SARA voting: easier-to-describe MAS
2016-10-26 14:29 GMT-04:00 Michael Ossipoff email9648742@gmail.com:
On Wed, Oct 26, 2016 at 1:07 PM, Jameson Quinn jameson.quinn@gmail.com
wrote:
My overall point is that, while VSE does not directly and precisely
measure the true "goodness" of the election outcome, if done carefully and
if assumptions are varied, it gives us the best and least-biased measure we
have of that goodness.
It isn't a measure of goodness at all, as my homeless-man/billionaire
example shows.
OK, do you prefer the technical term "estimator"? I was trying to avoid
jargon, but that's what I meant.
Estimator is a better word. I included D1 (sincere Score) among the methods
by which I compared the candidates in Forest's example.
We might disagree on what voting system to use in Utopia. ...a situation
we aren't likely to have to deal with anytime soon.
I completely agree with, applaud and appreciate CES's advocacy of Approval
and Score. I appreciate the work that they do.
I consider Approval & Score the best methods.
...except that there's at least sometimes a case for using MMPO, because of
its CD.
Reasons to not propose MMPO:
-
It has no use precedent.
-
It has some strongly-felt criticisms that would have to be answered,
and which would give an advantage to opponents, who have more media-access
than reform-advocates have.
I suggest proposing Approval, Score, &/or Bucklin.
I suggest that the best initial proposal, to the public &/or an initiative
proposal committee, is a several-methods proposal that describes and
advocates several methods, offering them all as proposals
That proposal should include Approval, Score, & Buickliln.
...and maybe MMPO, only if you're willing to include, in the proposal, the
objections to MMPO, and the answers to them.
But maybe, because MMPO has no use-precedent, and because pairwise-count
methods have no use precedent, and because of the need to answer those
objections to MMPO, it might be better to leave MMPO out of the proposal,
even at the start.
Of course the familiar objections to Approval should be answered in the
proposal too. CES's website does a good job of that.
Michael Ossipoff
....
On Wed, Oct 26, 2016 at 4:12 PM, Jameson Quinn jameson.quinn@gmail.com
wrote:
2016-10-26 14:29 GMT-04:00 Michael Ossipoff email9648742@gmail.com:
On Wed, Oct 26, 2016 at 1:07 PM, Jameson Quinn jameson.quinn@gmail.com
wrote:
My overall point is that, while VSE does not directly and precisely
measure the true "goodness" of the election outcome, if done carefully and
if assumptions are varied, it gives us the best and least-biased measure we
have of that goodness.
It isn't a measure of goodness at all, as my homeless-man/billionaire
example shows.
OK, do you prefer the technical term "estimator"? I was trying to avoid
jargon, but that's what I meant.
With the 300/100 dollar example, the gamble is the same as saying that two people would prefer to take a gamble so that one of them wins $301 and the other wins nothing than they both take $100 each guaranteed. Not all people would hold the same utility values, so that particular example wouldn't work for everyone, of course.
But in any case, a maxmin approach doesn't work because most people would take some sort of gamble like the above. Almost all pairs of people would take a coin flip gamble for $1000 rather than a guaranteed $1. Maxmin suggests that the $1 for each is better.
With the yacht example, it could be that even a whole yacht isn't sufficient to equal a dollar in the poor person's eyes. But there are real life examples a little bit like this anyway. Would you ban fairground rides? Each person who has a go on one has only a small increase in their overall lifetime utility for having a ride on one of these, and a very small proportion of people get injured or even killed on them. It would be very hard to justify their existence using any sort of maxmin approach, and most people seem to be taking more of an average utility approach when engaging in some sort of activity that comes with a small risk of injury.
From: Michael Ossipoff <email9648742@gmail.com>
To: Toby Pereira tdp201b@yahoo.co.uk; "election-methods@electorama.com" election-methods@electorama.com
Sent: Wednesday, 26 October 2016, 19:47
Subject: Re: [EM] SARA voting: easier-to-describe MAS
On Wed, Oct 26, 2016 at 10:28 AM, Toby Pereira tdp201b@yahoo.co.uk wrote:
Also, a given amount of money is worth more in utility to a poor person than a rich person, so Michael's analogy of taking a dollar from a homeless person and giving it to a billionaire doesn't work. I know he gave a previous example where the billionaire gets a yacht or something and it does get more debatable at that point.
Yes, I added that chance for that reason.
But I don't think the maximum minimum utility is necessarily even the best principle to use anyway. There are good arguments for maximising average utility.
I haven't heard one. The argument below doesn't successfully show that.
If I am given 100 dollars, then there is an amount of money that I would gamble that for on a coin flip, or where I'd call them equivalent. Let's say I decide that I'd gamble it for anything more than 300 dollars. That's the same as me saying that for me the difference in utility between 0 and 100 dollars is the same as the difference between 100 and 300 dollars. It is also equivalent to saying that if there are two people with the same utility ratings as me, it's as good to give one of them 300 dollars as it is to give each of them 100 dollars.
It results in the same total utility. To say that that means its just as good, circularly assumes what you seek to show.
You haven't shown that giving one person $300 and giving the other person nothing is as good as giving $100 to each. ...you've shown only that the total utility is the same.
