Drama over Rob Richie's comment

22 views
Skip to first unread message

Leon Smith

unread,
May 20, 2013, 7:01:44 AM5/20/13
to electionsciencefoundation
Ok,  I haven't really kept up with this list in a while,  but I did do some reading yesterday and wanted to interject some points:

1.   Relax.  I doubt the judges take into account the comments on a video.   Perhaps they do,  but I would be slightly surprised.

2.   I really don't know how good our chances of winning.   Yes,  I've looked at probably 2 dozen promising looking entries and I haven't seen anything else that's both presented attractively and has a concrete suggestion that isn't totally pie-in-the-sky.    But this is politics,  and politics isn't about being correct,  it's about being perceived as correct by other people,  in this case the judges. Perception is (political) reality.

In any case,  I don't think we have any realistic chance at winning the public choice award.  Sorry.   Other people's perception just isn't there yet.   The American public is currently more interested in unworkable,  "common-sense" aphorisms to the problems they perceive,  whether or not that problem is even really a problem or just a symptom of a deeper dysfunction.

3.   It seems that Rob Richie was the first to bring up the topic of IRV in the discussions,  though perhaps comments by Clay and by Abd kind of hinted at it.   Just wanted to point this out for later.

4.   Though I've never met Rob Richie or had a conversation with him, even over the internet,  I do think I understand his position pretty well,  in a way that few people here seem to.   And that revolves around the mechanics of a political campaign, not debating the finer points of voting theory.   

I do thank FairVote for providing me my first introduction to the issues surrounding voting systems.  From late 1999 to mid 2000,  I read every one of their web pages,  twice.   (at least!)    And FairVote made it clear that they were a political campaign devoted to the adoption of PR and IRV,  not a debate of trying to figure out which system was best.    Now I became disenchanted with their arguments after a few years,  and drifted towards Condorcet (without any particularly strong like or dislike for Approval) for a while.   Then I saw Ka-Ping Yee's visualizations which really made me reconsider approval more carefully,  and I saw Warren's claims about Arrow's Theorem which were intriguing at first and quite a revelation later.

But I digress,  the point is that people tagging along with a political campaign derailing the conversation the campaign is putting forward and non-constructively attacking the campaign is a real problem,  and from what I gather from GtV, this is Rob Richie's real complaint about this group's past behavior.    It's not at all surprising to me that he has finally showed us a modicum of respect with the production of this video,  as it was a constructive effort of trying to get our message out on our own instead of attempting to ride FairVote's coattails,  so to speak. 

And I believe that comment was made in good faith and more-or-less accurate.  Of course I don't think the Later-no-harm criterion is terribly relevant,  and certainly not worth the high cost of IRV,  but that's the problem of axiomatic approaches to judging voting systems.   And the comments regarding the CUSG elections I tend to interpret as an honest mistake;  I think Abd is correct in that we need to take some future opportunity to calmly explain the situation and see how Richie reacts.  

Now,  one could perceive this as a (small) example of Richie trying to ride our coattails and derail the conversation.   And I certainly tend to see it that way.   My suggestion is to brush this particular incident off.   Consider it a little lesson in the way Richie has (probably) viewed us for years.

And be careful not to overplay your hand when responding.   I'm glad that Clay's comment is on the second page.   But live and learn;   the only way you won't make mistakes is by not trying.

Best,
Leon

Abd ul-Rahman Lomax

unread,
May 21, 2013, 12:39:01 PM5/21/13
to electionsciencefoundation
At 06:01 AM 5/20/2013, Leon Smith wrote:
>Ok, I haven't really kept up with this list in a while, but I did
>do some reading yesterday and wanted to interject some points:
>
>1. Relax. I doubt the judges take into account the comments on a
>video. Perhaps they do, but I would be slightly surprised.

Right. However, if the comments are from the submitter, or from the
organization that submitted the video, some of the comments might
give them pause. Hence my concern, particularly about *our* comments,
not those of others, which would be unlikely to have any effect
unless they show some kind of fraud or deception on our part, that
the judges could easily check. Nothing was there like that.

>2. I really don't know how good our chances of
>winning. Yes, I've looked at probably 2 dozen promising looking
>entries and I haven't seen anything else that's both presented
>attractively and has a concrete suggestion that isn't totally
>pie-in-the-sky. But this is politics, and politics isn't about
>being correct, it's about being perceived as correct by other
>people, in this case the judges. Perception is (political) reality.

That's correct. I think we have a good chance. By no means is it a slam-dunk.

