← Gallery

About

Random Serial Dictatorship assigns objects fairly by randomizing the picking order: shuffle students, then each picks their top remaining choice in sequence. Simple, strategy-proof, but potentially unfair to those who draw late positions.

Algorithm Pseudocode 
1. Draw a random permutation of
   all students (the "priority order")
2. The first student picks their
   most-preferred available school
3. The second student picks their
   most-preferred remaining school
4. Continue until all assigned

Repeat many times to compute
expected probabilities.

Preferences

StudentPreference Order
s1A > B > C > D > E
s2B > A > C > D > E
s3A > C > B > D > E
s4C > B > A > E > D
s5B > C > A > D > E

Controls

Properties

✓ Ex-post efficient: each realization is Pareto-optimal
✓ Strategy-proof: no student benefits from lying
✗ Ex-ante unfair: luck of the draw determines outcomes
✗ Not ordinally efficient in expectation
Compare with Probabilistic Serial: PS is ordinally efficient and envy-free, but not strategy-proof. RSD is strategy-proof but less fair. This is a fundamental trade-off in mechanism design.

Event Log

Random Serial Dictatorship

Fair assignment through randomized priority — 5 students, 5 schools

Priority Order (Shuffled)

Monte Carlo Simulation — Probability Distribution

Run 1000x to see probability distributions