Seventeen Weeks Holds 2 856 Hours — the Limit Only Lets You Work 816
Working-time law almost never talks about a single week. It sets an average and then names a stretch of calendar to average it over, which is why the argument in an HR meeting is rarely about last week's timesheet and almost always about the block of weeks around it. Before anyone can say whether a limit has been breached, the reference period has to be turned into hours: the calendar hours it physically contains, and the far smaller number of those hours that may be spent working.
What the Regulations Are Counting
The 48 hours is an average, not a ceiling
The window can be stretched
Not every hour on site is working time
Rest rules usually bite first
Sizing a Reference Period Before You Average Anything
Two numbers make an averaging argument: how long the window is, and how many hours were worked inside it. This page supplies the first one and puts it beside the second.
Enter the reference period in weeks
Type 17, 26 or 52 into the left field and the calendar hours appear at once. Spaces are ignored and a comma works as a decimal point, so a part-period such as 8,5 weeks can be entered exactly as it was written down.
Set the working-hours ceiling beside it
Multiply the same week count by 48 to get the maximum permitted working hours — 816 over 17 weeks, 1 248 over 26, 2 496 over 52. Compare that with the total on the timesheets, not with any single week.
Reverse it when you start from an hours total
The swap button (↔) turns the page into h → wk, which is what you want when a payroll export gives 2 184 hours and the question is how many weeks of coverage that represents.
Copy the plain figure into the working-time record
The copy control on each field gives the digits with no unit and no spacing, ready for the column of a monitoring spreadsheet that an employer has to keep. Ctrl + C inside a field behaves the same way.
Reference Periods, Rest Rules and the Hours They Leave
Each rule against the stretch of calendar it applies to, the raw hours that stretch contains at 168 per week, and the working-hours limit it imposes inside them. The middle column is what the converter gives you; the right-hand column is what the law does with it.
| Rule | Period (weeks) | Calendar hours | Working-hours limit it sets |
|---|---|---|---|
| 48-hour average week, default period | 17 | 2 856 h | 816 h of working time |
| Special-case roles (continuity of service, security) | 26 | 4 368 h | 1 248 h of working time |
| Extended by collective agreement | 52 | 8 736 h | 2 496 h of working time |
| Night work, 8-hour average per 24 h | 17 | 2 856 h | 952 h of night work |
| Weekly rest, 24 uninterrupted hours | 1 | 168 h | 144 h remain available |
| Fortnightly alternative, 48 uninterrupted hours | 2 | 336 h | 288 h remain available |
| Daily rest, 11 consecutive hours in each 24 | 1 | 168 h | 91 h remain available |
The last three rows are the ones people forget. Daily rest alone removes 77 hours from every week, leaving 91 — so a worker who has signed an opt-out and has no average to answer to is still nowhere near able to work 168 hours, and a roster that looks legal on the average can fail on rest. Note also how the night-work figure is built: 8 hours across each of the 119 days in a 17-week period is 952 hours, averaged the same way as the 48-hour rule but on a per-24-hours basis rather than per week.
Handling the Numbers in a Working-Time Audit
A reference period sized in one keystroke
Type the week count and the calendar hours appear immediately, so 17, 26 and 52 can be compared side by side while an agreement is still being drafted.
Timesheet totals run backwards
Swap the pair and an accumulated hours figure from payroll turns into the number of weeks it covers, which is the form a monitoring record wants.
Fortnights are a unit in their own right
The searchable dropdowns include a fortnight of 14 days alongside days, months and quarters — handy when the rest rule you are checking is written per two weeks rather than per week.
Four-figure totals stay legible
Thousands are separated by a space in the result, so 8 736 hours across a year-long averaging period does not turn into an unreadable run of digits.
Questions About Averaging and Opt-Outs
What exactly does the 17-week reference period average?
Working time, summed across the period and divided by the number of weeks in it. The period rolls: it is not seventeen fixed blocks starting in January but the most recent seventeen weeks at any moment, so the average is recalculated continuously and a breach can appear or disappear as old weeks drop out of the window. Seventeen weeks of calendar is 2 856 hours; the 48-hour average caps working time inside it at 816.
I worked 60 hours last week — is that a breach on its own?
No. The limit lives on the average, and one week of 60 leaves 756 hours for the remaining sixteen weeks of a 17-week window, which is 47.25 a week — still inside. Four such weeks would leave 576 hours across thirteen, or 44.3 a week, which is also fine. What matters is whether the running total is on course, and the employer is the party required to keep records adequate to show it.
Does time spent on call count towards the average?
It depends where you have to be. On-call time at the workplace generally counts in full even when nothing happens, because the worker is at the employer's disposal and cannot use the time freely. On-call from home is usually counted only for the periods actually worked, unless the constraints are so tight — a very short response time, a required location — that the whole stretch is effectively unavailable to the worker. The distinction turns on how much freedom the worker really has, not on whether a payment was made.
How is the average worked out if I was off sick or on leave?
Those days are excluded rather than counted as zeros, which would otherwise drag the average down and mask a genuinely heavy schedule. For each excluded day the calculation borrows a working day from immediately after the reference period ends and uses its hours instead. The arithmetic is therefore total working time, plus the hours from the borrowed days, divided by the number of weeks in the period — which is why a plain spreadsheet average over seventeen rows will not match the legal figure.
Can an opt-out be withdrawn once it has been signed?
Yes. An opt-out from the 48-hour average has to be in writing, individual and voluntary, and it can be cancelled by the worker with at least seven days' notice unless the agreement itself specifies longer — and even then the notice cannot exceed three months. Signing away the average changes nothing about the rest entitlements or the separate night-work limit, and a worker cannot be forced into an opt-out as a condition of the job.
No comments yet. Be the first to comment!