Strategy Advisor based on Randomized Voter Order: Difference between revisions

From electowiki
Content added Content deleted
(Initial page)
 
mNo edit summary
 
(2 intermediate revisions by the same user not shown)
Line 6: Line 6:


# For every possible way to order the ballots, or a sufficiently large random sample of these:
# For every possible way to order the ballots, or a sufficiently large random sample of these:
## For voter v = 1 ... number of voters:
## For voter v = 1 ... number of voters, in the given order:
### If voter v has chosen not to let the method strategize, pass the ballot to the underlying method (turning ratings into rankings if required).
### If voter v has chosen not to let the method strategize, pass the ballot to the underlying method (turning ratings into rankings if required).
### Otherwise, determine a good strategic ballot (typically using the [[w:Tactical_voting#Myerson%E2%80%93Weber_strategy|Myerson-Weber strategy]]) given the election result so far. Submit this ballot.
### Otherwise, determine a good strategic ballot given the election result so far (typically using the [[Tactical_voting#Myerson-Weber strategy|Myerson-Weber strategy]]). Submit this ballot.
## Add each candidate's score in previous election to their total score.
## Add each candidate's score in previous election to their total score.
# Elect the candidate with the greatest total score.
# Elect the candidate with the greatest total score.

Latest revision as of 17:23, 14 May 2024

Strategy Advisor based on Randomized Voter Order or SARVO is a category of Declared strategy voting methods devised by Warren Smith.[1] All SARVO methods are cardinal single-winner methods: the voters report their utility should X win the election, for every candidate X, and choose whether the method should strategize on their behalf or not.

In the election-methods post, Warren defined SARVO for Range voting, but it can be straightforwardly applied to any weighted positional method.

The method calculates the winner as follows:

  1. For every possible way to order the ballots, or a sufficiently large random sample of these:
    1. For voter v = 1 ... number of voters, in the given order:
      1. If voter v has chosen not to let the method strategize, pass the ballot to the underlying method (turning ratings into rankings if required).
      2. Otherwise, determine a good strategic ballot given the election result so far (typically using the Myerson-Weber strategy). Submit this ballot.
    2. Add each candidate's score in previous election to their total score.
  2. Elect the candidate with the greatest total score.

This page is a stub - please add to it.

References

  1. Smith, Warren D. (2005-10-25). "declared strategy voting / Sarvo-range voting".