[EM] Condorcet strategy and anti-strategy measures

[MIKE OSSIPOFF]

[EM] Condorcet strategy and anti-strategy measures

James A. wrote:

	And oh, Mike, my name is James Green-Armytage. "Green" isn't my middle
name, but is rather a part of my surname. Hence "James A." is not an
appropriate abbreviation for my name.



I repliy:

Sure it is. The purpose of an abbreviation is to briefly identify. "James 
A." does that just fine.

You continued:

If you feel the need to abbreviate
it,...

I reply:

It's either that or write out the entire name every time I refer to you or 
speak to you.

You continued:

...I suggest "James G-A". There's nothing at all wrong with the name
"James Armytage", except for the fact that it's not my name.

I reply:

James G.A. isn't your name either. An abbreviation for your name is 
something that isn't your name, but is briefer.

The trouble with James G.A. is that we already have a James G. "James G.A." 
unnecessarily resembles "James G."

The only part of your initials that distinguishes you from James Gilmour is 
the Armytage part. Hence there's no reason to represent any other part in 
the abbreviation, whose purpose is just to identify you, not to accurately 
name you.

Your letter, to which I'm posting a few reply comments now, has been very 
well replied-to by others already. So i'll only make a few comments.

I should point out that I'm repeating, probably nearly verbatim, what I said 
when I replied to these very same statements of yours before.


I'd said:

>First, I don't believe that Condorcet wv has a serious strategy problem. 
>We've often discussed the defensive truncation strategy. Here's how I
>often word it: The only way you can steal the election from other voters is 
>if they were trying to help you. Doesn't that make you proud of yourself?

	I do believe that Condorcet wv had a serious strategy problem.
	While it is unknown how likely a successful execution of the burial
strategy would be in a public election, I believe that if it did occur, it
would produce an *extremely* negative reaction towards both the result and
the election method itself.
	Imagine if Kerry beat Bush by a few hundred thousand votes, and yet Bush
actually won the election because most of the Bush voters strategically
ranked Nader second, creating a fake Nader > Kerry victory to knock Kerry
out of the running.

I reply:

As I said at least twice when we were discussing this before, wv's problem 
when offensive order-reversal is being attempted is no worse than the 
problem that IRV has without anyone attempting offensive strategy.

You can call wv's offensive order-reversal problem a serious problem if you 
want to, but then you should also say that all the other voting systems have 
a more serious problem. A similar problem that exists all the time, instead 
of just when some voters attempt a very risky offensive strategy.

With IRV, you'll often need to bury your favorite in order to keep your last 
choice from winning instead of the CW. With Condorcet, the CW will win 
unless someone actively steals the election from hir. And attempting to do 
so is quite risky, requiring impossible predictive knowledge.

But if the offensive strategizers were sophisticated and well-informed 
enough to know when to try that, how do you justify an assumption that the 
other voters won't be sophisticated and well-informed enough to counter it?

And no, they don't have to know that the offensive order-reversal is being 
done. They can do the defensive truncation, and, without the offensive 
order-reversal, the defensivse truncation will have no effect, no harm.

If you feel that the electorate, or at least some segment of it, is so 
devious that you expect them to be inclined to order-reverse against your 
favorite, then have some representatives of your faction publicize that 
they'r e going to defensively truncate, not rank anyone whom they like 
better than their favorite, whom they believe to be the CW. And remember 
that their judgement and polling information availability is just as good as 
that of the would-be offensive order-reversers.

In fact, if I didn't like a candidate, I just wouldn't rank hir. That's out 
of principle, but it also protects against offensive order-reversal by hir 
voters. The candidates whom  I don't like seem more likely to encourge their 
voters to offensively order-reverse, or to have voters inclined to 
offensively order-reverse.

And, as Adam pointed out when we discussed this here before, truncation is a 
much milder thing to do than order-reversal. Much easier and more likely to 
do spontaneously. Much easier to convince someone to do as a strategy.

And, as Adam pointed out, a successful offensive order-reversal, on a scale 
sufficient to change the outcome, really depends on some public organizing, 
which couldn't remain undetected by its intended victims. Why should they 
help the candidate whose voters are planning to try to steal the election 
from them? ? All they need do is not rank hir, and the offensive strategy 
will backfire badly.
That would be understood by the would-be offensive reversers.

In fact, the mere fact that, in a known devious electorate, many Democrat 
voters are likely to cautiously truncate the Republican,  aside from any 
specific knowledge of a planned reversal, would, by itself, be enough reason 
for the Republicans to not try the reversal.

