Loading page
About this tool
Find the Condorcet winner from pairwise comparisons, or expose a voting cycle.
The Condorcet Election Judge takes ranked ballots written one per line as "A > B > C" and runs every head-to-head matchup, declaring a Condorcet winner only if one option beats all the others one-on-one — that is, wins all n-1 pairwise contests. Committees, product teams and clubs get a full comparison table showing how many ballots preferred each side of every pair, plus an explicit "no Condorcet winner" verdict when the group's preferences form a cycle. It counts what you enter; it does not run the election or handle ballot secrecy.
Open Condorcet Election Judge on AltFTool — it loads instantly in your browser.
Type one ballot per line in the Ranked ballots box with candidates separated by >, for example A > B > C, or load the Five ballots preset.
The verdict recomputes as you type: either a named Condorcet winner or 'No Condorcet winner (tie or cycle)', with a caption counting the candidates and ballots found.
Read the pairwise table columns - Candidate A, Candidate B, A preferred, B preferred and Pair winner - then use Copy, or Download to save it as condorcet-election-judge.txt.
When no option beats every other, it says so outright instead of picking an arbitrary winner, which is the case a plurality count silently misrepresents.
The table lists both sides of each matchup with how many ballots preferred each, so anyone can recheck the arithmetic that produced the verdict.
Rankings are typed as A > B > C lines with no fixed candidate list to configure first, so you can paste what people actually sent you.
The option that beats every other option in a one-to-one comparison of the ballots. With four candidates that means winning all 3 of its head-to-head matchups out of the 6 pairs compared in total; if no option manages that, there is no Condorcet winner.
Because the group's preferences are cyclic or a pair tied. The classic case is A beating B, B beating C, and C beating A — majority preference can be intransitive even when every individual ballot is a consistent ranking, which is why the tool reports the cycle rather than inventing a result.
One ballot per line, candidates separated by the > character in order of preference, for example A > B > C. Names can be words rather than letters, and the candidate list is built from whatever appears across all lines.
A candidate that appears on a ballot is counted as preferred over any candidate that voter left off entirely, and a pair where neither name appears contributes nothing to that matchup. Truncated ballots therefore still count, but agree with your group on that rule before running the numbers.