[EM] Subject: Re: APR Steve's most recent response to Richard Fobes

Richard Fobes ElectionMethods at VoteFair.org
Sat Nov 8 18:11:14 PST 2014


On 11/6/2014 6:12 AM, steve bosworth wrote:
 > ...
 >> >>2S:Yes, both types of primaries help the <the final results> to be
 > better. However, APR’s primary seems to have the additional advantage of
 > allowing citizen to encourage more attractive candidates to be available
 > to be ranked later in the general election.

Your method is not the only one that would improve the quality of 
leaders entering and winning primary elections.

If U.S. primary elections used better ballots, say approval ballots, 
then wise problem-solving leaders would enter the primary elections. 
Now, they know they can't win, so they don't enter.

 >> ...
 >> >>2S:I understand this criticism of IRV and that using the
 > <Kemeny-Young method> as a corrective might even be practically possible
 > in single winner elections.However, am I correct in believing the this
 > method would be entire impractical for the large multi-winner 
elections ...

The calculations for proportional multi-winner elections are not 
excessively time-consuming, even when the Condorcet-Kemeny method is 
used for calculating the winner of each successive seat.  The VoteFair 
ranking software does those calculations as part of calculating VoteFair 
representation ranking results, and even in unusual worst-case scenarios 
the calculation times are very practical for election purposes.

As for a non-election scenario, if proportional calculations had to be 
done for every search-engine results page, then the calculation times 
would be an issue.

The people who claim that any method is slow are likely to be assuming 
that the only way to calculate the method is the same way the method is 
described.  In the case of Condorcet-Kemeny calculations an algorithm 
that has some similarities to the "insertion sort" algorithm is much, 
much faster than the "brute-force" algorithm that calculates every 
sequence score (as if all the scores had to be known in order to find 
the highest score).

 > ...
 > for which APR’s combined use of modifies STV and ‘Asset Voting’ would
 > seems to be the best practical method – best because it allows each
 > citizen to maximizes the chances that his or her vote will be added to
 > the ‘weighted vote’ in the legislature of the rep he or she trusts
 > most.  APR seems to be practical both because it is technically possible
 > and is relatively easy for ordinary citizens to understand.

I agree that APR -- Associational Proportional Representation -- is easy 
to understand compared to most PR (proportional representation) methods, 
and that is a big advantage.

I agree that being technically practical is very important.  Yet all of 
the methods seriously promoted here -- except instant-runoff voting used 
in something big like a close U.S. Presidential race (because every 
ballot must be gathered at a single counting location) -- is technically 
practical.

 > If my belief is correct, this is in sharp contrast to the relative
 > difficulty of understanding Condorcet-Kemeny (even if used in a
 > single-winner election) and the practical impossibility of using it for
 > a large multi-winner election.

Actually, the Condorcet-Kemeny method is much easier to understand than 
the Condorcet-Schulze method.

What's difficult to understand -- for most methods except instant-runoff 
voting and approval voting -- is how the calculations are actually done. 
  But voters do not need to understand how the calculations are actually 
done.  They only need to understand how the calculations _could_ be done 
(which in the case of the Condorcet-Kemeny method is the brute-force 
method).

As previously explained, the Condorcet-Kemeny method is practical, even 
within a multi-winner proportional method.

...

 >> >>2S:I don’t yet understand the answer you have given. Yes, the
 > <consultants/thugs> might pay a commission only if their candidate
 > wins. However, when their candidate loses, my point was that they would
 > find it more difficult, if not impossible, to discover whether this was
 > because their bribed voters failed to vote as promised or the other
 > voters who were not bribed but who were assumed to be supporters of
 > their candidate decided not to vote for that candidate.

If APR were used for electing members of parliament, it has a much 
higher benefit for each dollar of influence, compared to most other 
methods (with the notable exceptions of plurality and Borda count).

To understand this "influence-per-dollar" concept better, imagine that 
members of parliament (or Congress) are elected the same way as now, but 
then after the election each voter could do what your APR method allows, 
which is to let each voter decide which MP (member of parliament) "gets 
their vote" for the purpose of weighted voting on proposed laws.

In this arrangement, news-media owners would promote celebrity MPs, and 
hide the scandals of those MPs.  In contrast, they would highlight the 
scandals of MPs who have "opposition" opinions.  Under that media 
influence huge numbers of voters would give their votes to those 
celebrity MPs.  Of course those promoted MPs would be the ones who would 
vote the way the news-media owners want.  And remember that big 
corporations own TV networks specifically for the purpose of keeping the 
media silent about the scandals in those businesses.

