Probabilities

[Elisabeth Varin/Stephane Rouillon]

MIKE OSSIPOFF a écrit :

> Steph--
>
> You wrote:
>
> I almost fully agree with you about what can happen.
> I just really do not agree with those events probabilities...
>
> I reply:
>
> Wait a minute--how can you disagree with a probability? If you
> mean that you don't agree with my statements about those probabilities,
> then you're right: Of course you don't agree with my statements about
> those probabilities, because I've made no statements about those
> probabilities. You understandably don't agree with statements that
> haven't been made.

You have made no statement, but some different appreciations of the
"occurences" of particular events definitively appear by comparing our
relative points of view.... This is an example extracted from one of your
last e-mail:
"But, as we've discussed here many times, order-reversal isn't likely to be
used on a sufficiently large scale."

> You continued:
>
> Polls for preferential ballots will rarely be reliable.
>
> I reply:
>
> Some polls will be more reliable than other. Ignore polls other than
> by people whom you trust. But my claims about wv vs margins assume
> that polls are always reliable?

Adam's analysis does, or at least uses them considering they are reliable
enough.

> You continued:
>
> Voters do not vote in block.
>
> I reply:
>
> I'm not quite sure what you mean by that. Did I say that they
> vote in block? Often, however, there are more than 1 voter voting
> the same pairwise preference.

Yes , but they do not hold reunions like party fans do. Thus, it could be hard

to gather them in order to persuade them to modify their vote (by truncation
or slight order-reversals...)

> You continued:
>
> Violation of majority rule with margins or relative margins
> is incredibly improbable.
>
> I reply:
>
> Ok, it's you who are making probability claims. Unsupported claims.
> I've often posted plausible, ordinary examples in which margins
> and relative margins violates majority rule. No, I didn't calculate
> the probability, and you haven't suggested a way of estimating
> it.

Maybe you did not read all my mails...
>From two of my previous mails one sent to Adam via [EM] , the other sent to
YOU:

"But what is more realistic? Identifying and convincing 18% of the voters to
tilt their preferences
from poll predictions as in my case, or 49% in your case. As you said my case
is very unstable,
I bet I would convince nobody. But what about yours? Even if I admit it is
very more probable
than mine, how probable is it? 1% of the cases? We definitively need to
compute this number,
even for the not so simple 3 candidates case with a small number of voters.
Compared to the polls
precision of 9 different ballots (A, B, C, ABC, ACB, BAC, BCA, CAB , CBA) I
think you would still
convince nobody. Finally, with 4 candidates or more, forget polls precision,
forget identifying voters
with the same rankings, forget strategy at all. It has no mathematical
relevance. It is too small cases.

Since sincere preferences seem the most probable input, I think we should
consider an equal probability
of the 9 different ballots. I hope we can create a stealing CW case with few
voters, because only with
4 voters we have 6561 equiprobable simulation cases...

The next thing we will have to do is look at what happens when the number of
candidates or/and voters
inceases. Would the ratio (CW can be stolen cases / voting simulations)
increase or diminish in each case? "

and

"So both criteria can harm a Condorcet Winner. The question is: how big is the

majority rule violation in both cases? Give me numbers from an A, B, C, AB,
BC,
AC, BA, CB, CA distribution. I will do it myself, but for a small number of
voters. I sound like M. Carey, working with 3 candidates and 4 voters..."

> You continued:
>
> IRV will rarely elect dangerous extreme candidates.
>
> Just as I said above, about margins & relative margins, you're
> making another unsupported claim. Very many ordinary, plausible, typical
> examples have been posted in which IRV violates majority rule
> by electing an extreme candidate. But you say "dangerous" extreme
> candidate. The dangerousness of extreme candidates isn't something
> that you can make objectively verifiable claims about. But if you
> want to make such a claim, you should try to justify it, as with all
> the other unjustified claims in your postings.

Sorry, I misread your sentence:
"When you elect someone away from the voter-median, you violate majority
  rule. Voting systems, like IRV, that jump to extremes are dangerous."
IRV jumps, yes. But not always to extremes. When voters refuse to
give second and more preferences, the voter-median then becomes the
plurality choice and IRV finds it too... You do make unjustified claims too.

> You continued:
>
> Preventing unsincere truncation should not be the main issue
> of an electoral method.
>
> I reply:
>
> That's good, because I've never said that it should be or was the
> main issue. As I keep repeating, SFC & GSFC make no mention of truncation.
> Truncation will be common in elections however, and
> it will make a mess of those elections with margins and relative
> margins.

Only if their proportion is as big you think.

> You continued:
>
> Optimizing the wish of the electorate should.
>
> I reply:
>
> Then let's not use voting systems that force voters to reverse
> sincere preference and bury their favorite, as IRV, margins, and
> relative margins do.

It does not force people to do so. It only encourages them to do it
more, not to do it. You should understand the difference.
The mean gain of truncation without pre-election knowledge
is a loss, not a gain! So the estimation of a probability of making
truncations with those methods is the key factor.

> You continued:
>
> Still, thanks a lot for all the time you give me.
> And felicitations for your work on SFC.
> It is harder to build than to destroy...
>
> I reply:
>
> Is it that you believe that you've destroyed SFC? :-)
>
> Destroying or refuting things isn't as easy as you seem to think.

Gosh no, I am just finishing trying all I had to destroy it,
and now that I am finished. It is pretty obvious to me that SFC
proof is all good.I missed miserably. And I am glad I did.
You do extrapolate to much. It is harder building than destroying,
it does not mean I was succesful doing it! On the contrary, I am now
a real wv partisan when it comes to kill truncation. Ranked Pair (wv)
is now my second choice, and I will have to examine the two further
issues to see if it becomes first in fornt of Ranked Pair (rm).
1) what is that darn probability of truncation stealing a CW with
relative margins?
2)does wv protects from order-reversal as well, and if it does not,
does margin or relative margin do a better job?

> You've got to be a lot more careful with your meanings, careful that
> you're saying things that are relevant to the claims that you're
> making. You've got to justify your claims a lot better than you've
> been doing.
>
> Mike Ossipoff

You wrote:
"Governement isn't going to represent public wishes very
well if the voting system strategically forces people to not express
their wishes, to in fact reverse their real wishes on their ballot."

Again, rhetorical excess is in the eye of the beholder...

Steph.

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