Kleroterion | Hacker News Reader

Kleroterion

en.wikipedia.org

99 pointsvaleg22 comments

While Athens was not selecting ballots, but representatives directly, random ballot voting[1] is related and is one of a very few voting systems in which strategic voting (voting for other than your preferences in order to get the desired outcome) is never correct.

There are systems that make strategic voting not correct in most (if not all) real-world settings, but those are so complicated that IMO strategic voting would still happen just because people don't understand the system.

I would love to see random ballot voting for electing any sufficeintly large set of representatives, it's simple and provides proportional representation.

1: https://en.wikipedia.org/wiki/Random_ballot

Non-determinism isn't great though.

It could be useful to allocate a number of seats, but then deciding who filled those seats would be a matter of headache (or "trust the party!" goes into effect and removes direct voting)

Determinism can be readily added by means of a simple procedure (the name of which I have forgotten—hopefully a fellow commenter can chime in here): Along with their vote, each voter submits a random number between 1 and N, where N is the number of voters. To select which ballot should decide the vote, all numbers are publicly revealed and added together, modulo N: the resulting number selects which ballot should be used. It is not possible for any individual or group to fix the final number in advance, since even a single non-cooperating individual can alter the final number in an arbitrary way.

Obviously this simple form can only be used if N is known in advance, but I think that if you instead select a sufficiently large maximum (several times larger than the probable N) the randomness property should remain.

I'm not sure I understand "deciding who filled those seats would be a matter of headache" you just pick a random ballot and whomever is written on the ballot gets the seat. Repeat for each seat (possibly by district).

I've never heard of this!

What about seniority or as HN might call it 'institutional knowledge'? 60% 'safe' seat that used to be held by same Member for life, now 4 out of 10 times is not.

And how do you solve for problems with crazy people or super super extreme. Even in a large body like the US House, one member can do a decent amount of damage.

> And how do you solve for problems with crazy people or super super extreme. Even in a large body like the US House, one member can do a decent amount of damage.

Yes, the specifics of the parliamentary procedure used would have to be adjusted so that a single bad actor can't dominate the whole body. OTOH I've seen a single bad actor in smaller systems (think HOA[1] board) get controlled fairly easily with Robert's rules and a professional parliamentarian.

> What about seniority or as HN might call it 'institutional knowledge'? 60% 'safe' seat that used to be held by same Member for life, now 4 out of 10 times is not.

Some people think of this as a feature rather than a bug; there is a long history of suspicion of career politicians. I'm not so sure myself, as there is something to institutional knowledge.

You could solve this by having every region elect N representatives and each ballot list N candidates in order of preference. At N=4 a 60% popular politician would be elected 98% of the time. At N=2 a 60% popular politician could expect to serve for 2 terms over 70% of the time, which is in keeping with some suggested term limits.

1: https://en.wikipedia.org/wiki/Homeowner_association

haha look at Ted Cruz ;) !! though perhaps Senate isn't a big enough body I can find examples in the House too

yeah no one polls as bad as Congress, but I don't want some 'randomly' chosen Freshman as Speaker.

I've found that Reps that have actually competitive races are by far the most responsive by necessity to their District and I think that's a better 'fix' than this proposal.

I named a project after these! https://www.kleroteria.org

Saw it and signed up a few weeks back!

I love the messages from each of the people selected as they come in - some very interesting long scripts, others a single line, but all have made me smile.

I was selected a wee while back, but haven't gotten around to writing as yet - will add it to my task list.

Interesting idea/project, never heard of the listserve or anything similar. Site is clean and fast, nice work ;-) A few questions if I may. How many people are registered? If someone is selected (randomly), can he/she be selected again next week? Do you plan to release/make public available all previous emails?

Thanks!

There's around 1k members right now. I stole the 10k catchphrase from The Listserve. Consider it aspirational!

Yes, people can be selected multiple times, but they must submit a post before they can be selected again.

No, I don't currently plan to release the posts. I might change that at some point but I'd want authors to opt in -- even though the license allows it without their consent I doubt everyone understands that.

Hi Simon. I am a subscriber. I really enjoy the occasional random email from kleroteria.

What a cool idea! Is it still active? I ask because the last published mail is from March.

It is! I should make it clearer that the published email is just an example of what's sent out. There's been about a post a week since it's started.

edit: updated; thanks for the feedback!

Look at that! And then look at contemporary hiring. A phone screen, a video interview, an on-site, another on-site, a take-home exam, psychometric testing, dowsing, crystal pendulum, and it still doesn't work better than casting lots.

>it still doesn't work better than casting lots.

i don't know about that. Using all these tools you listed the companies have been able to "hire the best, the top 1% (or 2%)" while casting the lots they would definitely not be able to do that.

Is it actually true that FAANG hire the best, top 1%?

Seeing the products they create, I am not convinced.

> the companies have been able to "hire the best, the top 1% (or 2%)"

Citation strongly needed. Having worked with a lot of big companies, I'd say Top 20% is still generous.

1 or 20 doesnt matter - for any N>0 only one company can do "top N", not all of them who pretend to achieve that, and they wouldn't be able to pretend it if they used the random approach like the GGP suggested.

It's interesting that the people being chosen with a kleroterion and black and white dice were called 'dikastes', since the mechanic was literally 'die casting' (as in die rolling, not the manufacturing process).

This is cool, but why didn't they use 5 chests with tickets on them, drawing names round-robin from the chests?

Probably for technological reasons: they didn't use paper so the tokens had two problems: 1 - they were non-uniform so could in theory be biased and 2 - typically the tokens were made of dried clay, so could easily fracture.

For casting votes potshards were typically marked by the voter to indicate preference, then discarded.

Reply on news.ycombinator.com