[EM] Threat/Promise Approval strategy. Summary of strategies.

MIKE OSSIPOFF nkklrp at hotmail.com
Sat Feb 12 11:25:02 PST 2005


Choice of Approval strategies depends largely on what you prefer to 
estimate.

I suggested a  strategy that I called Expected Differences, which, for 
deciding whether to vote for candidate X, requires judging the probability 
that X will be in a tie or near-tie with someone better, or someone worse, 
and the expected utility difference between that better or worse candidate 
and X.

Forest suggested Better/Worse, which merely requires an estimate of whether 
the winner is likely to be better or worse than X.

But how about another simplification of Expected Differences:

Just ask whether the threat from worse candidates than X is greater than the 
promise of better candidates than X.

That sounds like Better/Worse, but, instead of asking whether someone better 
is more likely to win than someone worse, it asks whether the expected 
worseness is greater than the expected betterness. Arguably one has a feel 
of whether fear of worse is greater than promise of better, more than one 
has an estimate of whether worse is more likely than better.

Threat/Promise, along with Better-Than-Expectation,  is probably the next 
simplest thing than just asking yourself if you feel like voting for X.

Like Better/Worse, Threat/Promise is very similar to Forest´s 
Greater-Than-Expectation strategy, which asks if X is better than the 
expected outcome. In other words, Greater-Than-Expectation says to vote for 
X if you´d rather have X in office than hold the election.

It´s a small step from saying to vote for X if the expected worseness from 
someone worse winning is greater than the expected betterness of someone 
better winning to saying to vote for someone if s/he is better than the 
election´s expectation.

Usually you´ll know whom you want to vote for and whom you don´t want to 
vote for, but, when you don´t, that´s when strategies such as these and the 
others we´ve discussed will be useful.

If there´s a strong feeling about two candidates being the likely 
frontrunners, then maybe one would feel like using one of the Best 
Frontrunner (BF) strategies that have been discussed here.
Having recently posted a list of those BF strategies, and suggestions for 
choosing among them, I shouldn´t repeat that here. BF is important since 
it´s what voters say they´re using now, though they´re using it without 
reliable information.

Strategy A sounds very different from BF(1st), but it amount to the same 
thing. BF(1st) is basically a rewording and renaming of Strategy A. It seems 
to me that BF(1st)´s wording is simpler and more obvious and that its name 
more descriptive.

If there´s no information and no estimates other than the 
candidates´utilities, then use the 0-info strategy of voting for the 
above-mean candidates.

And lastly, if the candidates are in 2 sets such that the merit differences 
within the 2 sets are negligible compared to the merit difference between 
the two sets, then of course vote for the better set candidates and not for 
anyone else. That isn´t really a different strategy, since all the 
strategies lead to it. But when its premise is true it makes voting much 
simpler. I claim that its premise is true in our political system.

Even if you don´t agree with me about the last sentence of the paragraph 
before this one, it would be difficult to deny that our elections are 
0-info. So the two paragraphs before this one are my suggestions for 
Approval voting in the elections in the U.S., with the current political 
system, because the conditions match the premises of those two paragraphs.

I know I´ve recently said this too, but, in Plurality the 0-info strategy is 
to vote for your favorite.

And, in Plurality, if there are 2 sets of candidates such that the merit 
differences within the two sets are negligible compared to the merit 
difference between the sets, then vote for whichever better-set candidate is 
most likely to take victory from a worse-set candidate.

Mike Ossipoff

_________________________________________________________________
Is your PC infected? Get a FREE online computer virus scan from McAfee® 
Security. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963




More information about the Election-Methods mailing list