5 ms·
In case there is no Condorcet winner, there is actually a more principled way to randomize over alternatives than just choosing each with the same probability,
by DominikPeters 4y ago
In case there is no Condorcet winner, there is actually a more principled way to randomize over alternatives than just choosing each with the same probability, known as maximal lotteries (e.g., https://pub.dss.in.tum.de/brandt-research/fishburn_slides.pdf https://pub.dss.in.tum.de/brandt-research/fishburn_slides.pd...).
- creata 4y agoI love this idea: it feels very parsimonious, and it's optimal (in a very narrow sense) by definition. I can't find much criticism of it, though, aside from: https://groups.google.com/g/electionscience/c/wT1_hN38b8s https://groups.google.com/g/electionscience/c/wT1_hN38b8s
- c7b 4y agovery cool, thank you