From: "Kathy Dopp" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
To: "Stéphane Rouillon" <[EMAIL PROTECTED]>
CC: election-methods@lists.electorama.com
Subject: Re: [EM] New MN court affidavits by those defending
non-Monotonicvoting methods
Date: Thu, 6 Nov 2008 23:43:28 -0700
On Thu, Nov 6, 2008 at 10:49 PM, Stéphane Rouillon
<[EMAIL PROTECTED]> wrote:
> The spoiler effect is a special case of non-monotonicity.
>
> A general definition of a monotonic method is:
> no voter or group of voter could harm a candidate by expressing its full
> preference toward any higher preferred candidate.
No, not even close. That definition you are giving is not met by any
voting system I know of. Can anyone think of one?
Exactly, no electoral system can garantee coherence between the order of
preferences of a voter
and the impact the participation of that voter has on the result.
> While you restrict monotonicity definition to:
> no voter or group of voter could harm its favourite by expressing its
full
> preference.
Not even close. I simply use the mathematical definition of monotonic
functions. Look it up. It's very simple.
It is not so simple and you know it. For a monotonic function:
if pref (a) > pref (b) then result (a) > result (b)
For an election method a and b are not simple values but cover all
instanciations of every other ballots case. Because some of these
instanciations represent a collective incoherent will (containing a cycle),
generic monotonicity cannot be inferred for all cases.
>
> It is your choice. You chose to disregard the fact that winners, while
the
> voter expresses or not its full preferences, could both not be the
favourite
> of the voters.
I love the way you keep referring to voters as "its". It says
something about the way you view voters.
It says my mother tongue is french and I try not to discriminate between his
or her.
Huh? Your sentence above makes no sense to me. You want to try
restating it more precisely?
If, by voting or not for your favourite C, you could get either A or B
elected, I say the method is not
monotonic according to the general definition I use. Voting or not for C
should have no impact on the result between A and B.
>
> I do not understand why you want to consider the spoiler effect as a
> different problem.
Because it IS. Because Arrow and every other expert recognizes it as a
different problem than monotonicity. Because mathematically it is a
different problem, etc.
> As soon any voter would learn that its first choice has
These voter "its" again...
> no chance of winning, its second choice would become its new first
choice,
> the spoiler effect leading again to your personal definition of the
> monotonic dilemma...
Huh*!?
My personal definition? You mean the personal definition that ALL
mathematicians use? What do you think I'm so all-powerful that before
I was born I went back in time and forced all mathematicians to adopt
a definition for "monontonicity"?
OK I can see I'm wasting my time here.
If anyone wants to send me anything intelligent on this topic, please
email me personally, off-list.
Thanks.
Kathy
If you can't even recognize there is many definition to mononicity, you are
definitively wasting your time doing psephology. Search for mono-add-plump
or mono-add-top for example. By personal, I meant the one among these you
use. As you cannot provide one by yourself I had to. If as you say:
"I simply use the mathematical definition of monotonic functions. Look it
up. It's very simple."
Why not copy these simple lines?
As about intelligence, I suppose it is like beauty and disrespect, it's all
in the eye of the beholder...
Finally I remark you did not comment on auditing paper versions of STV.
Please do so on-list if you have good arguments, or off-list if you want
preventing this discussion to harm your legal case.
I hope it helped you prepare for your affidavits.
Stéphane Rouillon, ing. M.Sc.A. Ph.D.(in mathematics)
PS: Being a mathematician is not sufficient to always be logical, neither
right all the time.
----
Election-Methods mailing list - see http://electorama.com/em for list info