>In any case, I don't think we have any realistic chance at winning
>the public choice award. Sorry. Other people's perception just
>isn't there yet. The American public is currently more interested
>in unworkable, "common-sense" aphorisms to the problems they
>perceive, whether or not that problem is even really a problem or
>just a symptom of a deeper dysfunction.

The most popular video was one with zero practical suggestions, just
feel-good "punish the bastards" jokes.

>3. It seems that Rob Richie was the first to bring up the topic of
>IRV in the discussions, though perhaps comments by Clay and by Abd
>kind of hinted at it. Just wanted to point this out for later.

No. The sequence:

Aaron and Clay had commented, both CES board members. Jameson had
commented and I had commented, and others had commented. There had
been no mention of IRV. The closest was my comment, "The principle is
really of voting independently on all candidates, and that principle
can be extended to more sophisticated rating or ranking on ballots."
Which someone might think was a reference to IRV. It wasn't. It's
about Bucklin and Range.

Then Ryan C (I have no idea who he is) wrote this:

><http://challengepost.com/users/ryancwiklinski>Ryan C 14 days ago
>
>Great video and great concept. It's interesting to see it compared
>to Ranked Choice Voting:
><http://lookingatdemocracy.org/submissions/15145-who-s-your-favorite-president-how-ranked-choice-voting-works>http://lookingatdemocracy.org/submissions/15145-who-s-your-favorite-president-how-ranked-choice-voting-works

Aaron replied promptly:

><http://challengepost.com/users/ElectionScience>Aaron H 14 days ago
>
>Thanks, Ryan. As you can imagine, our focused support for Approval
>Voting over other methods is quite deliberate. It's hard to imagine
>a simpler improvement. Plus, you can always vote your favorite in
>Approval Voting. This is a feature that is simply unavailable to
>Ranked Choice Voting and other ranked methods. It's remarkable how
>much gain there is from Approval Voting just by letting voters pick
>as many candidates as they want.

Just because someone else invites comparison with IRV, doesn't mean
we should do it, not there. We can invite people to "the
conversation," and it's a complex one. Our video does not make that
comparison, was not intended to make that comparison, and they really
are apples and oranges. Aaron is here speaking officially for CES,
using "our" to reflect CES reasoning. Clay commented on the RCV video
page, actually signing it CES. I objected to that, and it seems Clay
did manage to edit that signature out.

I agreew with Aaron as to the substance, but not as to the context.
It, together with Clay's comment on the RCV video, invited what
showed up the next day, Richie's comment, followed by Aaron's
response to him. Aaron, if he'd realized the issue, could simply have
thanked Ryan C, and should Richie still have shown up, should have
thanked him for his recognition of the value of approval over
plurality, or some such.

Instead, we got the attack on Richie on the next page of comments,
with references that will largely be unintelligible to those who are
not familiar with the issues. Richie himself was out on a limb, and
it seems he was speaking more to us than to the judges, as Warren
seems to think.


>4. Though I've never met Rob Richie or had a conversation with
>him, even over the internet, I do think I understand his position
>pretty well, in a way that few people here seem to. And that
>revolves around the mechanics of a political campaign, not debating
>the finer points of voting theory.

I've had extensive discussions with Richie at various times, though
never in person or on the phone. I agree with Leon's analysis. Richie
is a fairly standard "type," a political activist.

>I do thank FairVote for providing me my first introduction to the
>issues surrounding voting systems. From late 1999 to mid 2000, I
>read every one of their web pages, twice. (at least!) And
>FairVote made it clear that they were a political campaign devoted
>to the adoption of PR and IRV, not a debate of trying to figure out
>which system was best. Now I became disenchanted with their
>arguments after a few years, and drifted towards Condorcet (without
>any particularly strong like or dislike for Approval) for a
>while. Then I saw Ka-Ping Yee's visualizations which really made
>me reconsider approval more carefully, and I saw Warren's claims
>about Arrow's Theorem which were intriguing at first and quite a
>revelation later.

Arrow himself was quite aware of the limitations of the Theorem.
Lot's of people are not; FairVote, in particular, uses the Theorem to
defend IRV, along the lines of "According to Arrow's Theorem, no
election method is perfect," so, it's implied, why not use IRV, and
it's almost the only method that satisfies the Later-No-Harm
Criterion (which certainly is not one of Arrow's study criteria, and
which is not widely accepted as desirable, which can be argued is
itself *harmful*, and which has been used to attack, say, Bucklin,
when Later-No-Harm, as such, was very likely not the explanation for
supposedly low usage of additional ranking in later Bucklin elections.

