Hacker News new | ask | show | jobs
by baoyu 1540 days ago
Here’s a quadratic voting demo that looks nice and provides more intuition: https://www.economist.com/interactive/2021/12/18/quadratic-v...
3 comments

This is great!

And it reveals another (probably unsolvable?) UX problem with quadratic voting: it takes so much longer to allocate votes by comparison to a thumbs-up / thumbs-down vote.

For myself, I did a pass through the issues, supporting or opposing each issue with a single vote. Then, I did another pass to add additional votes to each issue in proportion to how much I cared, while carefully paying attention to how many issues were left to be voted on, and my remaining "budget" of votes. Then, I did a final series of passes to rebalance everything, trying to make it as reflective of my positions as possible, while using as many votes as I could (so as not to leave any crumbs of democracy on the table, so to speak).

The whole thing was actually really complex, and not done in "linear time" like a traditional ballot. And it wasn't because I thought more about the issues, but purely because the rules of the system introduced so much more overhead.

Not saying I dislike the idea of quadratic voting, just that you can really feel the difference.

Why do the questions never include my purchasing power, given my current financial situation?

This is what 50% of the population cares about, yet their votes are stolen away by focusing on things like immigration and terrorism.

The mismatch in these questions is why lower social classes often vote for populist right wing parties, instead of in their own best interest.

Right.

The problem isn't how much votes cost, it's how many questions have to be answered at once.

thanks for sharing, I love the UI!
That UI really is good.

The large number of votes (100) and issues (10) seems to reduce the issue around not being able to "spend" all your votes as well. When I've only got 25 it feels particularly wasteful to not be able to spend 7. But with 100 and 10 choices I really had a lot more degrees of freedom.

I run a small ranking app and have previously landed on Schulze method for calculating the rankings. http://blog.forcerank.it/counting-votes-is-hard I'm going to have to think about adding quadratic as an option though.

For the example given, it seems like the majority criterion is actually not desirable.

Picking a talk topic that everyone will like a lot is a better outcome than picking another slightly preferred by most but completely uninteresting to the rest.