Re: [EM] help with advocacy

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [EM] help with advocacy

Forest Simmons
Nobody sharper than Steve Eppley in countering IRV propaganda ... for example see his response in the forwarded link below .. great person to have in your corner.

---------- Forwarded message ----------
Date: Thursday, April 15, 2021
Subject: Single-winner election data from the OpenSTV database
To: Forest Simmons <[hidden email]>




----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

[EM] Fixing Narkive (Re: help with advocacy)

Rob Lanphier-3
Hi Forest,

You linked to Narkive in a very recent message, referring to Steve
Eppley's rhetoric from long ago:

On Thu, Apr 15, 2021 at 4:19 PM Forest Simmons <[hidden email]> wrote:
> https://election-methods.electorama.narkive.com/8HqX0Gjr/single-winner-election-data-from-the-openstv-database

Your email led me to visit this URL:
https://election-methods.electorama.narkive.com/

It appears as though Narkive has stopped archiving EM-list since 2018.
I did a little bit of checking the website for more information:

https://narkive.com/legalese

I would love to figure out how to restore that, but I'm not sure how.
Can you (or anyone) find contact information for the maintainer(s) of
Narkive?  I have some ideas, and I'm planning to follow through on
them, but it could be that I'm missing out on an obvious FAQ that
tells me where to go for technical questions, but my hunch is that
it's deliberately vague, since the website is a small-time operation
(much like this mailing list)

Rob
----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Fixing Narkive (Re: help with advocacy)

Forest Simmons
 Rob,

I was trying to find Eppley's July 2007 article in the Pasadena weekly by Googling Eppley VPR, and that Narkive result came up.

Besides BTR-IRV some other minimal tweaks to greatly improve IRV (though not as viable as your idea) are Voting Published Rankings (VPR) and the Candidate Withdrawal option, ... both thoroughly expounded by Eppley.

Also we could mimic the movement of states pledging their Electoral College votes to the plurality winner ... by pledges to support the Pairwise Winner in the context of ranked choice ballots.

That doesn't confer Smith compliance, but the following version would do so for a small cost ... very small compared to the total cost of IRV:

If the base method is loser elimination like IRV, then as soon as there is a Pairwise Winner among the remaining candidates, the pledge takes effect. If there are not enough pledge votes to render further rounds moot, then the rounds continue until pledged votes are sufficient to give the (current) pairwise winner (among the remaining candidates) an outright majority of the votes still in play.

Ideally the technology should accommodate voter pledges as well as candidate pledges.

Way too much for the current Burlington situation, but a way to respect voter wishes when the base method is incapable of delivering as advertised: "... your second choice will be there to take effect should your first choice be eliminated..."

For that promise to be fulfilled at every stage for some voter, the candidates would have to be eliminated in the same order as the ranking order of that voter. If that promise is kept to most voters, then most ballots would have the same first choice, so nobody's second choice would matter. So IRV propaganda makes a promise that cannot be kept  to more than half of the voters except vacuously ... i.e. when the IRV winner is identical to the Plurality winner, i.e. when IRV does not improve on Plurality.

So any advantage IRV has over Plurality depends on a disingenuous promise ... it cannot give different results from Plurality without violating that promise on at least half of the ballots.

It's not that IRV fails to keep that promise once in a while ... pervasive failure is inevitable!

 It sounds to me like grounds for a law suit!

My Best,

Forest

On Thursday, April 15, 2021, Rob Lanphier <[hidden email]> wrote:
Hi Forest,

You linked to Narkive in a very recent message, referring to Steve
Eppley's rhetoric from long ago:

On Thu, Apr 15, 2021 at 4:19 PM Forest Simmons <[hidden email]> wrote:
> https://election-methods.electorama.narkive.com/8HqX0Gjr/single-winner-election-data-from-the-openstv-database

Your email led me to visit this URL:
https://election-methods.electorama.narkive.com/

It appears as though Narkive has stopped archiving EM-list since 2018.
I did a little bit of checking the website for more information:

https://narkive.com/legalese

I would love to figure out how to restore that, but I'm not sure how.
Can you (or anyone) find contact information for the maintainer(s) of
Narkive?  I have some ideas, and I'm planning to follow through on
them, but it could be that I'm missing out on an obvious FAQ that
tells me where to go for technical questions, but my hunch is that
it's deliberately vague, since the website is a small-time operation
(much like this mailing list)

Rob

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Fixing Narkive (Re: help with advocacy)

Forest Simmons
I agree with VoteFair Richard that voter friendly naming is important.

Round Robin and Head-to-Head are familiar concepts ... good for introducing Pairwise.

Beats All is another good descriptor.

If we ever need the concept of Smith, ... a Smith candidate has Beat Paths to all of the other candidates.

Shifting gears ...

IRV false claims should be challenged in court, and court injunctions issued against false and misleading verbiage...

Three of the most common misleading claims ...

1. IRV always finds a majority winner. In reality, this is impossible when majorities prefer A over B and B over C and C over A. It is true that the winner is preferred by a majority of the unexausted ballots over the last candidate to be eliminated ... but nothing more can be claimed truthfully.

2. The optimum strategy is to rank the candidates sincerely. ... This claim assumes zero information about preferences of other voters.

3. If you rank Y immediately after X, then in the event X is eliminated, your vote will be transferred to Y.  That holds only if X is eliminated before Y is eliminated. 

In my previous message I explained why (unless most ballots are exhaustied before the final round) this promise cannot be kept on more than half of the ballos if there is no majority first place candidate ... the only case that justifies ranked choice style ballots in the first place. So this claim is not just a tiny white lie that hardly ever occurs in practice ... its pervasive violation is inevitable (except perhaps when rankings are woefully incomplete).

