[EM] Median systems, branding, and activism strategy

[Abd ul-Rahman Lomax]

At 09:55 AM 6/12/2013, Jameson Quinn wrote:
>As voting reform activists, we must work 
>together as much as possible. In general, that 
>means that raising awareness should start with 
>teaching people about approval. Still, if 
>someone is unsatisfied with the expressivity of 
>approval, we should have a backup offering.

That is, we should have a plan for future 
improvement. Really, unless a place already has 
IRV in place, a more complex ballot is not the 
first step to take. It's like our decision to 
promote Approval. An obvious first step, a 
do-no-harm improvement. More complex ballots mean spending money.

The actual reform we suggest should depend on local conditions. They vary.

>Personally, I think that median systems offer 
>the best backup offering in that sense. That 
>doesn't mean I intend to undercut people 
>promoting Score or Condorcet; just that I think 
>median systems offer a good compromise between 
>expressivity and low rewards to strategy. From 
>my perspective, Score is great for honest 
>voters, but for strategic voters it has exactly 
>the same expressivity problems that approval 
>does. And Condorcet is too complex — not just to 
>describe in the abstract, but to present the 
>results of even a single election in a clear, intuitive form.

Uh, Score systems can ba amagalamated as median 
or as average or as sum. The latter fits most 
closely with democratic traditions, it is just 
Count all the votes, i.e., Approval, plus the allowance of fractional voting.

Median systems are introducing what appears to 
be, for public choice, an *entirely new* concept. 
So, the obvious question, why?

And, remember, we are talking about first reforms, not the *ideal*.

>But median systems have a problem. There are too 
>many of them, and even more names.
>
>Off the top of my head, I can think of the following names:
>    * Bucklin: A general class of median systems 
> which are implemented via a descending 
> threshold. Also used to describe various 
> specific ranked or hybrid ranked/rated systems 
> used during the progressive era. For instance, 
> some use "Bucklin" to mean "median using full 
> or truncated rankings, using the highest 
> majority as a tiebreaker"; others mean the 
> Grand Junction system of "median using 3+1 
> numbered rank/grades with skipping allowed and 
> ties allowed at the third grade only; highest majority tiebreaker".

Bucklin is a median system, that's true, but that 
is not the *concept.* And any poor performance of 
Bucklin would come, quite likely, from using 
median amalgamation. (I.e., a Bucklin ballot is 
really a Score ballot, and a score ballot can be 
used for Bucklin amalgamation, it is an obvious and simple extension.)

>    * Majority Choice Approval (MCA): I've seen 
> this applied to various 3-rank median systems, 
> but I think the canonical one is: "if there is 
> a majority top-rank, then the highest such; 
> otherwise, the candidate with the most non-bottom rankings."
>    * Majority Judgment (MJ): as defined by Balinski and Laraki
>    * Graduated Majority Judgment (GMJ): as defined by me.
>There are also a number of possible descriptive 
>"branding" terms for a median / Bucklin system:
>    * Instant Runoff Approval
I created that one, because that is exactly what 
it is, and it simulates a series of descending 
cutoff Approval elections, hence a way-cool 
implementation, as a primary and runoff method in 
a two-round system, thus, if it has N approval 
grades, it simulates 2*N approval runoffs in the 
full two-election set. Jurisdictions that want to 
ensure majority approval can probably reach that, 
even with many candidates, in a two-round Bucklin 
system, and I've suggested that if the Range 
ballot is a little deeper, i.e, includes 
unapproved ranks, it can also, in the primary, be 
analysed for a beats-Range top two (or 
beats-Bucklin top two) winner, using the lower 
preferences, thus making the *system* Condorcet 
compliant, if a Condorcet winner must always be 
in the runoff. If a majority is found, depending 
on the wishes of the jurisdiction, the runoff may not be necessary.

Increasingly, though, I've become aware of 
systems that *always* go to the general election. 
So the primary is just a unified primary, as 
distinct from party primaries. And with a good 
runoff method, and with write-ins allowed in the 
runoff, the electorate can still have three 
choices on the ballot, if certain conditions are 
met, plus write-ins that don't necessarily vote-split.

