

By Capt. Amit Singh FRAeS • Founder, Safety Matters Foundation

A pilot can follow every rule in the book and still be one of the sleepiest people in the sky. This is a plain-language walk through our new review of India’s pilot duty rules: what they get right, where they leave pilots tired, and what should change.

It is just after dawn. An aircraft is on final approach after a night duty that broke no rule. The pilot reported for work at 9 p.m., has been awake since the previous morning, and has flown straight through the hours when the human body most wants to sleep. Everything about the duty is legal. The question this review asks is simple: is the pilot alert?

India’s rules on pilot working hours are set by the Directorate General of Civil Aviation (DGCA) in a document called the CAR. The UN’s aviation agency, ICAO, says every country’s duty rules must be based on science, knowledge and real experience, so that pilots stay alert enough to fly safely. In 2024 DGCA rewrote the rules: night now runs from midnight to 6 a.m., a night duty may have only two landings and last at most 10 hours, and pilots get longer weekly rest. Those were real improvements. We wanted to know whether they are enough.

← Legal does not always mean rested. That is the whole story in one line.

The short answer: by day, mostly yes. At night and late in the evening, often not. A duty can follow every rule and still leave most of the crew sleepy.

Four donut charts: out of every 100 pilots, how many are sleepy as they begin the last landing. Europe, day flights, measured: 16. India, legal morning start: 12 to 21. India, legal night 21:00 to 07:00: 63. India, legal late day landing 23:30: 55 to 77.
Out of every 100 pilots, how many are sleepy as they begin the last landing. Europe’s figure is measured on real crews; India’s are predictions from a sleep model.

First, how to read “63 in 100”

Most numbers here look like “63 in 100 pilots sleepy”. Imagine 100 pilots flying exactly the same duty. People differ, so the model predicts that as they start the descent for the last landing, about 63 of them would rate themselves 7 or more on a 1-to-9 sleepiness scale. Scientists call it the Karolinska Sleepiness Scale; 7 means sleepy, 8 means fighting to stay awake.

It is a forecast for one kind of duty, like “63% chance of rain”, not a count of named Indian pilots. The yardstick throughout is Europe: when Europe’s air safety agency (EASA) measured real crews on ordinary day flights, about 16 in 100 were sleepy.

A weather forecast, not a head count. 16 in 100 is our yardstick.
Sketch of 100 pilot faces: 63 sleepy, 37 less sleepy, on a legal first night after days off. Annotation: this does not mean every pilot is unsafe; it means the risk is much higher.
Sketch for illustration. A legal 21:00 to 07:00 night with two landings.

Why nights are hard: a battery and a tide

Two forces drive tiredness. From the moment you wake, sleep pressure builds like a phone battery draining; only sleep recharges it. Separately, the body clock raises and lowers alertness like a tide, lowest between about 2 and 6 a.m., however well you slept.

Night flying combines a nearly flat battery with the lowest tide, at the same time. That is why the last landing of a night duty is the sleepiest moment of the job. And when sleep is cut short night after night, the shortfall carries over, like a debt.

Flat battery + low tide = the last landing of the night.
Sketch: a battery draining from 07:00 wake-up to 04:00 after 21 hours awake, and a tide-like body-clock curve that dips lowest between 2 and 6 a.m.

1. Nights: the biggest risk

A pilot who reports at 9 p.m. for a legal 10-hour night has usually been awake since morning and flies through the body clock’s low point. As they start the descent for the last landing, about 63 in 100 are predicted to be sleepy, and 40 in 100 at 8 or more, which means fighting to stay awake. That is about four times what Europe measured on day flights.

The rules allow two nights in a row (three once a week). Pilots sleep in the day between them, but daytime sleep is shorter and lighter, so the shortfall builds: 62 in 100 on the second night and 66 in 100 on the third. A 90-minute nap before the first night helps a little (55 in 100), but nothing in the rules requires one.

Key message: nights in a row do not get easier.

Did you know? Operators have asked DGCA to allow a third landing at night. The model predicts that would raise the share of sleepy pilots from about 63 in 100 to about 74 in 100.

2. Late evenings: the hidden risk

A duty that ends just before midnight counts as a day duty, so it may include up to six landings. Yet by the last landing the pilot has been awake 16 to 20 hours and is heading into the body clock’s low point. On a legal day landing at 11:30 p.m., 55 in 100 pilots are predicted to be sleepy with four landings, and 77 in 100 on a six-landing turboprop day.

Real records say the same thing. Across 452 working days flown by five Indian pilots, legal days ending between 8 p.m. and midnight left 25 in 100 crew sleepy on average, against 12 in 100 for days ending earlier. A four-landing day ending after 8 p.m. (36 in 100) was sleepier than the average night duty (32 in 100). Every one of the five pilots flew such days.

