How it works

Most polls and elections use plurality voting, where the candidate with the most votes wins. This works well if there are only two candidates or there is a clear global preference. However, sometimes there is a need for strategic voting when a preferred candidate is not the global preference. Much time and research has gone into better methods for selecting winners based on a more thorough voting system: preferential voting.

Top10r is a preferential voting tool. This means that it's like a poll except users can supply a complete picture of their preferences when electing a winner. How this happens is best introduced by example.

Suppose 5 friends on a road trip are feeling a little hungry and decide to stop for food. There are 3 fast food joints that they're considering: Arby's (A), Burger King (B), and Checker's (C).

Emilio, Ines, Rico, Manuel and Soledad each have their own ideas about where they'd like to eat. Emilio and Ines would like to go to Arby's, but prefer Burger King over Checker's, Soledad and Manuel would prefer Checker's, but can't stand those weird Arby's faux-meat sandwiches. Rico is the only one who would prefer Burger King, but is watching his waistline, so would really rather not go to Checker's.

If we are to consider only the first choice of each voter, it might seem that Arby's or Checker's are the two best choices. The voters may decide the overall winner by means of thumb wrestle or rock-paper-scissors, but if we dig a little bit deeper, we can discover more about their preferences.

  • 2 voters: A>B>C
  • 1 voters: B>A>C
  • 2 voters: C>B>A

2 People prefer A over B over C, 1 prefers B over A over C, and 2 prefer C over B over A. We can enter every pairwise comparison into a table. Because Emilio and Ines ranked A first and C last, we count that as 3 preferences: A over B, A over C and B over C.

Arby's Burger King Checker's
Arby's 2 3
Burger King 3 3
Checker's 2 2

This table is read row first, then column. That is, 2 people prefer Arby's over Burger King. It becomes more clear if we look at a graph:


We take the winner between each pair of arrows and calculate the difference between the preferences, to determine the strength of the beat between each pair:

Pairwise winners

Because both Arby's and Burger King beat Checker's, we can safely eliminate Checker's as a winner

Checker's is eliminated

Of the remaining options, more people prefered Burger King over Arby's, so Burger King is the winner. The overall election result becomes:

  1. Burger King
  2. Arby's
  3. Checker's

This simple example hopefully gives you a feel for the power of preferential voting. If you're interested in more details about the method and its properties, take a look at the wikipedia article on the Schulze Method or the research article.