I like this. But you are not ambitious enough. Why just 1 random partner? Why not many? Entire segments of the populace with whom you must come to a common decision! And you do not even need to discuss with them in advance - just make your decision known, and they will too.

For a brief but thorough discussion of alternatives to the modern, election-based form of democracy, read Against Elections, by David Van Reybrouck. There are less-convoluted ways than yours that could remedy the problems you refer to. And many of them have the added advantage that they have actually been put into practice in the real world. I think your system will absolutely fail to preserve the secrecy of the vote, which means it will be extremely straightforward for a suitably wealthy or powerful anti-democratic force to use intimidation and public reprisals to force the vote.

You assert that the problem with real-world democracy is "increasingly divisive politics," but I say that's a symptom. To take the U.S. as an example, there are many policies that are very popular across the spectrum, but aren't made into law because the political class mostly serves the ultra-wealthy (who oppose those policies). If it were just about counting votes, we would have different laws & leaders. Instead, we play games about which votes we count and how they are grouped. Not a problem of division, but of election subversion. Your "solution" is, oddly, more of the same.

For more, see Gilens & Page, whose study documents the fact that "economic elites and organized interest groups play a substantial part in affecting public policy, but the general public has little or no independent influence." Not a problem of voting, but a problem of self-reinforcing corruption. Which, again, is not addressed by your proposed alternative. (Article about G&P study: vox.com/2014/4/18/5624310/…)

Your fictional country has already failed. There is no freedom of association. Votes are not secret. (And they have have thrown overboard a perfectly workable democratic system for this nightmare, meaning that there is something fundamentally wrong with their electorate.) (P.S. Votes which don't match cancel out even in the ordinary democratic system. All you have made is destroy a good system for no gain.) (And collecting data about how people voted or will vote is difficult by design. It is not a bug. It is an essential feature.)

What this proposal addresses is not the voting, but the polarization of society. The process of needing to communicate honestly with a random stranger and come to an agreement changes the participants. They become more similar in outlook. We see the same thing happening in jury members. I recommend not needing to compare the votes, but simply have the process of matching and then only allowing them to vote when agreeing. The votes can be secret because the value is in the process, not the actual votes. I see this as building a more common society.

@DavidR Yes, more common society is the main goal eventually. 'Vote only when agreeing' removes anonymity, and leads to vote selling. Same but person can post a different vote after the conversation keep the system safe from intinidation. I agree that your way is preferred in an idesl world, but option i offer is more resilient against bad players, and that is needed too.

@Tom Poor are single issue voters. They consider 1 problem, pick a candidate for it, and vote this way. Rich people consider 100 problems at once, and pick their candidate based on best match on the total of their problems. This is inevitable. The only thing that can be improved is to explain this better. That being stubborn is the reason your vote means nothing. System I offer reduces the time till this realisation, even for poor people. They instantly, in an hour, will see why exactly they lose on average. No system of vote helps, system of education will. This is what I offer.

@Tom random person in power is a complete cop out. Sure, it sort of works, and is comparable to the modern system. But this will completely giving up on any progress. It will be a different system, but it will not be better, and worse still, it will never improve, as there is no decision mechanism. Decision is taken on random. I do not want to go down that road because I want to see the chance of a better system down the line. System I offer allows to test an idea that could be better, but more than that it allows a path for a whole rangw of new systems. This is what interests me. Evolution.

@Willk I think you miss the main point. Idea of not counting not-matching voices is very different from current system. Currenrly you just vote for top2 parties, because you have 0 options to change the mind of the whole society during the election. System I offer allows a chance to actually change the mind of your assigned partner, or change yours in the process. This is the difference, blind belief that you are right VS an actual conversation, where you can test your ideas, and test what matches and what doesnt with the other views.

Biggest problem come when there are more than two options. The only people who get any representation are those with the most homogenous beliefs. So 90% of society doesn't get any vote. A single voting block with homogenous beliefs can dominate. Minority opinions without high cohesion will eventually quit trying.

@DWKraus that is the point. Fringe beliefs are not compatible with each other. If you want your opinion to matter, find a way to make it compatible with others

@SurprisedSeagull welcome ! +1 interesting proposal ! but to keep this one open on WB, it will have to have a clear question (hint), all questions in your opening seem already answered.

@Goodies question is what will fail. People mention privacy, just compare it to existing system, or point to the ignoring of fridge opinions. Im interested if something else might fail

Nice edit ! I think there are 2 pitfalls that could make this system fail: when you would keep this system for too long, you'll get a very dull politics, with weak and undecisive government. That will reflect negatively on participation. I also found an issue when I was writing my answer.. in a multi-party democracy, it could be difficult to get enough votes to have a representative outcome.

@Goodies to make sure we are on the same page, lets consider this: probably majority of people just cant agree on anything, at all. So a minority of people will be making decision in the world I describe. This minority would consist of people who can change their opinion with the new data. Both of this, smaller decision group, and ability to change quickly, will make the politics more responsive as I see it. I agree that it wont be as explosive as some modern leaders are, but I dont see it as inherently bad thing. Decisions will be made and quickly, just less shiny.

I think participation is a difficult thing that could undermine your system.. for this type of democracy, you will need motivation to meet other people, to start with. You can't make participation too small: people will start to complain about elite democracy and not being represented. There must be incentives to meet, and reach a compromise. Say your traveling costs are covered, or there is some benefit attached to putting your vote together..

Who controls "random" selection? Control of that would be a prime target for corruption. If you can pair all your party members together and mix up member of other parties, only your party gets counted.

@Goodies I was thinking connecting the UBI with this, as an insentive. But thats a whole new story. But I agree, in a modern world talking with another person is an event worth compensating for. Sad but true

Yep.. it would certainly undermine yout system, I have put it as a last sentence in my answer.

@JamesFaix I agree that it could be exploited. But I would say it is harder to exploit than the modern system. The agreements, initial position and eventual decision is the most valuable information from this system. And it will show such corruption too. And if you have an access to such edits, you could just rewrite others votes or make new votes anyway.