What, I'd like to know, is not to like about this?

>    * Graded Instant Runoff
>    * Descending Threshold Approval
>    * Majority Threshold
>    * Grade Voting
>    * Majority-Based Grade Assignment
Instant runoff Approval fully satisfies 
expectations. It is what the name implies.

>It seems to me that Median and Bucklin advocates 
>should come to a consensus on what specific 
>system to promote, and what to call it. That 
>doesn't mean we should cease discussing the 
>different systems in fora like these; just that 
>when promoting systems to the public, we should be on the same page.

Bucklin is widely known as Bucklin, the Grand 
Junction system, and it was simply called 
"preferential voting" in many places. It was also 
called "American Preferential Voting," to 
distinguish it from "foreign" systems (i.e, STV methods).

>Which system is best? I think the clear choices 
>are MCA or GMJ, and I'd personally favor GMJ. 
>MCA is the simplest well-defined median system. GMJ is good because:
>    * Unlike the specific systems called 
> "Bucklin", it has no vestige of ranked 
> thinking, and thus requires no dishonest strategy.
That's naive.

>    * It's more expressive than MCA.
>    * Unlike MJ, GMJ can easily be expressed in 
> terms of a Bucklin-like descending threshold, a 
> single algebraic formula, or a graphing 
> procedure. (As far as I know, MJ can only be 
> expressed in one way). GMJ is also easier to 
> "program" into a spreadsheet in my experience.
>    * Unlike MJ, GMJ results can be succinctly 
> and unambiguously expressed as a single number for each candidate.
>    * Unlike MJ, the GMJ winner for a given 
> honestly-voted utility profile tend to be 
> stable as the number of evenly-spaced grading 
> categories varies, even in moderately 
> "pathological" examples (such as a 
> single-peaked versus a two-peaked candidate).
>    * However, the actual results will agree 
> with other median systems in almost all realistic cases.
>What should we call it? GMJ isn't a horrible 
>name, but if people prefer to use one of the 
>above "branding" terms or something similar, I'd be open to discussing it.

I *very much* dislike the use of letter grades 
that imply absolute ratings. It is essentially 
suggesting to the voter that they disempower 
themselves. The whole discussion about "dishonest 
strategy" is contaminated with judgment against 
the normal think we do with choices, our internal 
ratings are highly dependent on expectations *in 
the particular case*, and people who don't do that tend to be unhappy.

>But I think it's important for us to join our 
>voices in better harmony on this. Abd, in 
>particular: why do you continue to talk about 
>"Bucklin" (ie, the grand junction system) when 
>there are better-designed, more-clearly-defined 
>Bucklin systems available today?

The Grand Junction System is adequately defined, 
but the system I'd recommend now is Bucklin-ER, to start. Three ranks.

All three ranks are approved ranks. In *some* 
elections, two ranks might be used.

>Jameson
>ps. None of this message should be read as an 
>attempt to abandon the common effort to promote 
>approval voting as a first step.

This is very simple. Approval first. The obvious 
first and most significant objection to Approval, 
then, is addressed with ranked approval. There is 
a method, approval-IRV, that has the LnH behavior 
of IRV, if the voter wants that protection. But I 
actually think it harms voter behavior to encourage excessive fear of LnH.

Look at the arguments for GMJ above. They will 
mostly be unintelligible to ordinary citizens who 
are not specialists in voting systems.

So, I see this as one reform path.

1. Approval.
2. Bucklin-ER, 3-rank. (Traditional Grand 
Junction system plus equal ranking *allowed* in all ranks.)
3. Full range ballot for Bucklin [adds . (*Range 
reporting starts, and Condorcet analysis becomes more complete*)
4. Runoff with pairwise leader included if 
different. Runoff if majority not found. Or unconditional runoff.
5. Range amalgamation substitutes for Bucklin 
amalgamation. Resolution may be increased if 
needed. Approval information is always collected.

