How to assign referees fairly across a season

Almost never “we should all have the same number of games”. What people mean is closer to: I can see why I got what I got.

An official who gets fewer games than a colleague is rarely upset about the count. They are upset because they cannot tell whether it reflects their availability, their level, a judgement about their ability that nobody said out loud, or nothing at all. Uncertainty is what turns a scheduling outcome into a grievance.

Which means fairness is mostly a documentation problem. Two assignors can make identical decisions; the one who can explain them has no fairness complaints and the one who cannot has several.

Not everything competing for your attention has equal weight. Working through them in a fixed order makes assigning faster and the results more defensible.

Hard — these eliminate candidates

• Availability. Including the ceiling they gave you. Someone who said two games a weekend is not a candidate for a third, even though they are technically free.
• Level and certification. What the game requires, and what they hold.
• Conflict of interest.Their child’s team, their old club, the team they coach at another age group. Ask about this at registration; nobody volunteers it later.
• Physical possibility. A 9:00 that runs to 10:45 and an 11:00 across town is not an assignment, it is an apology waiting to be written.

Soft — these rank the survivors

• Workload so far. Both directions: who is short, and who is heading for a fifteenth consecutive Saturday.
• Development. Whether this game teaches them something, and whether they will have support in it.
• Travel. Distance, and whether the day can be built as one trip rather than three.
• Variety. The same official at the same club six times is fine for them and slowly corrosive for the club.

“Games this season” is the obvious metric and a bad one on its own. Six under-10 games at one local complex and six varsity fixtures at six different venues are the same number and nothing like the same load.

What is worth tracking separately:

• Games, by level. Two counts, not one. The competitive fixtures are what people compare themselves on.
• Days worked. Someone with three games on one Saturday has had one day out. Someone with three games across three weekends has had three.
• Distance, roughly. A tally of long trips is enough. Nobody needs mileage.
• Unsociable slots. The 8am Sunday and the Friday under lights. Track who takes them, because those are the ones people remember doing more than their share of.

The temptation is to sort officials into levels and assign strictly within them. It produces defensible schedules and a pool that never develops, because the only route upward is a promotion decision nobody is incentivised to make.

Two things work better:

• Stretch assignments with support. Put the newer official one level up, on a crew with an experienced colleague, rather than alone at a quiet game one level down. The quiet game teaches them nothing.
• Protect the difficult fixtures. Derbies, promotion deciders and anything with history are not development opportunities. Everyone in the pool understands this, and nobody resents it as long as it is the stated reason.

The failure mode to watch for is the opposite of unfairness: your three most capable officials taking every hard game, every week, until two of them quit.

Nobody sets out to give one person nine games in a fortnight. It happens because assigning is done a weekend at a time, and a weekend at a time is a window too small to see a pattern in.

Three cheap defences:

• Ask for a ceiling in advance and treat it as a hard constraint, not a preference.
• Keep the running count visible while you assign, not in a report you generate afterwards. A number you have to go and look up is a number you will not look up at 10pm on a Tuesday.
• Look at a month, not a weekend. Even a glance. The fifteen-consecutive-Saturdays problem is invisible at weekend resolution and obvious at month resolution.

Some officials confirm within the hour, arrive early, and have never cancelled. Others confirm on Friday, sometimes. It would be disingenuous to pretend this does not affect assigning, because it does, in every organization, and it should.

What matters is how it affects it.

• Reasonable: giving the important fixture — the one where a no-show is a crisis — to someone with a track record.
• Reasonable: not giving a fourth chance to someone who has cancelled on the morning three times.
• Not reasonable:quietly halving someone’s games because of a reliability judgement you have never told them about.

The line is disclosure. If reliability affects assignment — and it should — then say so as a policy, and tell the individual before their allocation changes rather than after they notice. An official who is told “you have cancelled late twice, so I am going to be careful with you for a few weeks” usually fixes it. One who is silently demoted just leaves.

Publishing every assignment decision would be exhausting and would invite negotiation on all of them. The useful amount is smaller:

• Publish the policy, once.A short note before the season saying what drives assignment — availability first, then level, then spread, and that difficult fixtures go to experienced officials. Half a page. It pre-answers most of the season’s questions.
• Keep the record, always. Who was assigned what, when they were told, and what they said. Almost every fairness dispute is settled by a record rather than an argument — and quite often the record shows the complaint was right, which is worth knowing.
• Answer the individual question directly.“Why did I only get four games in October?” deserves a real answer, and usually there is one: you were unavailable for three weekends, or you were third in line for the two varsity fixtures. Both are fine answers. “It just worked out that way” is not.

Be sceptical of anything promising to assign fairly on your behalf. Fairness in officiating depends on things that are not in any database — who is coming back from an injury, who had a rough game at that club last month, who is quietly ready for a step up. An optimizer that cannot see those will produce an allocation that is even and wrong.

What software is genuinely good at is the bookkeeping that makes your judgement possible: keeping the counts current, keeping the roster next to the games, and keeping a record of what was sent and what came back.

Ref Grid does not have automated fairness optimization, and does not assign anyone for you. It shows you what each official is already carrying while you assign, warns you when someone is double-booked, and keeps the assignment and communication history in one place so the record exists when you need it. The decisions stay yours, which is where they belong.

More on how that works: sports officials scheduling software.