[EM] Classifying 3-cand scenarios. LNHarm methods again.

[Kristofer Munsterhjelm]

Kevin Venzke wrote:

> The Burlington votes are inspiring. I'm amazed at how close the 
> first preference counts were, and that a fourth candidate even got
> 15%. Unfortunately the resolution is so stereotypical you could 
> think it was contrived to make a point.
> 
> What worries me is the possibility that every time we succeed in
> implementing an election method which can handle any number of
> candidates that we throw at it, we will mostly see scenarios with
> one or two strong candidates and a half-dozen losers that never
> coalesced into anything, so that we mostly will not be able to tell
> the difference in effect from just using FPP.

Even if the method would only produce one or two strong candidates, I'd 
imagine the information provided by the rank ballot input or by the 
outcome social ordering could be used by the "losers" to find out how to 
adjust their positions to have a better chance of winning.

Even in an FPTP two-party state, a new party may displace one of the old 
parties, such as with the Whig Party being displaced by the Republican 
Party. In FPTP, this is rare, though, because even if voters want to 
change parties, they have to consider the problem of the lesser of two 
evils. A better election method (such as a Condorcet method) would let 
the voters change parties more gradually, and so, such party shifts 
could happen more often (and the established parties would have to take 
that into consideration).

A greater problem would be if voters (or parties) would "mostly not be 
able to tell the difference", they could say "eh, why bother with this 
technological mess when it doesn't make a difference anyway?", and 
revert to FPTP. Then the subtle dynamics would never have a chance to 
come into being.