[EM] IRV strategy?

MIKE OSSIPOFF nkklrp at hotmail.com
Sat May 18 23:41:42 PDT 2002


Don wrote:

05/18/02 - Rank Your Lowest Poll Favorite as Number One:

Mike, you wrote about IRV: "...it eliminates your favorite before it lets
you help a lower choice."

If your favorite is lower in the polls then you can expect that it will be
eliminated before your second choice.

I reply:

But it might get eliminated right after it eliminates the only
compromise that can keep your last choice from winning.

But yes, with IRV one can expect the elimination of whatever candidate
currently tops fewest ballots, since that's what IRV does.

Don continued:

What you should do is rank number one whichever of your two candidates is
the lowest in the polls. If your favorite is the lowest, you rank him
first. In that way you may help your favorite to also avoid the
elimination that your lower choice is going to already avoid because he has
a higher poll rating.

I reply:

There's no reason to rank your 2nd choice over your favorite just
because you're sure that your 2nd choice will poll lower than,
and get eliminated before your favorite.

The necessary strategy is: If there's a candidate whom you believe
is the one most likely to be the best compromise you can get, the
one who can get a majority against those you like less, then rank
him/her 1st. Rank the rest in order of their liklihood of being
that candidate.

If you feel that there are candidates who are completely unacceptable,
though winnable, then rank all the acceptables in order of their
probability of being the one who can, with your help, deny victory
to an unacceptable candidate.

These strategies of course often will involve favorite-burial, something
that Approval will never give you need for, and which Condorcet(wv)
won't give you need for under the conditions described in
SFC, GSFC, WDSC, & SDSC. And, when there's a sincere CW, Approval &
Condorcet(wv) always have an equilibrium in which that CW wins and
no one order-reverses, while IRV & Condorcet(margins) have plausible 
situations
where the only equilibria are ones in which defensive order-reversal
is used.

Don continued:

But, the big question needs to be put to you: `Just how many candidates do
you expect to elect with your one vote in a single-seat election, or for
that matter, in a multi-seat election like STV?'

I reply:

I assume that by "vote", you mean ballot.

Amazing answer: a) One candidate is elected in a single-winner
election. But whether or not your voted preferences are counted
reliably depends on whether or not we're regrettably conducting the
election with IRV.

Mike Ossipoff


_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp.

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em



More information about the Election-Methods mailing list