A watch roster should be built around protected sleep, because people who are short of sleep watch badly, judge badly, quarrel and have accidents, and the quietest hours of the night are both the hardest to stay awake through and the ones an intruder may well pick. Count the longest unbroken stretch each roster leaves between watches: 4 hours for 3 people sharing one post, 6 hours for 4, 8 hours for 5, and compare it with the 7 to 8 hours of sleep the survival articles aim for.
Part of Watch and Patrol Scheduling
Sleep cycle management is the planning of rest around watch duty so that the people on the roster stay alert, safe and fit to make decisions.
What Is Known, What Is Recalled, And What This Article Will Not Say
This article gives no maximum safe watch length and no “vigilance span”. The parent article says 2-hour shifts are optimal and 4 hours the maximum; this corpus has no source for either, and I have not been able to confirm a figure I could stand behind. What I recall from sleep research, and flag as recalled and unchecked:
- Alertness follows a daily rhythm with its trough in the small hours, roughly 0300 to 0500 for someone on a normal schedule, about two hours before they would naturally wake.
- Sleep runs in cycles of roughly 90 minutes, and waking from the deeper part of a cycle leaves grogginess (sleep inertia) that can last from several minutes to over half an hour.
- Lapses of a few seconds (“microsleeps”) occur in tired people without the person noticing.
- In classic laboratory experiments on watching a monotonous display, detection of rare signals fell off early, within the first half hour. If that holds, shift length alone cannot fix watch quality, and the answer is variety, pairing and sleep.
The corpus’s own figure for sleep is 7 to 8 hours (Rest and Sleep Discipline), which I use as the benchmark.
Worked Example: Rest Left By Each Roster
One post watched round the clock on 2-hour slots, with N people sharing it evenly (each person’s watches equally spaced through the day):
| People N | Slots per person per day | Watch-hours per person | Longest off-watch stretch |
|---|---|---|---|
| 3 | 4 | 8 | 24 / 4 - 2 = 4 h |
| 4 | 3 | 6 | 24 / 3 - 2 = 6 h |
| 5 | 2.4 (round robin) | 4.8 | 8 h (see Rotating Schedules) |
| 6 | 2 | 4 | 24 / 2 - 2 = 10 h |
The off-watch stretch must hold the walk to bed, falling asleep, waking before the watch and the 5-minute early arrival at handover. Assuming 1 hour for all of that (an example, not a measurement):
- 3 people: 4 - 1 = 3 hours of sleep at most.
- 4 people: 6 - 1 = 5 hours.
- 5 people: 8 - 1 = 7 hours, at a different clock time each day.
- 6 people: 10 - 1 = 9 hours.
Only the larger rosters meet the 7 to 8 hour benchmark. The parent fatigue article treats two hours on, four hours off as sustainable (Fatigue Management); that is the 3-person row, and it leaves 3 hours of sleep. It is sustainable for a few days, not for weeks.
Clustering a person’s watches into one window gives them one long off stretch instead of several short ones, but it makes some watches longer than 2 hours and puts some people to sleep by day (see the clustered example in Rotating Schedules).
Fragmented Sleep Costs More Than The Hours Removed
Eight hours broken into two four-hour pieces is not the same as one eight-hour block. A person woken from the middle of a cycle for a watch loses the deepest sleep, then stands the watch in the grogginess of waking, then goes back to bed unable to settle. The tiredness the roster causes shows first in the quality of the watch: missed log entries, dull replies at handover, and the sentry who sits down “for a moment”.
Method: Building Sleep In
Step 1: Schedule sleep first. Mark each person’s protected block, then fit watches into the rest.
Step 2: Give the night sleeper a real night. No chores, no non-urgent waking, a dark quiet room; give the night watcher a dark quiet place by day too (Sleep Hygiene Without Electricity).
Step 3: Anchor the clock with morning light (Sunlight and Circadian Rhythm).
Step 4: Use naps as a supplement, following the nap rules (Nap Strategy). A person coming off a night watch sleeps when they can.
Step 5: Cut hours, not rest. Move to a lower threat posture or add mechanical and animal warning at night (Early Warning Systems).
Step 6: Watch for the signs in the group: missed or copied log entries, irritability, startling at normal noise (Hypervigilance and Startle Response), accidents. If two appear in the same person, take them off the roster for a few days.
Diagnosing A Failing Roster
- Sentry found asleep. The parent’s rule applies: fix the system first. Count their sleep in the last three days.
- Handover goes badly. The incoming sentry was woken minutes before; allow a longer wake-up period and a read-back.
- The same person always has the small hours. Rotate, within the limits of what the body clock will bear.
- Quarrels among the watchers. Look at the schedule before the personalities.
Common Mistakes
- Counting watch-hours, not sleep-hours. A balanced watch load can still be a wrecked sleep load.
- A “relief” watcher who is also on the day’s work crew. The tiredness adds up.
- Using stimulants or noise to mask fatigue. The parent fatigue article notes it hides tiredness and ends in a crash.
- Treating dozing as bad character. It is an expected result of the rota.
Safety
A drowsy sentry near height, fire or tools
A sleepy watcher on a platform can fall; one beside a stove or banked fire can be burned or breathe carbon monoxide; one using a blade or a firearm can injure themselves or another. Do not post a person who has had too little sleep at a height or beside a fire, and use the lower posture instead.
What’s Next
- Learn next: Rotating Schedules — the arithmetic of who stands which slot
- Combine with: Handover Procedures — waking and briefing the incoming watcher