"We run six on, two off" is one of those phrases everybody on a ward or in a warehouse understands, and which produces two different calendars when you ask two different people. The notation says less than it appears to: it describes the rhythm of work and rest, not who is standing there at six in the morning.

It is worth setting out what it actually means, because the choice of pattern decides far less than people think, and the direction of rotation far more.

What the notation means

The first number is consecutive working days, the second is rest days. So:

  • 5x2: five days on, two off. The classic week, with rest always landing on the same days;
  • 6x2: six on, two off. The cycle runs eight days, so the rest days drift through the week: someone off Monday and Tuesday this time will be off Wednesday and Thursday next time round;
  • 4x2: four on, two off, a six-day cycle. This one drifts too, and faster.

The drift is the whole point and the reason these patterns exist: a cycle that is not a multiple of seven covers every day of the week with the same number of people, weekends included, without anyone having to decide who works Sunday each time.

The downside is that a personal calendar becomes impossible to hold in your head. Nobody can answer "are you free on the 14th?" without checking, and that has real effects on life outside work.

How many people a post really needs

One calculation settles a lot of arguments. Covering a single post around the clock, seven days a week, is 168 hours a week. In eight-hour shifts that is twenty-one shifts.

On a 40-hour contract that needs 4.2 full-time people for that one post, and the 4.2 is before holiday, sickness and training. Anyone trying to cover a continuous post with four people does not have a pattern problem, they have a headcount problem, and no rotation fixes it.

It is worth doing this sum before choosing the pattern. Almost every rotation that "does not work" is a sound rotation applied to a rota that is short of people.

Direction matters more than the pattern

Here is the part that gets skipped. When shifts rotate between mornings, afternoons and nights, the direction of rotation makes a concrete difference to fatigue.

Forward rotation (mornings, then afternoons, then nights) follows the direction the body clock shifts most easily, and it is the one shift-work research points to as better tolerated. Backward rotation (nights, then afternoons, then mornings) demands the opposite adjustment, and produces more fatigue for the same hours.

The second element is the quick return: finishing an afternoon at 22:00 and starting again at 06:00. On paper they are two different shifts on two different days; in practice it is an eight-hour gap, under the threshold for rest between shifts. It is the join that makes an otherwise healthy pattern unbearable, and it is nearly always accidental.

How to choose

Four questions, in order:

How many hours do I actually have to cover? Not "we are open eight till eight", but how many people are needed in each slot. That gives you the number of shifts, and only then the pattern.

Should the cycle drift or not? If the business is closed on Sundays, a drifting pattern complicates life and gives nothing back. If it trades seven days, a fixed pattern forces a manual decision every single week.

How long can the run be? Six consecutive days is a lot, and it has to be read together with the rest that follows: that rest needs to satisfy the rules on weekly rest, not just be long enough to feel fair.

How do the nights fall? If the pattern hands them to the same people every cycle, you have a fairness problem built into the structure. The limits and thresholds are in night work.

The pattern is not the plan

The last point, and the most useful one. A rotation pattern gives you the rhythm, not the plan: it knows nothing about who holds a skill nobody else has, who has asked for leave, who cannot do Thursdays. In practice every real week is the pattern plus a dozen exceptions, and it is the exceptions that eat the time.

Anyone spending Sunday evening fitting those exceptions together has not picked the wrong pattern. They are doing by hand a job that can be described as a set of constraints.

In Sked Solve the rotation pattern is the starting point, and the exceptions (skills, availability, leave, nights already worked) are constraints the plan resolves on its own. See how it works at Sked Solve.