Rank instead of pick. The end of the wasted vote.
With a single box, people often vote for the lesser evil. By ranking the options, you give your true first choice without wasting your vote; the result picks the most broadly supported option. Here is how, step by step.
With a single box, the majority can end up with the wrong winner.
100 voters, three candidates. Adèle and Sophie are close; Camille is alone in her lane. Under first-past-the-post, Camille wins with 40 votes: the other 60 are split between Adèle and Sophie.
Yet 60 voters out of 100 prefer Adèle to Camille, and just as many prefer Sophie. Camille wins while losing every one of her head-to-head contests: this is vote splitting, which feeds the fear of a “wasted vote.”
A measured effect: in France, a third candidate staying in the second round costs the closest candidate the race about one time in five (Pons and Tricaud, 2018).
- Camille 40 · in the lead
- Adèle 35
- Sophie 25
- Camille , then Adèle , then Sophie 40 voters
- Adèle , then Sophie , then Camille 35 voters
- Sophie , then Adèle , then Camille 25 voters
Rank in order of preference.
Voters order the options by drag and drop. There is no need to rank everything, and two options can share the same rank.
Ties allowed. Place two choices at the same rank.
Partial ranking. Leave aside what you do not want to decide between.
The winner is the one who beats every other option.
In 1785, Condorcet proposed comparing the options two by two, like so many small runoffs read from the ballots. The option that beats all the others is the Condorcet winner.
On the same 100 ballots, Adèle beats Camille 60 to 40 and Sophie 75 to 25: she is the Condorcet winner. Camille, first under first-past-the-post, loses all her duels.
Sometimes, nobody beats everyone.
In another vote, preferences go round in circles: Camille beats Adèle, Adèle beats Sophie… and Sophie beats Camille. This is the “Condorcet paradox”: consistent individual choices can produce a circular collective preference.
- Camille , then Adèle , then Sophie 40 voters
- Adèle , then Sophie , then Camille 35 voters
- Sophie , then Camille , then Adèle 25 voters
A rare case in real votes.
Across nearly 10,000 online votes, a Condorcet winner exists in 98% of those with more than 100 voters (Myers, 2024). Cycles mostly affect small groups, hence the value of a clear rule.
A complete ranking, and almost never a tie.
To settle a cycle and rank every option, we use advanced mathematical methods, Schulze by default. They start from the same duels, always elect the Condorcet winner when one exists, and carry the ranking all the way through.
-
A cycle settled consistently. The least decisive victory (Sophie over Camille, 60 to 40) gives way to the other two: Camille wins.
-
A complete order. First, second, third… enough to elect several people in a single vote or designate alternates.
-
Unlikely ties. With Schulze and Ranked Pairs, their probability tends toward zero as the number of voters grows. In a small group, a tie remains possible; it is then displayed as is.
-
No element of chance. Same ballots, same result, recountable by anyone from the public register.
A reasoned case for ranking.
The most broadly supported choice
The Condorcet winner is preferred to each of the other options by a majority. First-past-the-post does not guarantee it.
The end of the “wasted vote”
Putting Sophie first no longer costs Adèle anything: your second choice still counts. With Schulze and Ranked Pairs, a candidate similar to another does not change the outcome among the others either.
Voting instructions carry less weight
An instruction to “put X first” is no longer enough: X must beat each rival by a majority, which no organized minority achieves alone. Negotiated withdrawals lose their point.
Less room for tactical voting
To skew the result, a group has to agree to misreport its preferences. Simulation studies show such maneuvers succeed less often with a Condorcet method than with first-past-the-post.
The limits, in all honesty.
- No voting system is immune to manipulation (Gibbard–Satterthwaite theorem). A coordinated group can try to “bury” a rival to manufacture a cycle; the maneuver requires broad coordination and can backfire on its authors.
- In rare configurations, ranking a second choice can hurt your first: no Condorcet method can rule this out.
- Our arguments about voting instructions rest on the structure of the methods; no empirical study measures them yet.
Two flagship algorithms, led by Schulze.
All our methods pick the Condorcet winner when one exists; they differ only on cycles and on the order of the options that follow. The method used is displayed with the result.
-
The Schulze method
DefaultThe most robust method: it compares options by the strength of their chains of victories, resists clones and makes ties unlikely. Our default choice, and our recommendation.
-
The transparent alternative: duels are kept one by one, from the most decisive to the closest, and the calculation can be followed by hand. Almost always the same winner as Schulze.
Secondary methods
With no real advantage over the previous two, they meet specific needs, often set by bylaws: simpler to explain, they were designed for a hand count.
- The Copeland method One point per duel won, one taken away per duel lost: simple to count by hand.
- The Dodgson Quick method The option closest to beating everyone, following Lewis Carroll’s idea.
What people often ask us.
Do I have to rank all the candidates?
Can I put two candidates at the same rank?
What if there is no clear winner (a cycle)?
Which tally methods are available?
Can the result end in a tie?
Does my ballot stay secret?
The works this page draws on.
| Type | Reference | What it establishes |
|---|---|---|
| Paper | Expressive Voting and Its Cost: Evidence From Runoffs With Two or Three Candidates Vincent Pons, Clémence Tricaud (2018) · Econometrica 86(5) | The measured cost of vote splitting in French elections. |
| Paper | The Frequency of Condorcet Winners in Real Non-Political Elections Andrew C. Myers (2024) · Public Choice Society | How often a Condorcet winner exists across nearly 10,000 real votes. |
| Paper | Can a Condorcet Rule Have a Low Coalitional Manipulability? François Durand, Fabien Mathieu, Ludovic Noirie (2016) · ECAI 2016 | Condorcet methods leave less room for coordinated maneuvers. |
| Paper | Manipulation of Voting Schemes: A General Result Allan Gibbard (1973) · Econometrica 41(4) | No voting system is immune to all manipulation. |
Try ranking on your next vote.
Free up to 20 voters, no account, no credit card.