For example, randomly sample N voters from the electorate, and run ML taking account only of those voters. Fails Condorcet, like RD, but it gives the behavior I specified. Politically infeasible, of course. And I’m not really sure I buy the argument-from-war for RD.
For example, randomly sample N voters from the electorate, and run ML taking account only of those voters. Fails Condorcet, like RD, but it gives the behavior I specified. Politically infeasible, of course. And I’m not really sure I buy the argument-from-war for RD.