An attempt at a “good enough” solution for human two-party negotiations
I previously wrote about the difficulty of fair two-party negotiations when there isn’t an authority to refer to (like an established market price).
It’s not perfect; the system is still exploitable with enough effort. But I’ve tried to leverage human biases and computational limits to make it such that exploiting it is hard and comes with only a low reward, so I expect most people will tend towards not thinking about it too much and just inputting their true prices.
A detailed writeup of how it works is available here. I’d appreciate any suggestions you may have on how I could improve it.