- VERSION 1.06
- Method Common Name: SmithSet IRV
use Vote::Count::Method::CondorcetIRV; ... # SmithSetIRV my $winner = SmithSetIRV( $Election ); say $Election->logv();
Provides Common Basic Condorcet-IRV Methods. These methods are simple and beat other Condorcet Methods on Later Harm.
This module exports the methods it provides which expect a Vote::Count object as an argument.
Identifies the Smith Set and runs IRV on it.
SmithSetIRV is exported and requires a Vote::Count object, an optional second argument is an IRV tiebreaker rule name (see IRV module). It will return a hashref similar to RunIRV, in the event of a tie the Vote::Count Object's Active Set will also be the tied choices (available for any later tie breakers you would implement). Events will be logged to the Vote::Count Object.
SmithSet IRV is easy to understand but requires a full matrix and thus is harder to handcount than Benham. An aggressive Floor Rule like TCA (see Floor module) is recommended. If it is desired to Hand Count, a more aggressive Floor Rule would be required, like 15% of First Choice votes. 15% First Choice limits to 6 choices, but 6 choices still require 15 pairings to complete the Matrix.
When there is no Condorcet Winner this method is Later Harm Sufficient. There might be edge cases where IRV's sensitivity to dropping order creates a Later Harm effect, but they should be pretty rare. When there is a Condorcet Winner the effect is the normal one for a Condorcet Method.
The easiest way to imagine a case where a choice not in the Smith Set changed the outcome is by cloning the winner, such that there is a choice defeating them in early Top Count but not defeating them. The negative impact of the clone is an established weakness of IRV. It would appear that any possible Later Harm issue in addition to being very much at the edge is more than offset by consistency improvement.
Smith Set IRV still has a significant Later Harm failure, but it has less Later Harm effect than other Condorcet methods.
Meets Condorcer Winner, Condorcet Loser, and Smith.
By meeting the three Condorcet Criteria a level of consistency is guaranteed. When there is no Condorcet Winner the resolution has all of the weaknesses of IRV, as discussed in the Later Harm topic above restricting IRV to the Smith Set would appear to provide a consistency gain over basic IRV.
Smith Set IRV is therefore substantially more consistent than basic IRV, but less consistent than Condorcet methods like SSD that focus on Consistency.
John Karr (BRAINBUZ) email@example.com
Copyright 2019 by John Karr (BRAINBUZ) firstname.lastname@example.org.
This module is released under the GNU Public License Version 3. See license file for details. For more information on this license visit http://fsf.org.