An 11:30 p.m. finish is still called a “day duty”. The body does not care about that label.

In plain words. The rules draw a hard line at midnight. The body does not: a day that ends at 11:30 p.m. is nearly as tiring as one that ends at 1 a.m., but it can carry three times as many landings.

It is the hour a pilot starts work that matters most. Run the longest legal two-flight duty from each of the 24 possible sign-on hours, and 9 of them leave more than half the crew sleepy. Only an early-morning start keeps it near Europe’s daytime level.

A 24-hour ring: each slice is a sign-on hour for the longest legal two-flight duty. Nine of 24 sign-on hours leave 50 or more in 100 crew sleepy; only early-morning sign-ons stay under 25 in 100.

3. Early starts: the plane leaves after the low, the pilot does not

For a 6 a.m. departure, Indian pilots told our 2022 survey they wake at about 3 a.m. They shower, drive to the airport and report at about 5 a.m. Three of the four hours of the body’s low point are spent awake.

Going to bed early does not fix it: trying to fall asleep at 9 p.m. is like trying to eat a big dinner straight after lunch. Pilots measured with wrist monitors slept about 5½ hours before starts between 4 and 5 a.m., over an hour less than before a 9 a.m. start.

Now repeat it. Getting up at 2:30 a.m. for five 6 a.m. departures with four flights a day, the share of sleepy pilots climbs every day, from 31 in 100 on day 1 to 55 in 100 on day 5, as about 19 hours of sleep debt piles up. India’s rules already stop that exact run. But a run of 7 a.m. departures is allowed with no limit.

The flight leaves after 6 a.m. The pilot’s day began in the middle of the night.
A 24-hour clock for a 6 a.m. departure: wake at 03:00, drive, report at 05:00, depart 06:00. The pilot is awake for 3 of the 4 hours of the 2 to 6 a.m. body-clock low.
Times from the SMF 2022 pilot survey, a published wrist-monitor study of airline pilots (Roach et al. 2012), and drive times to Indian airports. Not the roster of any named pilot.

In plain words. Sleep debt works like a credit card. Paying the minimum each night keeps the statement looking the same for a while, but the balance keeps growing, and the pilot is the last to notice. In laboratory studies, people allowed 4 hours a night for two weeks made as many attention slips as people kept awake for two whole nights, yet said they felt only a little sleepy.

What about long days?

If mornings are fine, is a 13-hour day fine too? Not quite, because two different risks are at work. The clock makes nights and late evenings sleepy. The length of the duty wears pilots down even in daylight: workload, sustained attention, and no margin left when the day runs late.

Our sleepiness model mostly measures the first. A 13-hour day starting at 6 a.m. scores low on sleepiness. But NASA’s 1996 guidance set 10 hours as the standard duty, the scientific evaluation for Europe’s regulator (2008) said 13 hours should be the exception, not the routine, and a US accident study (Goode, 2003), though resting on few accidents, found accidents over-represented in very long duties. India allows 13 hours as a routine day for any start up to 10:30 a.m., with no extra safeguard. 70 in 100 Indian pilots in our 2024 survey said duties over 10 hours deeply worsen fatigue.

Two risks: when you fly, and how long you fly.

4. What real rosters show

Thirteen pilots shared their real rosters and logbooks. We removed all names, operators, flight numbers and airports before analysis. Before the 2024 rules, 55 in 100 duties left at least a quarter of the crew sleepy. Under the 2024 rules, that falls to 42 in 100. The rules helped, but 42 in 100 duties are still well above what Europe measured on ordinary day flights.

Most duties stayed inside the rules. A few went past a limit, mostly by minutes around midnight or 6 a.m. Some long-haul night flights ran past the normal 8-hour night flying limit under an exemption for which no fatigue evidence has been published.

These are real duty records, not just a theoretical model.

Europe’s pilots, this month

A survey of 7,198 European pilots, published in October 2026 for the European Cockpit Association, found that 77 in 100 had at least one microsleep, head-nod or fight to keep their eyes open on the flight deck in the past four weeks, and 52 in 100 had flown in the past three months when, looking back, they should have reported unfit. Europe already requires airlines to actively manage the fatigue of night and late-finish duties, and Europe’s regulator now says “being compliant does not necessarily mean safe”.

Source: A fatigue survey of European Pilots, 2026 (Baines Simmons for the European Cockpit Association), quoting EASA. Its questions differ from our Indian surveys, so the figures are parallels, not a like-for-like comparison.

Different rules, the same warning: legal is not the same as rested.

Three facts that say it plainly

No model is needed for these. They come from the law, from pilots themselves and from India’s own investigators.

Zero alcohol allowed. Not zero fatigue.

