Blog
What a Fair Rota Actually Looks Like
Fairness in a rota is about totals over months, not about who gets the bathroom this week.
Ask a household why their rota collapsed and nobody says the allocation was biased. They say it stopped being on the fridge, or that somebody decided the kitchen did not count as done.
Rotas fail socially far more often than they fail mathematically, and almost all the writing about them addresses the mathematics.
The three real causes of death
First, invisibility. A rota that lives in a group chat has a useful life of about four days, after which nobody scrolls back and it has effectively ceased to exist. A printed sheet on a wall survives for months for no reason other than that it is in the room.
Second, ambiguity about completion. "Kitchen" means one thing to the person who wiped the surfaces and another to the person who came in afterwards, and a rota with ambiguous jobs produces arguments about whether a turn happened rather than about who has it.
Third, a single unresolved swap. The moment one person covers for another without it being recorded, the rota and reality have diverged, and every subsequent complaint is unresolvable because the two parties are looking at different ledgers.
- Print it and put it where the work happens.
- Define completion for each job in six words or fewer.
- Write swaps on the sheet, or do not allow them.
What people actually mean by fair
They do not mean unbiased. Nobody has ever complained that a rota was generated with a subtly non-uniform draw, and if you offered them a provably uniform allocation with lopsided totals they would refuse it.
They mean roughly three things: that the totals come out similar over a period they can remember, that the unpleasant jobs move around, and that nobody has quietly opted out.
Only the first of those is a property of the allocation. The other two are properties of the job list and of enforcement, which is why a better generator is rarely the fix for a rota that is failing.
The unpleasantness problem
Every household has one job that is worse than the others, and a rota that counts all jobs as one unit is measuring something nobody cares about. Four turns at the bins and four at the bathroom are not equivalent, and the person on the receiving end knows.
Splitting the bad job so it appears more often — and therefore rotates faster — is the fix that requires no argument. Weighting the draw against whoever did it recently also works and requires somebody to maintain the weights, which in practice means the person who was already keeping score.
Drawing it separately each week is the most honest option, because it makes the unpleasantness explicit rather than burying it in a grid. It also gives the household a moment to notice it, which is worth more than the fairness.
Why the mathematical fix is still worth having
None of the above is an argument against generating the rota. It is an argument about what generating it buys, which is narrower than people assume: it removes the suspicion that whoever built the grid arranged it to suit themselves.
That suspicion is real and it is corrosive, and it is almost impossible to disprove once it exists. A generated rota with visible totals answers it permanently and costs about two minutes.
It just does not answer any of the three things that actually kill rotas, and a household that fixes the allocation and none of the rest will be having this conversation again in six weeks.
The person who quietly stops participating
The failure nobody writes about is not an argument. It is one person gradually doing less, nobody raising it, and the rota continuing to list them.
This is corrosive precisely because it produces no event to respond to. There is no week where somebody refused; there is a slow drift where their turns get done late, or partially, or by somebody else who did not mention it. By the time it is raised it has been happening for two months and the conversation is about the accumulated total rather than about this week.
A rota with a visible completion record — a tick, a date, anything — surfaces it in week two instead of month three. That is not a fairness mechanism and it is the single change most likely to keep a household rota alive, which is a slightly deflating conclusion for a page about allocation algorithms.
Frequently asked questions
Why do most rotas fail?
Invisibility, ambiguity about what counts as done, and unrecorded swaps. Almost never a problem with the allocation.
How long does a rota in a group chat last?
About four days. After that nobody scrolls back and it has effectively stopped existing.
What do people mean by a fair rota?
Similar totals over a period they remember, unpleasant jobs moving around, and nobody quietly opting out. Only the first is about the allocation.
How should the worst job be handled?
List it more often so it rotates faster, or draw it separately each week so the unpleasantness is explicit rather than buried in a grid.
Do swaps need recording?
Yes, or the rota and reality diverge and every later complaint becomes unresolvable because people are reading different ledgers.
Is generating the rota worth it at all?
Yes, for one narrow reason: it removes the suspicion that whoever built the grid arranged it to suit themselves. That suspicion is otherwise unanswerable.