You say that offensive order-reversal would cause a scandal if it succeeded, 
and that people would blame the method. Half correct. It would cause a 
scandal. People wouldn't blame the method, because it would be understood 
that all method have strategy, and all simple methods have at least the 
potential for considerable defensive strategy need. And that all the other 
simple methods are worse in that regard.

The scandal would be directed at the perpetrators of the offensive 
order-reversal. If the Republicans did that to the Democrats, that would 
completely undermine and destroy the alliance between Democrat voters and 
the Republicans. Would it be worth it to the Republicans. The Democrats 
would know that the Republicans were able to steal the election from them 
because they tried to help the Republicans. I doubt that they'd rank the 
Republican candidate again.

You continued:

	If this was a single-balloting wv Condorcet election, and Bush won (as
above), there would be an enormous outrage, and it would be entirely
justified.

I reply:

Quite so. Outrage by Democrat voters against the Republican voters, the 
Republican Party, and the Republican candidates, whom the Democrats wouldn't 
rank again.

At the very least, at their very most generoius, they'd use "Tit-For-Tat": 
Retaliate against the Republicans' in the next election.  TFT says: Copy the 
other player's most recent play (co-operation or defection).

Even so simple a measure as TFT would discourage offensive order-reversal.

Someone also pointed out some interesting and effective ways of dealing with 
offensive order-reversal, measures that hadn't occurred to me--such as 
dividing the win between the winner and whoever would win if the winner were 
deleted from the rankings. Maybe, as that person said, give the 
vice-presidency to the one whom  the winner pair-beats. Or maybe divide the 
presidency between them, time-wise. Or divide it between them by making the 
presidency a committee consisting of the winner and those whom he pair-beat.

Or divide victory between them probabilistically, which is really just as 
good. Or use Random Ballot to choose between the winner and the candidate 
who'd win if the winner were deleted from the ballots. So RB would be used 
to solve a tie between the winner and the candidate who'd win if the winner 
were deleted from the ballots.

Random Ballot (RB):

Randomly choose a ranking. Whichever member of the tie is ranked higher on 
that ballot is the winner. If 2 or more are ranked equally highest, among 
the tied candidate, they're the new tie, and so we randomly draw another 
ballot to solve that new tie...etc.

[end of RB definition]

All these solutions should be added to the list of wv enhancements:

If there's a circular tie in which everyone has someone else ranked over hir 
by a majority, then:

Divide the presidency (time-wise, probabilistically by RB, or as a 
committee) between the winner and the candidate who'd win if the winner were 
deleted from the ballots.

Or make the candidate(s) who'd win if the winner were delted from the 
ballots the vice-president.

So many 1-balloting enhancements have been described for wv, that it 
certainly isn't necessary to hold a 2nd balloting.

I suggest that, therefore, the benefit of a 2nd balloting is unlikely to 
justify the advantage of one.

And I wouldn't suggest RMDD as a 1st wv proposal, due to its drastic way of 
getting rid of the LO2E problem. Maybe I'd offer it as a later proposal, 
though. Ideally, I'd like RMDD, unless, due to people running an identical 
candidate and repeating their offensive order-reversal, RMDD went though 
many ballotings without a winner. The risk then would be that the elections 
would break down. An alternative would be to, after 1 or 2 extra ballotings, 
just use the solution described above, or do an Approval balloting. Then 
SSSC would only be met in the 1st ballotings. Of course if it were known 
that the reversers would run an identical candidate and reverse again and 
again, which will result in the eventual need to relax RMDD and let an 
existing candidate win at some point, then SSSC isn't being met in a 
practically meaningful sense, even in the 1st balloting.

So I know that RMDD has its problems, as does every method. But its offer of 
SSSC compliance justifies giving RMDD a try at some point, as we eventually 
try to further refine Condorcet wv, and look for more and more perfect 
voting systems.


James A. continued:

	In that sort of situation, you would really want to hold a second
balloting.

I reply:

That's one solution. Others include the ones suggested in reply to your 
posting, such as dividing the presidency between the winner and the 
candidate(s) who'd win if s/he were deleted from the ballots and the ballots 
recounted.

And I and others have suggested a number of other 1-balloting enhancements.

James A. continued:

In between ballotings Nader would be given the opportunity to
drop out of the race, and hopefully he would do it

I reply:

Nader would not drop out of the race.

Nor will he do so this year.

