election-methods@mailman.electorama.com

Technical discussion of election methods

View all threads

Some easier burial resistance criteria

KM
Kristofer Munsterhjelm
Sat, Jul 17, 2021 5:38 PM

I find it kind of hard to reason about burial-resistant Condorcet
methods because DMTBR is such a difficult criterion to pass. Perhaps the
following criteria will be good steps on the way:

Mutual majority burial resistance (MMBR): If A is in the smallest mutual
majority set and wins, then voters who prefer B to A can't make B win by
burying A under a candidate X not in the smallest mutual majority set.

Pareto-dominated BR: If A wins and X is a Pareto-dominated candidate,
then voters who prefer B to A can't make B win by burying A under X.

DMTBR seems to rely on that there can't be more than two disjoint solid
coalitions of the same cardinality supported by more than a third of the
voters. Similarly, MMBR would have to rely on there being only one
smallest mutual majority set. I may have an idea for a method that
passes MMBR, but I would have to check that my reasoning is correct.

But as for Pareto-dominated BR, I don't yet have a clue.

-km

I find it kind of hard to reason about burial-resistant Condorcet methods because DMTBR is such a difficult criterion to pass. Perhaps the following criteria will be good steps on the way: Mutual majority burial resistance (MMBR): If A is in the smallest mutual majority set and wins, then voters who prefer B to A can't make B win by burying A under a candidate X not in the smallest mutual majority set. Pareto-dominated BR: If A wins and X is a Pareto-dominated candidate, then voters who prefer B to A can't make B win by burying A under X. DMTBR seems to rely on that there can't be more than two disjoint solid coalitions of the same cardinality supported by more than a third of the voters. Similarly, MMBR would have to rely on there being only one smallest mutual majority set. I may have an idea for a method that passes MMBR, but I would have to check that my reasoning is correct. But as for Pareto-dominated BR, I don't yet have a clue. -km