zlacker

I am curious, what would such a system look like?

I am thinking that a "promise" is not a quantitative term. It needs to be ratified into specific data oriented actions that can go through a litmus test whether it was fulfilled or betrayed.

After that, one idea is to have an accountability score tracked by bureaucracy and have that printed on the ballot along with their principle accomplishments in the supplement. Another idea is to have a penalty score of not meeting prior promises as a dilution factor to the number of votes. If a politician only met 90% of the promises, they will lose 10% of the voting power of the public (like a 0.X multiplier to the votes). Just thinking out loud, there may be major issues with these ideas.

>>system+(OP)
A control theory idea: use continuous voting. When the support falls below some threshold (e.g. 50%), kick the candidate out.

But now you'll get an unstable system where candidates get kicked out all the time and are too populist because they don't expect to live long.

So add a low-pass filter. When the moving average of the candidate's support falls below the threshold and a definite other candidate's support is high enough, replace the incumbent with that challenger.

You might even increase the duration of the moving average with time, like the doubling trick in multi-armed bandits. The logic is that a candidate who has shown that he can weather the initial period without getting voted out can be trusted with more long-term decisions, i.e. actions speak louder than words.

>>system+(OP)
> I am curious, what would such a system look like?

Switzerland, people can always overrule politicians decisions there, no need to wait until next election. This means that the opinions of politicians is no longer as important so this issue doesn't even come up. So you place more value on finding the politician best fit for running the country, not the politician with values most aligned to your own.

>>somegu+hp
Only on HN do you find a theory of politics based on Controls/DSP fundamentals. Nyquist plot of political instability would be nice. I applaud you.

>>somegu+hp
Not a bad idea but it probably depends on a level of convenience and ease of access to voting that does not seem to exist currently.

Unrelated side note: when you are talking about a hypothetical politician, be aware of your choice of pronoun.

You might be the type of person who picks between he/she with a precise 50/50 split but I'm going to assume you are not that type of person. Similar to the way you seem to have assumed that if a person is a politician they are also a man.

>>somegu+hp
If you have continuous voting it may be better to vote for decisilons instead of voting for people. Because politicians usually happily pretend to support whatever policies they think are popular. https://news.ycombinator.com/item?id=23377423

>>shuntr+Zz
Women aren't 50% of politics. It's not a terrible Bayesian prior to say he. (as someone who says "they" for hypothetical humans)

>>sooheo+HK
But when done systemically it likely creates the expectation that politics is done only or mostly by men, therefore discouraging women from entering politics now and in the future. Which, I would argue, is a substantial net negative.

>>system+Ss
Okay, so this is where I will bring my proposal for Modular Government.

Next ammendment: direct election of cabinet members.

Also, all changes to policies to be parameterized and adiabatic.

>>shuntr+Zz
FWIW, English is my second language, and my first language has grammatical genders that don't have anything to do with real genders. We were also taught that "he" is the default pronoun in English.

I guess some of that bleeds through: that I use "he" without reflecting on it because it wouldn't carry an implication of actual gender in my first language. I am definitely not assuming that politics is a men's only club.

>>a9h74j+d31
>Also, all changes to policies to be parameterized and adiabatic.

"No heat, only work". I like that proposal :-)