Social Welfare (Metrics Discussion)

[AngelaKTE]

Description

• Goal: Measure how well the voting design represents the preference of the majority of voter preferences
• Why here?:
  • If OP Retro Funding, the voting should reflect the majority of preferences, we should prefer a voting rule, that creates an outcome with minimal distance between the voting outcome and voter's choices on average

How we model it:

• We create a random matrix of votes, and produce a voting outcome
• We measure each voters L1 distance between vote and voting outcome
• We sum up all L1 distance values and divide by the number of voters (average L1 distance)
• We compare across all voting rules

Chart:

a) yes, I'd like to see the L1 distances boxplot! (not average)

b) I'd like understand if the effect of voting rule is equal in all scenarios (e.g. with a lot more/less voters, or a lot more/less, or a lot more/less total funding to distribute to projects)

c) I'd like understand if the effect of voting rule is equal across all funding tiers e.g. to see L1 distance translated into funding values:

e.g. the top funded project, and boxplot of L1 distribution

e.g. the least funded project, and boxplot of L1 distribution

[linear[bot]]

GOV-28 Social Welfare (Metrics Discussion)

### Description * **Goal:** Measure how well the voting design represents the preference of the majority of voter preferences * **Why here?:** * If OP Retro Funding, the voting should reflect the majority of preferences, we should prefer a voting rule, that creates an outcome with minimal distance between the voting outcome and voter's choices on average **How we model it:** * We create a random matrix of votes, and produce a voting outcome * We measure each voters L1 distance between vote and voting outcome * We sum up all L1 distance values and divide by the number of voters (average L1 distance) * We compare across all voting rules **Chart:** * I'd like to see the aggregated L1 distance * I'd like to see the average L1 distance