How it works, precisely
An order is fixed. Drawing lots is the sensible default, because any other order needs justifying — seniority, need and merit are all legitimate, but they are political decisions.
Each person in turn takes their preferred option among those still available. A single pass is enough; the result is immediate and can be checked line by line.
Two proven properties: the outcome is Pareto-efficient (no rearrangement can help someone without hurting someone else) and the mechanism is strategy-proof (declaring your true preferences is always optimal).
A worked example
Four people, four assignments, order drawn at random: Chloé, Ali, Bruno, Dana.
| Person | 1st choice | 2nd choice | Gets |
|---|---|---|---|
| Chloé | Audit | Redesign | Audit |
| Ali | Audit | Support | Support |
| Bruno | Redesign | Audit | Redesign |
| Dana | Support | Training | Training |
- Chloé goes first and takes Audit, her first choice.
- Ali wanted Audit: no longer available, he takes Support, his second choice.
- Bruno gets Redesign, his first choice; Training is left for Dana.
Three of four people get their first choice. The quality of the result depends entirely on the draw — which is why the order must be announced and verifiable before the assignment.
Where it comes from
The procedure is as old as sharing, but social choice theory formalised it as « serial dictatorship »: at each step one person decides alone — hence the term, which describes the algorithm and not a regime.
It structures North American sports drafts, where the order is reversed from the standings to rebalance the teams; the draft lottery, introduced by the NBA in 1985, added chance to discourage deliberate losing.
Economists studied it as the house allocation problem, posed by Hylland and Zeckhauser in 1979. The proof that it is one of the rare mechanisms both efficient and immune to manipulation came later, with Abdulkadiroğlu and Sönmez (1998) and then Svensson (1999).
Where it is used
- Sharing out assignments, slots, desks or equipment within a team.
- Sports drafts and player selection.
- Allocating rooms in halls of residence or shared housing.
- Choosing internship or dissertation topics among students.
Limits and pitfalls
- Impossible to game: ranking sincerely is always best
- Very easy to explain and accept
- Verifiable draw, no favouritism possible
- The last in line inherit the leftovers
- Luck weighs heavily on the outcome
- Ignores preference intensity
Frequently asked questions
Should the order be drawn at random?
It is the most defensible default, because it requires no justification. Any other order — seniority, need, merit — is a judgement call to own publicly before the assignment, never after.
Is there any point in lying about my preferences?
No, never, and that is proved. When your turn comes you take the best remaining option: declaring anything else can only hurt you. This is what makes the mechanism strategy-proof.
Serial dictatorship or maximum satisfaction?
Serial dictatorship is transparent and incontestable but depends on the draw. Maximum satisfaction optimises the group total, at the cost of a computation nobody can redo in their head. Choose according to what the group must accept: clarity or the optimum.
Sources
The primary references this page relies on.
- Hylland, Aanund et Zeckhauser, Richard, The Efficient Allocation of Individuals to Positions, Journal of Political Economy, 87(2), 293-314, 1979. DOI ↗
- Abdulkadiroğlu, Atila et Sönmez, Tayfun, Random Serial Dictatorship and the Core from Random Endowments in House Allocation Problems, Econometrica, 66(3), 689-701, 1998. DOI ↗
- Svensson, Lars-Gunnar, Strategy-proof allocation of indivisible goods, Social Choice and Welfare, 16(4), 557-567, 1999. DOI ↗