[EM] LaterNoHelp

[Kristofer Munsterhjelm]

On 02.06.2022 19:35, Forest Simmons wrote:
> LNhelp is a double edged sword. If lower ranked votes cannot help your
> opponent, they cannot be counted on to help you either. Yet one of the
> biggest talking point claims of IRV advocates is that it satisfies
> Reliable Later Help: "If your first place choice is eliminated, then
> your second place choice will still be there."

LNHarm + LNHelp is just burial immunity. Like Plurality :-)

The FBC is what makes it safe to always vote your sincere favorite
(co-equal) at top... which is more about compromising.

So if the IRV advocates suggest that LNH implies you can vote your
sincere favorite top, then they're doubly wrong!

In another post, Kevin Venzke mentioned that he found some Condorcet
methods to be considerably more susceptible to compromise than others.
My impression was that since some Condorcet methods are *almost* FBC
compliant, and that it's possible to make them completely FBC compliant
at the cost of losing Condorcet compliance (e.g. Minmax and MMPO),
Condorcet methods as a whole were pretty close to FBC in some sense, in
that they generally aren't susceptible to compromise strategy. James
Green-Armytage's results also suggest this, as the Condorcet methods he
investigated all had relatively low compromise incentive.

But apparently it's not that simple, as Kevin's results indicate!

-km