Strategic Voting in the Context of Negotiating Teams

Leora Schmerler and Noam Hazon

Link to the paper: https://arxiv.org/abs/2107.14097v2

This paper is about:
Internal strategic voting inside negotiating teams – groups of agents that must first agree on a joint bid before facing an external opponent.
Schmerler & Hazon model the team’s decision stage as a positional-scoring vote (e.g., Borda, 3-approval) and ask:

Can one or more team-members misreport their preferences so that their favorite option becomes the unique outcome of the subsequent VAOV negotiation game?

They give two polynomial-time algorithms:

C-MaNego (single manipulator) – decides whether one voter can force her preferred outcome under any positional-scoring rule by computing the first round where the Rational-Compromise (Bucklin-style) intersection is a singleton and adjusting her ballot accordingly.
CC-MaNego (coalition of k) – extends the idea to a coalition of k manipulators.

*******************************************

You are part of a group of voters.
Your group will negotiate with another group for a single outcome.
Your goal is to manipulate the vote of your group so that your preferred candidate will win.

Enter your team profile (one ranking per line, e.g., A > B > C):

Live count: 0 candidates

Please enter the profile using '>' to separate ranked candidates.
Each line represents a voter. Example:
A > B > C
B > A > C

Next, choose the voting rule your team will use to pick its negotiation offer.

Borda – each candidate gets points according to their position in the ranking (e.g., 3 points for 1st, 2 for 2nd, etc.).
Plurality – only the top candidate gets points (1 point) [1,0,0,...,0].
Veto – all candidates except the last one get a point [1,1,1...1,0].
X-Approval – candidates in the top x positions get a point [1,1...1,0...0].

plurality and veto are actually x_approval but with a specific condition:
veto: x = m-1
plurality: x = 1.

Voting rule:

Now, you need to enter the preference of the other team, so we can see which candidate would be the outcome of the negotiating between the two groups.

Opponent ranking:

Preferred candidate: (the candidate you want to win the negotiating)

how many people in the manipulating group (k):

Example - loads a predefined example profile with a preferred candidate and opponent order.
Example K=2 - loads a predefined example profile with a preferred candidate and opponent order, but with k=2 manipulators.
Random Example - generates a random profile with a random preferred candidate and opponent order, you can choose the number of candidates and/or voters.
Total Random - generates a random profile with a random preferred candidate and opponent order, and a random number of candidates and voters (everything is random).

Load an example: Example Example K=2 Random Example Total Random

Press Load after making a choice and filling all the necessary fields.

About this project