Here, as a related story, I'll share that, years ago, when celebrity 
Martha Stewart was falling from grace, a doctor said that one of his 
patients requested that the doctor prescribe "the drug that Martha 
Stewart uses."  In case you don't know the story, she did not use the 
drug, rather (as I recall) she dumped stock based on inside information 
about the drug not passing the latest government approval process, 
before this fact was publicly announced.  The point is that people have 
oversimplified associations between celebrities and whatever the 
celebrity is linked to (such as drugs, shoe brand, etc.).

It should be easy to imagine news-media's influence on voters giving 
their support to celebrities, without realizing that the celebrity 
politician isn't doing what the voters really want.  Instead those 
celebrities would be picked for their puppet-like behavior.

...
 >> >>2S:As I see it, this would not at all be clear.If <lots of voters in
 > lots of districts (associations) vote for a specific … politician> that
 > might simply mean that he is trust by many people nationwide, i.e. that
 > he is their most trusted rep.

When Ronald Reagan was running for president, some voters were asked for 
their opinions about political issues, and even when they had very 
different political positions compared to Reagan, many of those people 
enthusiastically said they were going to vote for Reagan -- as if their 
choice was based on personality rather than principles.

...
 >>  In other words, the more carefully someone has to word the measure by
 >>  which their method is best, the less likely it will achieve the real,
 >>  underlying goals.
 >
 >> >>2S:I would like to understand the point you are making here.You seem
 > to be saying that <careful> wording is always dangerous because it
 > conceals rather than reveals reality.In contrast, I see careful wording
 > as the only way one has a chance of describing reality.
 >
 > In your case, your words:
 >>
 >>  > ..." their vote will proportionately increase the voting power of a
 >>  > representative" …
 >
 >> >>2S: Which of these words or their combination do you see as
 > concealing reality?

Your words "the voting power of a representative" are the most troubling 
words, for the reasons just explained.

Yes, careful wordings are important, but not if they are used to hide 
significant disadvantages.

I'll put it this way: When you look up a completely unfamiliar word in a 
dictionary, does the definition always clarify its meaning?  Of course 
not.  That's why it's important for people to use the word in a sentence 
in order to better understand it (or as a test to verify that someone 
understands its meaning.)

 >>  Experience is the ultimate judge.
 >
 >> >>2S:I would prefer to say, <ultimately>, <experience> and rational
 > thought can only provide us with provisional truths.

OK, I agree that a combination of experience and rational thought is a 
better combination, provided that "rational thought" is not limited to 
words.

...
 >> >>2S:Do you think it would be <a great method> for electing the French
 > National Assembly, the British House of Commons, or the US House of Reps?

I think APR would be a fine method to try within an organization where 
the stakes are not high.  But at a national level the method is 
vulnerable to the easy influence of money.

As an example of where it might be useful, the Wikimedia Foundation 
currently (last I knew..) uses single-winner elections for all its 
seats, and they need a more proportional method.  The result of 
mis-using a single-winner method is causing Wikipedia to be controlled 
by word-smiths (word-focused editors) and formatters, to the exclusion 
of subject-matter experts.  I have learned from experience (when I was a 
technical writer specializing in documenting especially complex 
technology) that this combination produces disastrous results. 
Accordingly Wikipedia is suffering from a loss of subject-matter 
experts, and the remaining "editors" are eroding the quality of what was 
earlier written by subject-matter experts, and it has led to a failure 
to improve articles that can only be improved by subject-matter experts. 
  A proportional method such as APR would make it possible for 
subject-matter experts to regain influence over Wikipedia content.

I think that covers your latest questions.  Thanks for your interest in 
my opinions.