Kerry has no ownership claim on the votes that I and many others would 
prefer to give to Nader.

That many don't want to vote for Kerry is no one's fault but Kerry's. Does 
it occur to you that if Kerery were honest, maybe we'd vote for him?  It 
isn't because of Nader that we don't vote for Kerry--it's because of Kerry. 
I, for one, wouldn't vote for Kerry even if Kerry and Bush were the only 
candidates allowed in the race.

Nader is a candidate for people who have no use for sleaze, who will not be 
resigned to voting for bribetaking Republocrats. Nader won't drop out for 
you. Nader is a candidate for people who will onlly vote for someone who is 
honest.

You're of course going to vote for Kerry. Of course not because _you_ like 
Kerry better than Nader. No, it's someone else, most people, you say, who 
are going to vote for Kerry (It's always someone else), and so you're going 
to follow them. Except that they're probably doing so because they're 
following everyone, including you. You're like a herd of sheep or cattle 
following eachother into a slaughter house.

Who's splitting the vote? The people who will only vote for an honest 
candidate, who won't vote for a sleazy bribetaker? Or the people who instead 
vote insincerely for the bribetaker?

Also, you attribute all the the Bush administration is doing to Bush. You 
forget that the media and various corporations, and lots of Democrats, 
including Kerry, were right behind him with the Patriot Act and the 
obviously fraudulently-justified bloody war of conquest against the Iraquis. 
Kerry voted for the Patriot act and the invasiion of Iraq. He talks 
progressive now, because he knows that's what people want to hear. He knows 
that you're so desperate that you need and want to believe him. It's the 
same in every election. The Democrats are progressives in the campaign and 
Republicans in office, because they get their votes from one segment of the 
population and their money and instructions from another (smaller) segment 
of the population.

Returning to your topic, and replying to your summary: You exaggerate 
unenhanced 1-balloting wv's strategy problem when you ignore the worse 
strategy probems of all the other simple methods, including IRV and 
Plurality.  And when you disregard all of the 1-balloting enhancements we've 
proposed.

You continued

	Mike, in your analysis of Condorcet strategy, you seem to be assuming
more coordination, more communication, and more time for people to respond
to each other's strategies than what I think is necessarily realistic. In
a previous e-mail on this topic, you made many statements similar to the
following:

I reply:

As others have pointed out too, the offensive order-reversal will need a lot 
of public discussion and organizing to succeed. Whether it's official 
organizing, or just bar-room conversations, letters to the editor, bus-stop 
conversations, etc., the intended victims are going to hear about it, and 
they won't vote for the perpetrator's candidate, and the perpetrators will 
wish they didn't attempt the offensive strategy.

You're assuming that the perpetrators have much better judgement, 
sophistication, and polling inforamtion than the intended victims, and that 
the intended victims won't hear any of the discussion about the intended 
reversal.


I'd said:

>To the extent that there's any perceived danger of offensive
>order-reversal, defensive truncation will be employed as a
>countermeasure.

	The thing is, the people who would be employing the "countermeasure" will
not know whether the strategy they are attempting to counter has taken
place, until after the election!

I reply:

They'll have heard discussion about intent to offensively reverse. And if 
the intended victims won't know exactly what's going to happen, neither will 
the would-be reversers. That's the part that you're missing. They'll know 
that there might be counterstrategy, the simple counterstrategy of not 
ranking the reversers' candidate. You're assuming that the reversers have 
better information than the intended victims.

You continued:

	As you know, there are various examples where if voters truncate against
a strategic threat that in fact does not exist to begin with, it can
result in the election of the car crash candidate in itself.

As Adam said, do you have an example? Assuming (excuse me while a I raise 
the tablecloth to puke) that Kerry is CW, can Nader win because the Kerry 
people didn't rank Bush?, if Kerry is CW and other people vote sincerely?

But sure, truncation can backfire in any method, because no method can help 
someone who doesn't vote for a compromise that s/he needs. Defensive 
strategy can hurt you in any method. The moreso in IRV or Plurality, for 
instance. I sometimes point out that Approval needs twice as many mistaken 
compromisers than IRV and Plurality need, to give away an election.

Look, the whole problem with voting strategy is that we don't haver perfect 
info. If we did, the CW would win every time, in any method. So then, let's 
not be criticizing wv for the consequences of not having perfect info, when 
those consequences are worse with other methods, such as IRV & Pluralilty.

