[EM] Proportional representation methods with very many candidates

[Juho Laatu]

On 3.12.2013, at 14.49, Kristofer Munsterhjelm <km_elmet at t-online.de> wrote:

> For one, because there are so many "candidates", most of them will be tied in a Plurality elimination stage like the one in STV.

In systems like Google search engine one can easily use the general popularity of different candidates to break the ties (= to have a small weight in the election).

> There is of course the compromise solution of just using sequential PAV or proportional Range voting. In "worse-is-better" style, it wouldn't care very much about theoretical guarantees as long as the method is good enough.

I note that these elections are not competitive. This means that one can use also e.g. such cardinal style methods that do not work in competitive (political) elections.

One more observation is that the user can give feedback. The method could e.g. allow the user to tell which hits are no good, and the results would change accordingly.

Juho