(Those were low-knowledge elections, i.e., party primaries, where
people will knee-jerk vote-for-one; there were still about 11% or so
additional approvals, but they nornally did not shift outcomes.
FairVote's analysis was *quite* shallow, and only applied to that
narrow slice of Bucklin elections. One of the biggest errors of
FairVote was, in fact, getting involved with "debate ... trying to
figure out which system is best," without taking advantage of known
election science. In scientific journals, the problems with the Ware
system (IRV!) were well known before 1900.

It's very likely, from the analyses I've seen, that IRV was adopted
in Australia as a spoiler-killer. Regardless, the effect was not to
encourage minor parties. IRV does work to protect major parties from
spoilers, as compared with Plurality. Unfortunately, as a minor party
gains traction, it breaks down with Center Squeeze.

>But I digress, the point is that people tagging along with a
>political campaign derailing the conversation the campaign is
>putting forward and non-constructively attacking the campaign is a
>real problem, and from what I gather from GtV, this is Rob Richie's
>real complaint about this group's past behavior. It's not at all
>surprising to me that he has finally showed us a modicum of respect
>with the production of this video, as it was a constructive effort
>of trying to get our message out on our own instead of attempting to
>ride FairVote's coattails, so to speak.

Well, without buying that construction of CES or preceding activity,
yes, this could be how he sees us. And so his comments, in line with
some other comments here by a FairVote counsel, can be seen as
raising a possibility of cooperation, and, for that to work, we are
going to need to restrain some of our more enthusiastic members. And
board members, especially.

>And I believe that comment was made in good faith and more-or-less accurate.

Yeah, that matches my analysis. The possible inaccuracies were in how
the presented facts could be used, but he did not actually use them
that way. He pointed to issues. That's entering a conversation. Isn't
that what we were inviting with the video?

> Of course I don't think the Later-no-harm criterion is terribly
> relevant, and certainly not worth the high cost of IRV, but
> that's the problem of axiomatic approaches to judging voting
> systems. And the comments regarding the CUSG elections I tend to
> interpret as an honest mistake; I think Abd is correct in that we
> need to take some future opportunity to calmly explain the
> situation and see how Richie reacts.

Yes. However, this is an obscure comment page. It may be much more
important for Richie to address the official FairVote position paper
on Approval Voting, which is pretty bad.

Given the number of jurisdictions holding elections where an improved
voting system could be of service, it's crazy to regard IRV and
Approval as "competitors" in a zero-sum game. It would be, in fact,
important to have implementations of various systems, to study and
see how the systems work. My position as a CES supporter is that we
want deciders to have fair and accurate information about what kinds
of performance can reasonably be expected, about risks, about the
history of voting systems, about costs, and all that.

>Now, one could perceive this as a (small) example of Richie trying
>to ride our coattails and derail the conversation. And I certainly
>tend to see it that way. My suggestion is to brush this particular
>incident off. Consider it a little lesson in the way Richie has
>(probably) viewed us for years.

I don't agree with the anaysis, it's simply not consistent. Richie
showed up to defend IRV, but along the way, he also commented
positively on our work or on what he appeared to perceive as a
refreshing attitude. However, that some of us interpreted Richie
differently says a great deal about our own knee-jerk habits.

>And be careful not to overplay your hand when responding. I'm glad
>that Clay's comment is on the second page. But live and
>learn; the only way you won't make mistakes is by not trying.

Yes. That's a Basic principle. The big problems with mistakes arise
when one is incapable of seeing them, out of a need to be right and
an inability to see things from other points of view.

Bruce Gilson

unread,
May 21, 2013, 3:27:51 PM5/21/13
to electionscience Foundation



On Tue, May 21, 2013 at 12:39 PM, Abd ul-Rahman Lomax <a...@lomaxdesign.com> wrote:

[...]


Given the number of jurisdictions holding elections where an improved voting system could be of service, it's crazy to regard IRV and Approval as "competitors" in a zero-sum game. It would be, in fact, important to have implementations of various systems, to study and see how the systems work. My position as a CES supporter is that we want deciders to have fair and accurate information about what kinds of performance can reasonably be expected, about risks, about the history of voting systems, about costs, and all that.

Although I do not much like Abd's prolixity in his posts, including the one I'm replying to now, and I usually disagree with what he says, in this he has expressed a position which is identical to my own.

We really need to try out all the proposed improved voting systems to see how they work in real elections. Warren Smith's Bayesian Regret simulations help point at likely value judgments, but they cannot substitute for real-world trials.

