Single transferable vote: Difference between revisions
→Notes
(→Notes) |
|||
Line 182:
=== Methods of transferring excess votes ===
If the quota is 100 votes and a candidate wins with 250 votes:
* Random transfer: 100 ballots are randomly chosen from the pool of 250 and "spent", with the other 150 being transferred to their next-highest-uneliminated preferred candidate.
* Fractional transfer: 100/250=40% of all 250 ballots is spent, with all 250 ballots now supporting their next-highest-uneliminated preferred candidate, but with only 60% ballot weight/power. There are several possible variations of [[Surplus Handling]]
=== Ways of choosing a candidate to eliminate ===
Line 240 ⟶ 242:
STV and IRV can be visualized using Sankey or flow diagrams. If, in a given round, a candidate reaches the quota (a majority in IRV), they win, and if more seats are to be filled, then a quota of their votes are spent. This could be visualized by arranging the candidates from most votes to fewest in each round, and then showing a threshold for how many votes the candidate with the most votes in that round needs to win, or even further, showing the quota needed for the top (number of seats to be filled) candidates to win. Here is one example: https://www.reddit.com/r/dataisbeautiful/comments/8we8ep/sankey_diagram_of_results_from_maines_democratic/
STV passes [[PSC]] regardless of the method used to decide which candidate is to be eliminated next, because for a solid coalition comprising k quotas, once all but k of the candidates in the coalition are eliminated, at least one of the k remaining candidates will have a quota of 1st choices and win, with their surplus transferring if necessary such that another of the k candidates wins, etc. It passes the [[Droop proportionality criterion]] when the [[Droop quota]] is used, making it one of the few commonly discussed PR methods that guarantees that a majority will always win at least half of the seats.
STV has several variations that can be discussed. For example, Meek and Warren STV are variants of STV that attempt to make the process fairer, but at the cost of needing to be computerized to compute the result.
| |||