[SPAM: #] [EM] Approval Strategy

[Neal Finne]

I don't know if this is considered within the scope of what is 
considered legitimate content on this list, but I'll send it anyway...

Forest Simmons wrote:

>Simple answer for the man on the street:
>
>Approve the candidate that you would vote for under Plurality, as well as
>every candidate that you like better.
>
>Civic minded voters can learn refinements of this basic (and perfectly
>adequate) strategy as easily as they can learn the rules of football and
>soccer.
>
I'd be reluctant to use any electoral system that requires voters to 
vote strategically. Still, within those systems, approval is a huge 
improvement over plurality or IRV.

>(1) Election outcomes are not the only "results" of elections: ask the
>Nader supporters in Florida and other swing states who didn't dare vote
>for him.
>
Or better yet, look at the results of Washington's 7th District 
congressional elections in 2000. Said district is a very liberal 
district, including Seattle proper plus maybe 100,000 people outside of 
Seattle. In that race, the Republicans didn't run a candidate, so the 
election became a contest between Jim McDermott (D) and Joe Szwaja (G), 
a highly qualified Green candidate (for whom I probably would have 
voted). McDermott won 70%-20%. I think that election provides ample 
evidence that the electorate as a whole prefers Democrats to Greens.

>Why do you suppose that Nader got the greatest percentage of votes in
>states like Alaska where he was so unpopular, and Bush so popular, that
>there was no chance of him spoiling things for Gore?
>
And Nader got 10% there, or 25% of the Nader-Gore vote (and surely less 
than that if the Bush voters had the chance to express second-choice 
votes, third choice votes, etc.).

>(2) Also, two party politics have conditioned us to think that there can
>only be two viable candidates.
>
>(3) Primaries are unnecessary under Approval voting; approve as many
>candidates from your party as you like.
>
>(4) When probabilities are involved, "almost surely" is safer than
>"surely."
>
>For example, polls may be misleading. It isn't impossible that (according
>to the polls) the two leading contenders A and B are running neck and neck
>(49 and 51 percent, say) while another candidate C is actually the
>favorite of 52 percent of the electorate, but out of favor with the
>corporate pollsters.
>
>20 A >> (various unapproved)
>29 C ... A >>
>23 C ... B >>
>28 B >>
>
>Since the voters have nothing to lose by approving favorite, under
>approval candidate C wins, even though every voter (following the basic
>strategy) approved one of the supposed front runners as well.
>
>Far fetched?  An informal Time Magazine Website poll put Nader far ahead
>of both Gore and Bush for the last several weeks leading up to the
>presidential election.  It was of the form: "Which candidate do you like
>the best?"  rather than "Which candidate do you intend to vote for?"
>
Online polls are meaningless. They do not represent the populace as a 
whole, or even the Internet-using populace as a whole, because they have 
self-selecting samples. I think I have also read that Greens are 
overrepresented on the Internet, but I can't provide a citation. Your 
example does look plausible, but not with Nader as C.

>The "for whom will you vote?" style polls always found Nader near the
>bottom.
>
>If the last US presidential election had been conducted under Approval,
>assuming that those who voted for Nader under plurality would approve only
>Nader, and assuming that more of the disenchanted (effectively
>disenfranchised) would have turned out to vote, the results could well be
>Nader 40% approval, Gore 37% approval, Bush 36% approval.
>
>The Time Magazine poll would have put Nader well above this.
>
>Nader was (arguably) right when he said Gore was the spoiler.
>
Let's look at this from a Condorcet perspective. As I believe I 
established above, Gore would have beaten Nader if they had been the 
only two candidates. Polls showed that among those who actually did vote 
Nader, 50% would have voted for Gore had Nader not run, 20% would have 
voted for Bush, and the remainder would not have voted. That means Gore 
would have beaten Bush. Neglecting other candidates, Gore would have 
beaten Bush, and Gore would have beaten Nader, so Gore was the clear 
Condorcet winner. Technically, you could call Gore a spoiler if Nader 
also would have beaten Bush in a two-way match-up, but I find that 
incredibly unlikely--just about all Gore voters would have had to prefer 
Nader to Bush--and even in such a case, voters preferred Gore to Nader.

Neal Finne
shadowdragon at softhome.net
"Si je savais quelque chose que me fût utile et qui fût préjudiciable à 
ma famille, je la rejetterais de mon esprit. Si je savais quelque chose 
utile à ma famille, et qui ne le fût pas à ma patrie, je chercherais à 
l'oublier. Si je savais quelque chose utile à ma patrie et qui fût 
préjudiciable à l'Europe, ou bien qui fût utile à l'Europe et 
préjudiciable au genre humain, je la regardais comme un crime." -- 
Montesquieu