Why it's fair
A method with proofs, not promises
Prefera doesn't invent its own rules for who deserves what. It applies your priorities through a matching method that economists have studied for sixty years, and shows its working.
The method
Deferred acceptance
David Gale and Lloyd Shapley published deferred acceptance in 1962. Alvin Roth later used it to redesign real matching markets, and in 2012 Roth and Shapley shared the Nobel Prize in economics for this work.
The idea is simple: nobody's place is final until everybody's is. People apply to their choices in order, and each option keeps provisionally the best applicants by priority.
Round 1
Everyone applies to their 1st choice. Each option holds on to its highest-priority applicants, up to its number of places, and turns the rest away.
Round 2
Anyone turned away applies to their next choice. Each option compares new applicants with the ones it is holding and again keeps the highest priority.
Until settled
Rounds continue until nobody is turned away or people run out of choices. Only then are places confirmed, so nobody gains by applying early.
Proven guarantees
Three things every participant can rely on
No justified envy
If you didn't get an option you ranked higher, every place in it went to someone with higher priority than you. Nobody jumps the queue.
Honesty is safe
Your best strategy is to rank what you really want, in the order you want it. No misreport can get you a better result.
Best fair outcome
No place is wasted, and each person gets the best result that any allocation without justified envy could give them.
US medical residency match
Places tens of thousands of doctors a year
New York City high schools
About 90,000 students a year
Amsterdam secondary schools
7,650 pupils matched in 2025
Several placements
Rotations: priority order with a guarantee for everyone after you
When each person does several placements with required specialties, Prefera goes through people in priority order. Each gets the best timetable for their ranking that still leaves everyone after them able to meet the rules.
Still strategy-proof
Your ranking only affects your own turn, so ranking honestly is still your best move.
Exact, not approximate
Each step is checked with an exact solver. A poll whose rules can't all be met refuses to open, so nobody is left without a valid timetable.
Your part
Fair method, fair inputs
The method applies your priorities exactly. Whether those priorities are fair is the organiser's decision, so Prefera makes them visible.
Ties settled by a recorded draw
Equal priority is decided by a random draw by default, or by earliest submission if the organiser chooses it before the poll opens. The draw's seed is saved with the run.
A person signs off
Results are reviewed by a manager before release, and every participant is told who to contact if they want their result looked at again.
Anyone can verify
The run report re-runs the allocation from the saved inputs and confirms it matches. The verification file lets an independent auditor do the same.
Tested on every change
Each version of Prefera re-checks the guarantees on hundreds of random polls before it ships.
Want an independent look at a past allocation?
We can re-run a previous rotation and report whether it was stable and consistent.