India grounds a pilot for any trace of alcohol and suspends the licence above 0.009%. But after 17 to 19 hours awake, people perform about as badly on some tests as at 0.05% blood alcohol (Williamson and Feyer, 2000): more than five times the level that costs a pilot their licence. A legal late-evening day can keep a pilot awake that long by the last landing.

Tired before the next flight even starts.

In our 2022 survey of 542 Indian pilots, 43 in 100 said they were already fatigued for their next flight. 71 in 100 said they had often been too tired for cockpit duty, 66 in 100 had fallen asleep unplanned or had microsleeps on duty, and 97 in 100 said DGCA is not doing enough.

Mangalore, 2010: 158 people killed.

The aircraft was landing at dawn after a night return flight. The Court of Inquiry found the captain had slept for much of the flight, and that waking just before landing may have impaired his judgment, made worse by flying in the body clock’s low point.

A pilot with a trace of alcohol is stopped at the door. A pilot impaired by the roster is not even asked.

5. When the rules were switched off

In December 2025, five weeks after the new night rules took full effect, DGCA suspended the most important ones for one large operator after it cancelled flights on a massive scale, until 10 February 2026. Every kind of duty that suspension allowed leaves more pilots sleepy than a normal day flight. The government told Parliament that operators’ safety risk assessments were checked by DGCA before any relaxation was allowed, but none has been published, so nobody outside can see what fatigue evidence was used.

Operators now want to manage fatigue with their own data-driven systems (FRMS) instead of fixed limits.

If the evidence is not public, outsiders cannot check it.

Why it matters. ICAO’s audit of India gave a good overall score, about 86 out of 100. But its lowest mark, about 51 out of 100, was for having enough qualified technical staff: the very inspectors who would have to check an operator’s fatigue data. Allowing FRMS now would be like letting students mark their own exams before there are enough teachers to check the marking.

What this means if you fly

  • This does not say flying in India is unsafe. Most duties keep pilots reasonably alert, and pilots, procedures and the second pilot in the cockpit all add layers of protection.
  • The thin margins are on flights that land late at night or in the early morning, after a long day, a run of nights or a run of very early starts.
  • Higher risk is not the same as unsafe. It means the safety margin is thinner than it needs to be, and that is what the rules should fix.
  • The fix is in the rules, not in passengers’ hands: duty limits that look at the clock, and a real chance to sleep before every duty.

What should change

Fix the rules first, collect Indian evidence next, and only then consider operator-run fatigue systems. That is the order ICAO’s own guidance sets.

  1. Set duty limits by the time a duty starts, as Europe, the US and Australia do, not by one fixed “night” band.
  2. Limit late-evening days that end close to midnight, for example to four landings.
  3. Protect runs of nights and early mornings, where sleep debt builds.
  4. Guarantee an 8-hour chance to sleep in every rest period.
  5. Count jet lag the way the body does: gradually, not as a 48-hour on/off switch.
  6. Measure Indian pilots’ fatigue the same way everywhere, and run an independent study of real pilots.
  7. Any exemption must come with fatigue evidence and an end date.
Fix the rules → measure in India → only then FRMS.

Does India pass the international test?

Partly. On science: partly, because the night rules follow it but late evenings and runs of early starts do not. On knowledge: partly, because India is as strict as other countries on some rules but not others. On real experience: not shown, because no measurements of how tired Indian pilots actually are have been published. And the aim, alert pilots, is not met at night.

How this was done, and how sure we are

The numbers come from a well-established sleep-and-body-clock model, tuned on laboratory sleep-loss studies and adjusted to match what Europe measured in real working pilots. It also carries sleep debt from one day to the next. It was run on duties at India’s legal limits and on thirteen pilots’ real rosters and logbooks. These are predictions for an average crew, not measurements of Indian pilots; individuals vary. We tested the model 13 different ways, and nights and late evenings came out well above Europe’s day level in every test. Pilots told us the same thing first, in our national surveys of 542 pilots in 2022 and 530 in 2024; the model reached the same answer independently.

Read the full guide.Download the 15-page illustrated guide, Are India’s Pilot Duty Rules Keeping Pilots Alert? (PDF, 6.6 MB). The full evidence, with methods, statistics and references, is in the companion report, Science-Based Review of India’s Flight Duty Time Limitations, available on request.

Capt. Amit Singh FRAeS is a senior commercial pilot and founder of Safety Matters Foundation, an independent, not-for-profit aviation safety organisation in India. No one commissioned this review and it received no external funding. It is not legal advice and does not accuse any operator, pilot or official. Sleepiness figures are model predictions for an average crew, rounded. Spotted a mistake? Write to admin@safetymatters.co.in; confirmed corrections will be acknowledged and published.


Discover more from Safety Matters Foundation

Subscribe to get the latest posts sent to your email.

Discover more from Safety Matters Foundation

Subscribe now to keep reading and get access to the full archive.

Continue reading