Condorcet methods

Ballots rank candidates, and every pair is compared: how many voters prefer A to B, and B to A. A candidate who beats all others in these pairwise contests is the Condorcet winner, and a Condorcet method always elects that candidate.

The marquis de Condorcet published the idea in 1785 and observed that majority preferences can form a cycle (A beats B, B beats C, C beats A), so no Condorcet winner need exist. Methods differ in how they resolve cycles; well-known ones include Copeland, minimax, ranked pairs and the Schulze method. Instant-runoff and Borda counts can elect someone other than the Condorcet winner.

• The Debian Project's constitution (§A.5) counts ranked ballots by pairwise defeats, finds the Schwartz set and drops the weakest defeats until a winner emerges. It governs general resolutions and project leader elections; a 2003 amendment vote proposed this clone-proof rule.
• Public elections rarely use Condorcet methods.

Counts are harder to explain than plurality or instant-runoff. Pairwise comparison ignores intensity of preference, some methods are open to strategic "burying" of a rival, and how often cycles occur in mass elections is debated.

Mostly confined to small and online organisations, where members are comfortable with ranked ballots. No large government has tested them at scale.