[EM] Markus & Manipulability

[Forest Simmons]

My thinking about Nurmi and Bartholdi's measure of manipulability is
similar to Mike's.

Consider the example of Perverse Random Ballot: the winner is the
candidate at the bottom of the list on a randomly chosen ballot.

The optimum strategy is completely obvious: vote your favorite last.

[As a side note, this is like the shower valve that has "hot" and "cold"
reversed. Richard's point about isotone versus monotone applies.]

According to Nurmi and Bartholdi this method is highly manipulable because
the voter can easily subvert the system to increase his expected utility
by voting strategically.

But notice that under Perverse Random Ballot the voter cannot be easily
manipulated into voting against his own best interest.

Manipulating a voter into voting against his/her own best interest is the
anti-democratic kind of manipulation.  Nurmi and Bartholdi's levels do
not address this most important kind of manipulation.

This example shows clearly that Nurmi and Bartholdi are worried about
voters "manipulating" the system to increase their expected utilities,
i.e. to vote in their own best interest, as though voters' utilities had
nothing to do with social utility.

It is the prerogative of the voter to maximize their own utilities,
whether anybody else thinks they have social value or not. That's
democracy. We don't try to use voting methods to protect the public
against the public will. We use voting methods to ascertain the public
will. 

A more realistic worry is pollsters and pundits manipulating the voters by
fooling them into voting against their own best interest. The more
complicated the strategy and the more sensitive good strategy is to
information (like the other shower valve with hysteresis near its hot/cold
boundary) the easier the "experts" can manipulate the vote of the
gullible voter.


Forest


On Sun, 3 Feb 2002, MIKE OSSIPOFF wrote:

> Markus, we went over this a long time ago on this list.
> 
> Markus wrote:
> 
> Due to Nurmi and Bartholdi, the more information you need about the
> opinions of the voters resp. the more accurate this information must
> be to be able to calculate a strategy, the less vulnerable to strategies
> the used election method is. To my opinion, this argument by Nurmi and
> Bartholdi is plausible.
> 
> I reply:
> 
> Markus, there's nothing wrong with copying & repeating what you've
> heard from academic authors, but it would be good to also question
> a what you read from them.
> 
> When academic authors write about methods with regard to strategy,
> they tend to speak of vulnerability to strategy. Strategy is
> taken to mean offensive strategy, by which someone is manipulating
> the method.
> 
> It's (almost) common knowledge on this list that strategy is important
> as an undesirable _need_ for voters, rather than as a manipulation
> opportunity. However that's something that most academic authors
> , and those who worship them, remain ignorant of.
> 
> When there's a problem with offensive strategy, the problem is
> the problem that that offensive strategy causes for others, the
> defensive strategy dilemma that it causes for them.
> 
> I'm almost certain that it was Nurmi who, in one of his books or
> articles, rated the methods on their vulnerability to strategy.
> IRV rated best or near best in that regard, because offensive
> strategy is difficult in IRV. But what good does that do, when
> defensive strategy is necessary in IRV, regardless of whether or
> not anyone is using offensive strategy?
> 
> Is it possible that the academics & their loyal are unaware that
> Plurality has a strategy problem not because someone can offensively
> manipulate it, but because Plurality forces voters to use a drastic
> defensive strategy, a kind of strategy needed to protect a sincere
> CW, or to enfore majority rule?
> 
> With any non-probabilistic voting system, a majority can get their
> way about something that they all agree on. If there's an alternative
> that they all want to win, they can make it win. If there's an
> alternative that they all want to lose, they can make it lose.
> The former is usually easy, but the latter, with most methods,
> can require members of that majority to vote something else over
> their favorite.
> 
> Strategy that's needed to protect the win of a sincere CW, or to
> enforce majority rule--I call that "defensive" strategy. It's the
> old "lesser-of-2-evils problem", and it's the reason why most of
> us want a better voting. And still the academics & their obedient
> copiers don't seem to understand that, and still consider
> "strategy problem" to mean vulnerability to manipulation.
> 
> Markus once said that Plurality's strategy problem is a vulnerability
> to manipulation, because if Bush ought to win by 1st choices, then
> the Nader preferrers who vote for Gore are manipulating the
> voting system when they elect Gore. That sort of thing is why
> I sometimes say that the academics have their heads up their ass.
> 
> (Markus, you know that doesn't strictly apply to you, if you aren't
> an academic, but merely an academic loyalist).
> 
> Plurality's problem is not that Nader preferrers can make Gore win
> even though Bush is favorite of most. It's problem is that Nader
> preferrers need to dump Nader & vote for Gore to enforce the
> majority against Bush.
> 
> Mike Ossipoff
> 
> 
> 
> 
> 
> _________________________________________________________________
> Join the world’s largest e-mail service with MSN Hotmail. 
> http://www.hotmail.com
> 
>