You continued:

	You seem to assume that a strategy of Dean voters to bury Kerry would
involve someone in the Dean high command calling up various supporters and
telling them to vote against their conscience. Then of course, the Kerry
high command would find out about it, and call up the Dean high command
and say "Hey! We know what you're up to! So if you don't tell your guys to
vote sincerely, we'll tell our guys to truncate against Dean!" In which
case it would be fairly cut and dry.

I reply:

As I said, successful reversal is very unlikely without discussion, and that 
discussion can't be kept from its intended victims. And yes there is 
communication, when the CW's people post on the Iternet, and write letters 
to the editor, and demonstrate in front of TV cameras that they're going to 
not rank Bush. And talk about it on streetcorner conversations, etc.

You continued:


		What it may come down to is: To what extent do voters on each side have a
propensity to gamble with burial strategies?

I reply:

Not much, when you consider the risks, which are arguably more than just 
risks, and are fairly llkely consequences.

You continued:

To what extent do voters on
each side expect voters on the other side to engage in burial strategies.

I reply:

Obvioiusly that depends on the electorate. In our EM poll I didn't expect 
order-reversal, though I allowed for it by saying that after the results are 
announced, people have a week to truncate or promote candidates to 1st 
place.

If the electorate is known to be devious (and deviousness wouldn't stay 
unknown long), or if people overhear people in cafe's talking about 
insincerly downranking a likely CW, then or course they'll have reason to 
expect reversal. The reversers should know that and should know better than 
to go through with it. And no, when the would-be reversers decide not to 
reverse, that doesn't cause a penalty for defensive truncation or defensive 
equal ranking.

And additionally, all your talk about the game of chicken disregards the 
defensive equal ranking that Adam described.




	The result of such a situation is unpredictable even if you know the
sincere preference rankings, and that unpredictability is very dangerous.

I reply:

Without perfect information, that unpredictabiilty is present in all voting 
systems, including IRV & Plurality. It isn't the result of the voting 
system. But some voting systems suffer from it more than others do. IRV & 
Plurality suffer from it more than wv & Approval do.

You continued:

It is quite possible that successful burial will take place without anyone
on the other side stopping it.

I reply:

Unlikely, for the reasons described. But it's perfectly possible that IRV 
will make you regret voting sincerely. That's what you should be worrying 
about. Yes, a 2nd balloting can improve wv. So can lots of 1-balloting 
enhancements that we've described. For a 1st wv proposal, the 2nd 
balloting's benefit therefore isn't worth its acceptance disadvantage. As 
Markus said, a big selling point of rank methods is that they don't need 2 
ballotings.


It is also possible that a combination of
mutual burial and mutual truncation (due to either aggressive or paranoid
voting) could lead to the election of the car crash candidate.

I reply:

Mutual truncation won't do that. And, given a particular set of polling 
data, there won't be candidates who'd both try offensive order-reversal 
against eachother.

You continue:

Both
possibilities are chaotic and fundamentally undemocratic.

I reply;

Again, you're holding wv up to a standard by which IRV & Plurality are much 
worse.

You continued:

I think that 2 balloting wv Condorcet (2nd balloting
only in the event of a cycle

I reply:

Don't forget that 2nd balloting methods should only trigger a 2nd balloting 
when there' s a circular tie all of whose membes are _majority_ beaten. And 
don't just say "cycle". An election result can have a cycle without having a 
circular tie. It can have a BeatsAll candidate, and some losers who are in a 
cycle. So it's better to say "circular tie" or "top cycle" than just 
"cycle".

In your 2-balloing system, if the triggering conditions occurred, I'd be 
among those telling Nader not to drop out. Why just have bribetakers in the 
election?

And this discussion has been assuming that Kerry is CW. Let's remember that 
that is only an assumption for the sake of discussion. There's no particuar 
reason to believe that it's true, except for what your tv commentators have 
been drumming into you all this time.

They speak of "the middle", as a point somewhere between the Republicans and 
Democrats. No, there's no reason to believe that the voter median is there.

That's where the media "mainstream" is, sure.

You continued:


	Going back to the two-ballot procedure, I still think that the second
balloting should be a Condorcet balloting. I prefer a Condorcet balloting
to an approval balloting here for all the reasons that I prefer Condorcet
to approval in general.

I reply:

But the information from the 1st balloting would give people the information 
to deal with offensive order-reversal in an Approval vote. And Approval 
would never give us the conditions that call for another balloting. What if 
you use Condorcet for the 2nd balloting, and again there's a circular tie 
with all majorilty defeats?

Mike Ossipoff

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