Try to find even one reasonable scenario with complete rankings where this claim holds true on most ballots.



On Thursday, April 15, 2021, Forest Simmons <[hidden email]> wrote:
 Rob,

I was trying to find Eppley's July 2007 article in the Pasadena weekly by Googling Eppley VPR, and that Narkive result came up.

Besides BTR-IRV some other minimal tweaks to greatly improve IRV (though not as viable as your idea) are Voting Published Rankings (VPR) and the Candidate Withdrawal option, ... both thoroughly expounded by Eppley.

Also we could mimic the movement of states pledging their Electoral College votes to the plurality winner ... by pledges to support the Pairwise Winner in the context of ranked choice ballots.

That doesn't confer Smith compliance, but the following version would do so for a small cost ... very small compared to the total cost of IRV:

If the base method is loser elimination like IRV, then as soon as there is a Pairwise Winner among the remaining candidates, the pledge takes effect. If there are not enough pledge votes to render further rounds moot, then the rounds continue until pledged votes are sufficient to give the (current) pairwise winner (among the remaining candidates) an outright majority of the votes still in play.

Ideally the technology should accommodate voter pledges as well as candidate pledges.

Way too much for the current Burlington situation, but a way to respect voter wishes when the base method is incapable of delivering as advertised: "... your second choice will be there to take effect should your first choice be eliminated..."

For that promise to be fulfilled at every stage for some voter, the candidates would have to be eliminated in the same order as the ranking order of that voter. If that promise is kept to most voters, then most ballots would have the same first choice, so nobody's second choice would matter. So IRV propaganda makes a promise that cannot be kept  to more than half of the voters except vacuously ... i.e. when the IRV winner is identical to the Plurality winner, i.e. when IRV does not improve on Plurality.

So any advantage IRV has over Plurality depends on a disingenuous promise ... it cannot give different results from Plurality without violating that promise on at least half of the ballots.

It's not that IRV fails to keep that promise once in a while ... pervasive failure is inevitable!

 It sounds to me like grounds for a law suit!

My Best,

Forest

On Thursday, April 15, 2021, Rob Lanphier <[hidden email]> wrote:
Hi Forest,

You linked to Narkive in a very recent message, referring to Steve
Eppley's rhetoric from long ago:

On Thu, Apr 15, 2021 at 4:19 PM Forest Simmons <[hidden email]> wrote:
> https://election-methods.electorama.narkive.com/8HqX0Gjr/single-winner-election-data-from-the-openstv-database

Your email led me to visit this URL:
https://election-methods.electorama.narkive.com/

It appears as though Narkive has stopped archiving EM-list since 2018.
I did a little bit of checking the website for more information:

https://narkive.com/legalese

I would love to figure out how to restore that, but I'm not sure how.
Can you (or anyone) find contact information for the maintainer(s) of
Narkive?  I have some ideas, and I'm planning to follow through on
them, but it could be that I'm missing out on an obvious FAQ that
tells me where to go for technical questions, but my hunch is that
it's deliberately vague, since the website is a small-time operation
(much like this mailing list)

Rob

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] help with advocacy

Forest Simmons
In reply to this post by Forest Simmons
A couple of points:

When the Condorcet Candidate gets the fewest top rank votes it is almost surely because of the "center squeeze effect" which is easy to illustrate in one dimension, and is easy to see in Yee Diagrams in the case of a two dimensional issue space.

The voters who prefer the CW to the sincere IRV winner have a strong incentive to rescue the CW from the center squeeze by insincerely ranking her in top place. Of course the same gambit would work equally well with Plurality ballots ... but wasn't adopting IRV supoosed to save us from these Plurality manipulations?

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Help with Advocacy

Forest Simmons
In reply to this post by Forest Simmons
All Condorcet compliant methods yield identical YEE diagrams, because the candidate closest to the center of symmetry of the (multivariate Gaussian) voter distrbution is always a Pairwise Winner.

So BTR-IRV, Smith//IRV, Kemeny Young, Ranked Pairs, Dodgson, Copeland, MinMax, etc. all have identical ideal diagrams with convex win regions... unlike plain IRV whose win regions may be fragmented, and even when connected, non-convex and riddled with holes.

So a dramatic visual demonstration of IRV's chaotic deviation from Condorcet when candidates are randomly distributed among regularly distributed voters is to show their YEE pictures side by side.

So claims that IRV will usually pick the Condorcet Candidate are actually not true according to these YEE diagrams. How do we explain reported high levels of Condorcet Compliance in practice? 

Unlike the absolutely sincere YEE voter V who ranks candidate X over Y when X is closer to V than Y in issue space, the flesh and blood voter pays attention to horse race polls and bends accordingly ... making pragmatic adjustments to their conscience.

Continuing Rob's airplane analogy,  without the body English and arm flapping of the passengers, the 299 out of 300 success rate would be unlikely.

The link below leads to some examples on Warren Smith's website. Note the win regions for IRV are rarely connected, let alone convex. So as candidates move along straight lines from one point of their win regions to another, their fortunes change erratically ... in total contrast to the Condorcet method pictures ... and most other well known reputable methods.

 it is important to know that Yee did not design his "electoscope" with the intent to make IRV look worse then any other method ...  it just turned out that way!

---------- Forwarded message ----------
Date: Saturday, April 17, 2021
Subject: RangeVoting.org - Yee Pictures
To: Forest Simmons <[hidden email]>


https://rangevoting.org/IEVS/Pictures.html




Sent from my MetroPCS 4G LTE Android Device


----
Election-Methods mailing list - see https://electorama.com/em for list info