You can't justify taking a dollar away from a homeless man who badly needs it, and giving another yacht to a billionaire.
I have the same intuitions about rich/poor people, but you need a logically consistent framework as well as intuitions.
I don't deny that logic can't apply to some ethical/moral questions. But logic has nothing to do with the basis of ethical and moral choices.
The basis of ethics & morality is intuitive & subjective, not logical.
Michael Ossipoff
(I have no way to delete the text below)
Not having that prize-dollar isn't great disutility. The situation most
likely is exactly as satisfactory for you with or without that dollar, and
that's why you're so willing to gamble it for $1000.
Taking the dollar from the homeless man is quite different. Then it's a
matter of going from bad to considerably more seriously bad. The loss of
his dollar is a lot more than your spending a dollar for a lottery-ticket,
when you don't really miss the loss of a dollar.
Spending your rent money on lottery tickets would be a better example.
The willingness of people to risk their lives on a fairground-ride says
something about how they rate the utilities of the disaster & the fun.
That's all it says. To each their own.
And the fact that the homeless man chose to be on the wrong end of the
economic system? ....Oh wait, he didn't choose it. Nor did he decide to
enter a lottery with a small probability of being a billionaire and a much
larger probability of being homeless. And when a lottery for high
probability of disaster in return for a tiny probability of a huge reward
is available, people very rarely take it.
People riding a fairground ride do enter a lottery for a tiny risk of
losing everything, for a small payoff. But people don't show willingness to
enter lotteries with a high probability of something really bad, in return
for a tiny probability of something extravagantly desirable. (Like spending
your rent money on lottery-tickets).
In the former case, the probability of disaster is small. In the latter
case, it's large.
It's a more than a little callous to say that hurting a homeless person to
help a billionaire is ok. But that's what BR-minimization or VSE
maximization, presented as a measure of social-rightness, says.
Lotteries are used to judge utilities. They don't say or imply what you're
trying to use them to say.
Michael Ossipoff
Michael Ossipoff
On Thu, Oct 27, 2016 at 5:29 AM, Toby Pereira tdp201b@yahoo.co.uk wrote:
With the 300/100 dollar example, the gamble is the same as saying that two
people would prefer to take a gamble so that one of them wins $301 and the
other wins nothing than they both take $100 each guaranteed. Not all people
would hold the same utility values, so that particular example wouldn't
work for everyone, of course.
But in any case, a maxmin approach doesn't work because most people would
take some sort of gamble like the above. Almost all pairs of people would
take a coin flip gamble for $1000 rather than a guaranteed $1. Maxmin
suggests that the $1 for each is better.
With the yacht example, it could be that even a whole yacht isn't
sufficient to equal a dollar in the poor person's eyes. But there are real
life examples a little bit like this anyway. Would you ban fairground
rides? Each person who has a go on one has only a small increase in their
overall lifetime utility for having a ride on one of these, and a very
small proportion of people get injured or even killed on them. It would be
very hard to justify their existence using any sort of maxmin approach, and
most people seem to be taking more of an average utility approach when
engaging in some sort of activity that comes with a small risk of injury.
From: Michael Ossipoff email9648742@gmail.com
To: Toby Pereira tdp201b@yahoo.co.uk; "election-methods@electorama.com"
election-methods@electorama.com
Sent: Wednesday, 26 October 2016, 19:47
Subject: Re: [EM] SARA voting: easier-to-describe MAS
On Wed, Oct 26, 2016 at 10:28 AM, Toby Pereira tdp201b@yahoo.co.uk
wrote:
Also, a given amount of money is worth more in utility to a poor person
than a rich person, so Michael's analogy of taking a dollar from a homeless
person and giving it to a billionaire doesn't work. I know he gave a
previous example where the billionaire gets a yacht or something and it
does get more debatable at that point.
Yes, I added that chance for that reason.
But I don't think the maximum minimum utility is necessarily even the best
principle to use anyway. There are good arguments for maximising average
utility.
I haven't heard one. The argument below doesn't successfully show that.
If I am given 100 dollars, then there is an amount of money that I would
gamble that for on a coin flip, or where I'd call them equivalent. Let's
say I decide that I'd gamble it for anything more than 300 dollars. That's
the same as me saying that for me the difference in utility between 0 and
100 dollars is the same as the difference between 100 and 300 dollars. It
is also equivalent to saying that if there are two people with the same
utility ratings as me, it's as good to give one of them 300 dollars as it
is to give each of them 100 dollars.
It results in the same total utility. To say that that means its just as
good, circularly assumes what you seek to show.
You haven't shown that giving one person $300 and giving the other person
nothing is as good as giving $100 to each. ...you've shown only that the
total utility is the same.
You can't justify taking a dollar away from a homeless man who badly needs
it, and giving another yacht to a billionaire.
I have the same intuitions about rich/poor people, but you need a
logically consistent framework as well as intuitions.
I don't deny that logic can't apply to some ethical/moral questions. But
logic has nothing to do with the basis of ethical and moral choices.
The basis of ethics & morality is intuitive & subjective, not logical.
Michael Ossipoff
(I have no way to delete the text below)