[EM] A more typical example. How the various methods do.

[Michael Ossipoff]

There are reasons why the CW is the most likely candidate to have the
highest Topcounts:

…

1. The highest candidate in a ranking is ranked over more candidates.

…

2. In an unskewed normal-distribution, max coincides with median.  Yes,
median-crowding can spoil that, but our elections don’t have
median-crowding, because the crowding is at the phony center between Dem &
Repub.  …though Bernie is probably an instance of Republocrat strategic
median-crowding.

…

3. Polls confirm it.

…

That’s why I propose Fav-Defeater as now my favorite autodeterrent method:

…

Elect the highest Topcount (or Midcount) candidate who pairbeats the
highest Topcount candidate.

…

Offensive strategy can’t affect anyone’s Topcount.

…

Anyway, Chris’ example:

…

46: A (the CW)

…

44: B>C  (burying A)

…

10: C (the Bus)

…

…is quite extreme, with the CW nearly being a majority-CW. I wanted
something more typical.

…

So I suggest the following example as typical.

…

CW is the Topcount-winner, because, as I mentioned, the CW is the most
likely Topcount-winner.

…

The BF faction is larger than the Bus faction, because I didn’t want to
favor Bus.

…

Given those requirements, I wanted the factions as nearly equal in size as
possible, because, that seems the center about which they’ll vary.

…

Here’s my example:

…

35: CW>BF

…

33: BF>Bus

…

32: Bus>CW

…

With only one exception, all of the intended autodeterrent methods that
I’ve proposed here, including CTE(Coombs(whole)), elect BF.  i.e. they fail
autodeterrence in this very typical example.

…

Failing in that example seems to definitely say something
evaluation-meaningful.

…

But Fav-Defeater elects Bus, & thus succeeds in this example…directly from
its definition…which was the intent of that definition.

/…

A few methods’ choices in that example:

…

DAC & DSC: CW

…

Margins-Sorted Topcount:  CW

…

Fav-Defeater: Bus

…

MinMax(wv), Schulze, & MAM (RP(wv)): Bus

----

Fav-Defeater, MinMax(wv), Schulze,  MAM also elect Bus in Chris’ example.

…

So far then, considering the typicalness of my example, & how different
Chris’ example is, Fav-Defeater, MinMax(wv), Schulze & MAM seem the best &
most versatile ranked-methods..

-----

I don’t understand the motivation & justification of the double-sorting
methods such as Margins-Sorted Topcount. Neither would the would the
public, & it’s something essential for public-acceptance.

-----

If we dare to be optimistic, we can hope that the public will use the good
ranking methods well, & in that case the best Condorcet methods, & even
Hare, would be fine.  But the voting behavior demonstrated by the many,
many people completely wedded to the protection of their precious
lesser-evil isn’t encouraging at all.

…

I used to say that we need Condorcet, to reassure the lesser-evil voters.
But it likely won’t reassure them enough. I told why, even with the best
Condorcet methods, I’d rank the Acceptables in order of winnability instead
of merit, because I regard candidate-merit as dichotomous.  …Then why
wouldn’t our lesser-evil progressives, for the same reason, do the same, &
rank the Democrats over the progressives, topvoting the Dems with the most
media-promotion.

…

So I don’t have confidence in the lesser-evil progressives with a
ranked-method.

…

The pure Cardinal methods, Approval & Score, are the most foolproof
methods. They never give any possible incentive to not topvote Favorite.   …or
any incentive to vote anyone over Favorite.

…

They’re the public political voting-systems that I want for our timid
lesser-evil electorate.

…

The ranked-methods are a morass.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20240203/a95b4790/attachment-0001.htm>