What are people's thoughts on using proportional sampling on the

Question

What are people's thoughts on using proportional sampling on the

outcome of a vote in order to determine what action is actually taken?

Example: Suppose a vote finishes at 51% for A and 49% for B.

Typical majority rules would say A is the victor. In this alternative system, one implementation might be to say that we will choose A with 51% probability and B with 49%. We then do a single random sample (onchain) according to those finalized probabilities. The result of that sample is the action we take.

Now, why would we ever consider such a system? The idea behind it is to prevent majority tyranny. If there is a group that has a very slight but very stable majority, you could get in a situation where, say, 51% of the population is always dictating what 49% of the population doesn't want. When you sample, you ensure that the small majority only dictates 51% of the time. So in expectation it is a fairer and more representative of the population.

There are some obvious concerns...we don't want 1% ideas, possibly terrible ideas to have a real chance of passing. Similarly, super majorities should perhaps generally win out. This is where the exact proportionality we use is an implementation detail.

The simple impl is that the sampling probability is equal to the voting outcome probabilities. So a 1%, 5%, 94% voting outcome implies sample probabilities of 0.01, 0.05, and 0.94.

Another choice is that we say there is a minimum voting outcome required, but still linear otherwise. So for example, you might use zero sampling probability for voting outcomes in the 0-33% range....then a linear increase in sampling probability, hitting 1 when the voting outcome reaches 66%. Taking the same example, a vote outcome of 1%, 5% and 94% would translate to sampling probabilities of 0.0, 0.0, and 1.0....effectively guaranteeing that the super majority in this case gets its outcome. (This would need to be tweaked a bit to use relative outcome probs above the quorum floor...for example, you'd want a vote outcome of 30% 35%, 35%...to go to sampling probabilities of 0, 0.5, 0.5...but ideally most votes we would do are for yea/nay proposals.)

Summary: to avoid tyranny of small but stable majorities, we should sample the voting outcome in a weighted fashion, while still ensuring that super majority ideas win out and very low choice ideas get zero weight.

#cryptocurrency #english #frax

0

25.05.2023

2 ответов

8 просмотров

TreeClimbingBear Автор вопроса

C2tP
There's no way. That just leaves randomness as wel...

With sufficient agreement from people still?

0

11 месяцев назад

C2tP · Accepted Answer

C2tP

There's no way. That just leaves randomness as well as an extra attack vector. An attacker no longer needs majority so they just keep attacking until it goes through

0

11 месяцев назад

209 похожих чатов

What are people's thoughts on using proportional sampling on the

2 ответов

Похожие вопросы