All of this works with runoff voting. We should 
*not* attempt to kill runoff voting as an *easy 
target*, and we should oppose FairVote attempts 
to do this, there is basic information about this 
to be included, and providing information for intelligent decisions is our job.

(If it were practical, approval voting with 
repeated Bucklin ballots until a majority is 
found is a *highly sophisticated system*.)

We don't know the Bayesian Regret performance of 
Bucklin, because the testing that was done, 
AFAIK, did not understand a Bucklin ballot as 
being a Range ballot, but treated it as a ranked ballot.

The Bucklin ballot I'd see at the 3rd step could 
even have letters attached, it's really a Range 4 
ballot, just as the step 2 ballot is that ballot 
without the elevated but unapproved rating of "1: 
included. So A is 4.0, as people expect. But I'd 
want to see the method in actual use *first*, to 
avoid misleading voters. If you really want to 
give all politicians a bad rating, then you 
should totally realize that you are disempowering 
yourself, not casting a full vote, to *make a choice.*

Bucklin could be reported, by the way, as a net 
grade point average for candidates, the values 
averaged would be the numbers of 0 plus 2-4. If 
the additional unapproved level of 1 is included 
on the ballot, it's then a complete Range set. 
And candidates can be given that single "grade" 
that Jameson wants. Voting is *choice*, not absolute rating.

If I could put full and accurate labels on a 
Bucklin ballot, the ratings would convey this:

4. Favorite(s) or Best
3. Good, Preferred over others.
2. Barely Accepted, better than holding a runoff. 
(If runoffs are triggers by majority failure.)
1. Disapproved, but better than the Worst.
0. Worst. No support in any way.

(If one sees a nice candidate, with the 
misfortune to be in a field of better candidates, 
hey, give him a zero, or maybe a 1 if there is 
significant write-in activity, but ... send him a 
donation to encourage him to keep plugging away! 
"I rated you 1, but only because there really 
were some great other candidates. But next 
election, I could vote for you. Keep up the good work!)

A rating of 1 will not ordinarily elect a 
candidate, per se, but *might* get the candidate 
into a runoff if pairwise analysis is done. It is 
a vote against the candidate, not any kind of 
approval of the candidate. It is possible that if 
other voters, by a majority, approve of that 
candidate, the vote of 1 might edge the candidate 
to first position in, say, Range amalgamation. As 
I'd have it, without majority approval, nobody is 
elected and the ballot becomes a method of 
providing ballot position in the runoff.

Now, the kicker, full circle. The most 
sophisticated system proposed here can be 
improved with a vote-for-one ballot, to create an 
assembly or other body to handle single-winner 
elections deliberately. Asset Voting. The chosen 
electors -- total free choice by the voters, no 
worry about wasted votes -- unless, of course you 
vote for a wasted idiot --  could easily use *any 
ballot method* with repeated elections, until a 
decisive and true majority winner emerges. And 
that election, on principle, *could be revoked at 
any time.* Asset. Radical, a complete revolution 
in how we think about politics and elections. 
Most people still hold on to expectations from 
the present system, when considering Asset.

They imagine, routinely, *the same political 
behavior seen with winner-take-all elections.* 
They imagine that candidates must still try to 
get as many votes as possible in the general 
election. They imagine that the need, then, to 
buy expensive media coverage will continue. They 
imagine the continuance of the common 
oppositional obsession, when Asset rewards 
nothing but cooperation. If electors cooperate, 
they elect more seats. If they don't, they elect fewer. No big deal.

I have no idea how far the single-winner reform 
path would go before an Asset or STV reform. 
Approval, yes, *immediately*. Bucklin could come 
vary soon, and, before I close this, I want to 
remind readers that Bucklin, aside from top-two 
runoff, is the most widely tried advanced voting 
system in U.S. history, aside from what I just discovered about Vermont.

Vermont had pure vote-for-one *majority required* 
going for a long time, they simply kept voting 
until they got it right. That is a *highly 
advanced system*. It's easily improved, made more 
efficient, with Approval or Bucklin methods. That's what's never been done.