Richard Fobes


On 11/6/2014 6:12 AM, steve bosworth wrote:
>
>
> ------------------------------------------------------------------------
> From: stevebosworth at hotmail.com
> To: election-methods at lists.electorama.com
> Subject: Subject: Re: APR [EM] Steve's most recent response to Richard Fobes
> Date: Thu, 6 Nov 2014 14:04:49 +0000
>
> 11/6/2014
>
> Topic:APR -- Steve Bosworth’s most recent response to Richard Fobes.
>
> (Steve’s most recent responses begin with >>>2S:).
>
> Note: Steve's article explaining how APR works will be emailed to you
> upon request.
>
> stevebosworth at hotmail.com <mailto:stevebosworth at hotmail.com>
>
>>  From: election-methods-request at lists.electorama.com
>>  Subject: Election-Methods Digest, Vol 125, Issue 5
>>  To: election-methods at lists.electorama.com
>>  Date: Sun, 2 Nov 2014 18:21:42 -0800
>>
>>  ----------------------------------------------------------------------
>>
>>  Message: 1
>>  Date: Sun, 02 Nov 2014 18:21:37 -0800
>>  From: Richard Fobes <ElectionMethods at VoteFair.org>
>>  To: election-methods at lists.electorama.com
>>  Subject: Re: [EM] Steve's response to Richard Fobes in
>>  Election-Methods Digest, Vol 124, Issue 28
>>  Message-ID: <5456E6B1.5070405 at VoteFair.org>
>>  Content-Type: text/plain; charset=windows-1254; format=flowed
>>
>>  On 11/2/2014 3:58 AM, steve bosworth wrote:
>>  > ...
> *>>S: Perhaps you did not notice that APR's 'primary' election is
>>  > different from the ones currently used in the US. ...
>>
>>  Yes I did notice that your recommended kind of primary is different.
>>
>>  We both agree that the better the primary election, the better the final
>>  results will be. This applies regardless of what kind of grouping --
>>  political party or association or whatever -- is involved between the
>>  primary election and the general/runoff election.
>
>> >>2S:Yes, both types of primaries help the <the final results> to be
> better. However, APR’s primary seems to have the additional advantage of
> allowing citizen to encourage more attractive candidates to be available
> to be ranked later in the general election.
>
>>
>>  > ...
>>  > Your counting method definitely has the focus-on-the-current-top-choice
>>  > "blinder" approach that I've already described.
>>
>>  >>S: I don't remember seeing this earlier description. Please, could
>>  you explain it again?
>>
>>  Apparently my first attempt to explain this concept was forgettable, so
>>  this time I'll say that it should become clear by comparing
>>  instant-runoff voting (IRV) with Condorcet-Kemeny voting (which Markus
>>  Schulze in Wikipedia has named the Kemeny-Young method). IRV looks at
>>  the ballots one choice at a time (the current "top" choice), which
>>  causes it to fail lots of important fairness criteria. In contrast, the
>>  Condorcet-Kemeny method looks fully into every ballot before identifying
>>  the winner, and therefore seldom fails those important fairness criteria.
>
>> >>2S:I understand this criticism of IRV and that using the
> <Kemeny-Young method> as a corrective might even be practically possible
> in single winner elections.However, am I correct in believing the this
> method would be entire impractical for the large multi-winner elections
> for which APR’s combined use of modifies STV and ‘Asset Voting’ would
> seems to be the best practical method – best because it allows each
> citizen to maximizes the chances that his or her vote will be added to
> the ‘weighted vote’ in the legislature of the rep he or she trusts
> most.APR seems to be practical both because it is technically possible
> and is relatively easy for ordinary citizens to understand.
>
> If my belief is correct, this is in sharp contrast to the relative
> difficulty of understanding Condorcet-Kemeny (even if used in a
> single-winner election) and the practical impossibility of using it for
> a large multi-winner election.
>
> …………………………….
>
>
> *>>S: I need more explanation of why you think APR would make bribing
>>  > easier. I see APR as being at least as secret and secure as any other
>>  > system. In fact, it might be argued that it would be even harder for a
>>  > 'briber' to infer whether the citizens who agreed to vote as the briber
>>  > wants had actually done so. This is both because each citizen has the
>>  > secret opportunity to rank any candidates in the whole country (as you
>>  > have also said) and each candidate has the structural possibility of
>>  > receiving votes from any citizens in the country.*
>>
>>  This concept is not obvious. It took me awhile to understand it.
>>
>>  The issue is not that anyone can know how a specific voter has voted.
>>  Rather a "consultant" or thug would have to deliver lots of "bought"
>>  votes. The "payment" to the consultants/thugs might be based on whether
>>  the candidate wins. Or the payment might be similar to the "commission"
>>  approach used to determine how much income goes to someone in a sales job.
>>
>>  This isn't a complete answer, but it's as much as I currently have time
>>  to write.
>
>> >>2S:I don’t yet understand the answer you have given.Yes, the
> <consultants/thugs> might pay a commission only if their candidate
> wins.However, when their candidate loses, my point was that they would
> find it more difficult, if not impossible, to discover whether this was
> because their bribed voters failed to vote as promised or the other
> voters who were not bribed but who were assumed to be supporters of
> their candidate decided not to vote for that candidate.
>>
> *>>S: I can see that people in such neighborhoods might be more likely
>>  > to agree to take such bribes but I do not yet see how APR makes it any