Stephen Unger

unread,
May 21, 2013, 4:15:41 PM5/21/13
to electionscience Foundation
I appreciate the spirit of Bruce's approach, which, as I understand
it, is that we should welcome having different jurisdictions
experiment with various alternative voting schemes.

But I think that position ignores the fact that there are serious
costs when particularly bad schemes are deployed. And I regard IRV as
about as bad as it gets. There are at least 2 downsides.

One is that it is obviously detrimental to the people of any area if
they have to use a seriously faulty system.

The other, with worse consequences, is that, if people are
persuaded to abandon the system they, and almost everyone else, has
been using for their whole lives, and use another method of voting,
the consequences if the new system works out very badly would be very
serious. It would certainly be very hard to get those people to then
try out a third system. And those in other jurisdictions asked to adopt
a new system (say approval) would be much more likely to reject the
idea.

I believe we ought to do the best we can to educate people about the
faults of all systems (including those we advocate). IRV has to get a
lot of attention in this respect because it is so bad, is the main
alternative offered to plurality, and because its promoters are so
skillful.

Steve
............

Stephen H. Unger
Professor Emeritus
Computer Science and Electrical Engineering
Columbia University
............
> --
> You received this message because you are subscribed to the Google Groups
> "The Center for Election Science" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to electionscien...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
> �
> �
>
>

Leon Smith

unread,
May 21, 2013, 4:42:04 PM5/21/13
to electionsciencefoundation
Well, the Political and Electoral Reform Working Group of Occupy Wall Street has been doing some interesting work in this area,  conducting polls and exit polls using a variety of different methods,  especially Range, Approval, and IRV.   But of course honest Range ballots can also be reasonably interpreted as Majority Judgement ballots,  and honest IRV ballots can be reasonably interpreted as any of the ranked voting systems as well,  like Bucklin or Condorcet.   And given that polls are more likely to be honest,  and that unfamiliar voting systems are more likely to be honest (or at most using something like naive exaggeration strategies),  this doesn't seem like an unreasonable assumption to make.

I'd say that downsides of this methodology compared to computer simulations is the expense, that you can't study as many different situations, and studying strategic behavior is difficult because the stakes of polling is lower and because the systems are novel and unfamiliar.   But of course it's real data instead of studying voting systems relative to some hypothetical model of voter behavior,  and studying strategic behavior is difficult in either case.

Warren D Smith

unread,
May 21, 2013, 4:56:22 PM5/21/13
to electio...@googlegroups.com
http://RangeVoting.org/RangePolls.html

(and pages linked from there) collects
much about real-world elections under different voting systems.
(The more recent elections usually have more data about more systems.)


--
Warren D. Smith
http://RangeVoting.org <-- add your endorsement (by clicking
"endorse" as 1st step)

Abd ul-Rahman Lomax

unread,
May 21, 2013, 8:50:24 PM5/21/13
to electionsciencefoundation
At 03:42 PM 5/21/2013, Leon Smith wrote:
>[...]But of course honest Range ballots can also be reasonably
>interpreted as Majority Judgement ballots, and honest IRV ballots
>can be reasonably interpreted as any of the ranked voting systems as
>well, like Bucklin or Condorcet.

Bucklin is often assumed to be a ranked system. In fact, it's a
rating system, it uses a Range ballot.

Original Bucklin did restrict the top two ranks as vote-for-one, but
the third rank allowed multiple approvals, and skipping the second
rank was allowed. A more generalized Bucklin would simply use a Range
ballot. A simple version that was otherwise identical to original
Bucklin would allow multiple voting in all ranks.

An IRV ballot cannot be converted to a Bucklin ballot without
assuming constant preference strength. In fact, IRV ballots only show
ranking, not preference strength. Bucklin ballots definitely consider
preference strength, which is indicated in two ways:

Low preference strength, equal ranking.
Medium preference strength, ordinary ranking.
Higher preference strength, skipping rank(s).
Maximum preference strength, Bullet vote.

So, yes, a Range ballot can be interpreted as a Bucklin ballot,
easily. Majority Judgment is Bucklin, really.

A range ballot can *also* be interpreted as a ranked ballot.

The reverse is not true, or, more accurately, information is lost by
interpreting Range votes as ranks.

I think that simulations have assumed that Bucklin was a ranked
ballot and did not recognize that skipping ranks was allowed and
actually practiced. As Bucklin is "instant runoff Approval," it
really is median Range.

Reply all
Reply to author
Forward
0 new messages