[EM] Ease of Voting

[MIKE OSSIPOFF]

I agree with most of what you said. Condorcet is the easy-voting method. 
Little or no need for any kind of strategy. Just rank sincerely, as we did 
in the recent EM poll.

It's true that IRV has strategy problems that will sometimes be serious.

Approval isn't so difficult. Its strategy is no more complicated than that 
of Plurality, and Approval doesn't have the major dilemmas that Plurality 
has.

I ususally suggest these Approval strategies:

I. If you rate the candidates in 2 sets such that the merit differences with 
the 2 sets are negligible compared to the merit difference between the 2 
sets, then vote (only) for the better set.

II. Otherwise:

A. If it's a 0-info election (no information about other people's 
preferences or which candidates are most likely to be the top 2 contenders), 
  then vote (only) for the above-mean candidates.

B. If it isn't 0-info:

1. Best Frontrunner strategy:

This is how people usually vote in Plurality: Vote for whichever of the 2 
expected top 2 contenders you prefer to the other. But, since it's Approval, 
vote also for everyone whom you like better than hir.

At the electionmethods.org website, at the Approval Strategy pages, I 
describe refinement and elaboration of Best Frontrunner.

2. Better Than Expectation:

Vote for every candidate whom you'd rather have in office than hold the 
election. That is, every candidate who is better than your expected value of 
the election, _given the assumption that the candidate currently under 
consideration is one of the top-2 contenders_.  (It's clearer to say it that 
way, but more precisely, one of the 2 candidates between whom your ballot 
can make or break a tie).
Remember that if that candidate is, for instance, a progressive, then the 
electorate must be rather progressive, and that affects your estimate of the 
value of the election under the assumption that s/he is a top-2 contender. 
Is s/he better than the value of the election under that assumption? Then 
vote for hir.

3. Better/Worse:

Pb is the probability that if X is one of the top 2 contenders, the other 
top 2 contender will be someone better than X. Pw is the probability that 
the other top-2 contender will be someone worse than X.

Db is the expected amount by which the other top-2 contender will be better 
than X if s/he is better than X.

Dw is the expect amount by which the other top-2 contender will be worse 
than X if s/he is worse than X.

Vote for X iff:

PbDb < PwDw

That can be rewritten:

Pb/Pw < Dw/Db

Either of those is good; it's just a matter of whiich seems easeir to 
estimate.
Again, though, the estimates for Pb, Pw, Db & Dw should take into account 
your assumption that X will be one of the top 2 contenders.

If X is a progressive, and is a top-2 contender, then of course other 
progressives have an improved chance of being the other top-2 contender and 
that affects your estimates for those 4 quantities.

Mike Ossipoff

_________________________________________________________________
FREE pop-up blocking with the new MSN Toolbar – get it now! 
http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/