>>  > easier for a briber to enforce such agreements.*
>>
>>  The cross-district candidate voting, with resulting published numbers,
>>  is what makes it easier to measure the "success" of buying votes.
>
>
>> >>2S:Given that APR would allow voters an opportunity for many more
> <cross-district> or cross-association rankings, I do not yet see how any
> bribed citizen’s vote could be reliably inferred not to have kept his
> promise.If bribes were offered, a citizen might take the bribe even
> though he plans not to vote for the briber’s candidate.He might do this
> believing that he has a chance of gaining any way, i.e. if the briber’s
> candidate wins without his vote.
>
>>
>>  As a related point, if lots of voters in lots of districts vote for one
>>  specific celebrity politician, that indicates that the voting method is
>>  allowing lots of unrepresentative legislators to get elected.
>
>> >>2S:As I see it, this would not at all be clear.If <lots of voters in
> lots of districts (associations) vote for a specific … politician> that
> might simply mean that he is trust by many people nationwide, i.e. that
> he is their most trusted rep.
>
>
>>  > ......................................
>
>
> *>>S: Of course, this may be so, but I?m still looking for any flaws in
>>  > APR.I have not yet found one from the point of view of a citizen who
>>  > wants ?**an electoral system that would allow them to guarantee that
>>  > their vote will proportionately increase the voting power of a
>>  > representative (and) they most trust in their country?s legislative
>>  > assembly, [and also a system] that will give them every incentive to
>>  > vote, to vote for candidates they like, never having to vote tactically
>>  > or negatively.?**
>>
>>  Ultimately what voters want is wise and fair laws and effective tax
>>  rates (etc.). (A higher-level goal might be the overall economic
>>  prosperity of the nation/region, but that is difficult to correlate with
>>  legislative decisions.)
>>
>
>> >>2S:Yes, I hope most citizens do <ultimately what … wise and fair
> laws>.However, what exactly is <wise and fair> can be rationally
> determined, if at all, only through the widest possible and well
> informed debates.It seems to me that APR would help provide the most
> promising electoral and legislative assembly conditions for such
> debates.If so, APR, itself, might justly claim to be an essential part
> of what is <wise and fair>.
>
>
>>  It's possible that a voting method can focus attention on achieving
>>  specific numbers -- such as how many wasted votes there are, or how many
>>  "unrepresented" voters there are -- and yet fail to achieve the higher
>>  goal of wise and fair laws and tax rates.
>
>> >>2S:Of course, no electoral system or constitution by itself can
> guarantee <wise and fair laws>.All we can do here is try to discover
> which electoral system would maximally facilitate these achievements.
>>
>>  In other words, the more carefully someone has to word the measure by
>>  which their method is best, the less likely it will achieve the real,
>>  underlying goals.
>
>> >>2S:I would like to understand the point you are making here.You seem
> to be saying that <careful> wording is always dangerous because it
> conceals rather than reveals reality.In contrast, I see careful wording
> as the only way one has a chance of describing reality.
>
> In your case, your words:
>>
>>  > ..." their vote will proportionately increase the voting power of a
>>  > representative" …
>
>> >>2S: Which of these words or their combination do you see as
> concealing reality?
>
>
>>
>>  are the ones that other, possibly better, methods might fail to achieve.
>
>> >>2S:I have not explained my advocacy for APR well if I have given the
> impression of my mind being closed.I am very open to receive better
> argument in favour of a <possibly> better method.I look forward to your
> suggestions along these lines.
>>
>>  Again, each time someone carefully crafts a definition, others can
>>  ignore that definition, or come up with a competing definition that
>>  favors a different method.
>
>> >>2S:Of course, but I would like to consider your <definition> or the
> prime value that currently motivates your favoured system.Would you not
> want an electoral system that would enable you to guarantee that your
> vote would be added to the rep you trust most?
>>
>>  Again I'll say that I understand why the method appeals to you, and it
>>  might be a great method in some situations.
>
>> >>2S:Do you think it would be <a great method> for electing the French
> National Assembly, the British House of Commons, or the US House of Reps?
>
> But I suspect it will
>>  become corrupt over time.
>
>>  Sometimes a great sounding decision-making method turns out to have lots
>>  of non-obvious flaws that make it unworkable in government situations.
>
>> >>2S:Of course, even the best is subject to corruption.Currently, APR
> would be <corrupted> for me if someone found a practical way of making
> it secretly elect reps other than those favoured by the electorate and
> counted as described.
>
> I see some of my above answers as explaining some of the reasons why I
> think APR might be more resistant to corruption than other methods.Do
> you still see how it might be more vulnerable?
>
>>
>>  Experience is the ultimate judge.
>
>> >>2S:I would prefer to say, <ultimately>, <experience> and rational
> thought can only provide us with provisional truths.
>>
>>  Richard Fobes
>>
>
>
>
> ----
> Election-Methods mailing list - see http://electorama.com/em for list info




More information about the Election-Methods mailing list