Condorcet Winner — Definition, Formula & Examples
A Condorcet winner is the candidate who would win a one-on-one majority vote against every other candidate. If you ran every possible two-candidate race, this candidate would win all of them.
Let be a finite set of candidates, and let each voter hold a strict ranking of . For , let be the number of voters who rank above . A candidate is the Condorcet winner if for every with . A profile has at most one Condorcet winner, and it may have none.
Key Formula
Where:
- = The candidate being tested
- = Any other candidate
- = Number of voters who rank x above y
Worked Example
Problem: 100 voters rank three candidates. 40 voters rank A > B > C, 35 voters rank B > C > A, and 25 voters rank C > B > A. Is there a Condorcet winner?
Step 1: Compare B with A. The 35 voters in the second group and the 25 voters in the third group rank B above A. Only the first group of 40 ranks A above B.
Step 2: Compare B with C. The first two groups rank B above C. Only the third group of 25 ranks C above B.
Step 3: B beats both rivals, so B is the Condorcet winner. For contrast, compare A with C. A also loses to C, so the plurality leader A loses every head-to-head race.
Answer: B is the Condorcet winner. B beats A 60 to 40 and beats C 75 to 25. A has the most first-place votes (40), yet A loses to both B and C.
Why It Matters
Social choice courses use the Condorcet criterion to judge voting rules. A rule satisfies the criterion if it always elects the Condorcet winner whenever one exists. Plurality, the Borda count, and instant-runoff voting can all fail this test, while methods such as Copeland and Schulze are built to pass it.
Common Mistakes
Mistake: Assuming every election has a Condorcet winner.
Correction: Majority preferences can form a cycle. Suppose three voters rank A > B > C, B > C > A, and C > A > B. Then A beats B, B beats C, and C beats A, each by 2 to 1. No candidate beats everyone. This is the Condorcet paradox, so check every pairwise race before you name a winner.
