Cumulative voting: Difference between revisions

Add clone independence failure.
(→‎Quadratic voting: - properly linking to Quadratic voting and w:Quadratic voting (rather than using the title located at #Quadratic voting))
(Add clone independence failure.)
Line 40:
 
The optimal tactical vote depends heavily on <math>p</math>. For instance, vote-splitting is a problem in standard cumulative voting but not in Range voting. On the other hand, the Burr dilemma is a problem in Approval voting (and thus in tactical Range voting) but not in standard cumulative voting.
 
=== Clone independence ===
 
Every form of <math>\ell_p</math>-norm cumulative voting except <math>p=\infty</math> (plain [[Range voting]]) is vulnerable to vote-splitting.
 
For continuous cumulative voting, suppose each candidate can be given any score between 0 and 1. Then let <math>x = \frac{\sqrt[p]{2} - 1}{2\big(\sqrt[p]{2} + 1\big)}</math> and consider the election
 
* <math>\frac{1}{2}+x-\epsilon</math>: A (1) B (0)
* <math>\frac{1}{2}-x+\epsilon</math>: A (0) B (1)
 
for some <math>\epsilon</math> where <math>x > \epsilon > 0</math>. A wins by majority rule no matter the value of p.
 
Clone A into A1 and A2 so that everybody rates the two clones equal. The <math>\ell_p</math> normalization then leads to the first faction giving each A candidate <math>y = \frac{1}{\sqrt[p]{2}}</math> points each. As a result, the two A candidates tie if <math>\epsilon=0</math>, and B wins for any <math>\epsilon > 0</math>, which demonstrates the clone failure.
 
=== Example ===
 
Suppose we want to find a clone failure for <math>p=2</math> (quadratic voting). Then <math>x = \frac{\sqrt{2}-1}{2+2\sqrt{2}} \approx 0.086</math> and <math>y = \frac{1}{\sqrt{2}} \approx 0.707</math>. Let <math>\epsilon=0.001</math> to compensate for the roundoff error. Before cloning, the ballots are
 
* 0.585: A (1) B (0)
* 0.415: A (0) B (1)
 
where A wins 0.585 points to 0.415.
 
After cloning:
 
* 0.585: A1 (0.707) A2 (0.707) B (0)
* 0.415: A1 (0) A2 (0) B (1)
 
A1 gets 0.413 points, A2 the same, and B gets 0.415 points, thus making B the winner.
 
== Notes ==
1,202

edits