Wednesday 10 September 2014

CSM Matias Otero resigns, CSM Asayami Dei appointed


For those avoiding EN24, reddit, and the forums, Matias Otero, the Brave Newbies representative has been inactive for sometime, and has recently made his resignation official.

Matias has  issued his apology on reddit including
...  I've decided to resign from the CSM rather than let my seat go to waste. It was a difficult decision, as I felt that I could at least have spent some of my energy on the all-important issue of new player attraction and retention, but in the end I had to admit to myself that it wasn't going to happen. ...


CCP then had the decision on whether to replace Matias or leave a seat vacant.  If they were going to replace him, how?

So lets look at the options https://en.wikipedia.org/wiki/By-election#In_multi-member_constituencies is a starting place.

Option: chose from losing candidates at previous election via recount

The voting software that CSM uses easily lends itself to excluding the failed candidate.  This has the advantage of allowing voters, especially those who voted for the resigned representative, to still have their say via their voting preferences.

If CCP chose this method, then Asayami Dei is elected, defeating Awoxing Pizza-Spymaster McBlueshooter by (my count) of 181.

The (my count) qualification is that I used droop with the warren rules, which is probably not quite right.  However Sephira Galamore on the forums posted that with CSM's own voting software, that this is also the same result.

Most systems that perform this type of recount keep all incumbents in their seat and use the recount solely for electing the vacancy. (Addressing Sephira's concern on the forums)


Option: Leave the seat vacant

Do nothing is always an option, but not really one that benefits anyone.  While the actual goals of what the CSM is for are perceived differently by some, very few of these goals are met by fewer CSM members.

Option: replace with a similar candidate by fiat

Many electoral systems will appoint a member from the same political party where available.  Whether this appointment is done by a Governor, an elected body themselves, an effort is generally made in many systems to retain the existing balance by replacing 'like with like'.

Eve does not (yet) have formal political parties.  We are close but not there yet. I do not see how this would work until we have formal tickets formally acknowledging who is in the same group.



Option: replace with the last excluded candidate

This is the option that was taken by CCP in the interests of speed.  Umm.  No disrespect to Asayami Dei but please CCP, never do this again.

This has the exact opposite effect of 'replace with a similar candidate' and generally will disenfranchise those that voted for the recently vacated representative.

For an example in the extreme.  Take a common voting scenario
  • Right wing candidate gets 49% of vote
  • Left wing candidate gets 35% of vote
  • Center Right candidate gets 14% of vote
  • Others get 2%
So on the initial vote (regardless of method) the right wing candidate is elected, and 1/2 way through their term resigns.

Under this method the left wing candidate would be elected, causing the right wing electors (with 63% of vote) to be disenfranchised.

The reason that Asayami is appointed under both this rule and recount method is that in the original count Asayami was only just excluded (80 votes).

Speed was not an issue.  I found and installed and worked out how to drive similar software to perform a reasonable count within a couple of hours.  Seripha found the software that CCP uses and re-ran the vote in the same amount of time.

Summary

To minimise accusations of favouritism and whim, I recommend that CSM ask CCP for a formal method of replacing candidates, and to formalise when a representative should be replaced.

My personal preference is a recount method.

I think it is fortunate that Asayami managed to be elected on both the whim based 'last candidate excluded' as well as a more rigorous recount method.


Links

drool warren based on original vote
Original vote modified to work with drool

drool warren with matias excluded
Original vote modified to work with drool and then Matias excluded

Original vote as provded by ccp

7 comments:

  1. This comment has been removed by the author.

    ReplyDelete
  2. Recounting could be done with trivial effort, by loading the ballot file into Notepad and deleting the code of Matias. So an "X, Y, Matias, Z" vote changes into "X, Y, Z".

    Then re-run the vote counter program.

    The reason why their "first excluded" gave the same as a recount can be seen on my cross-vote table: http://greedygoblin.blogspot.hu/2014/05/csm9-election-analysis-rewrite.html

    Matias voters had no particular second preference, those who voted for Matias, voted for other candidates randomly. So Matias was the only candidate from his "party".

    ReplyDelete
  3. @Gevlon : Not quite.

    Mcblueshooter goes from being knocked out at round 13 of 33 to being defeated at round 32 of 32

    For workings : compare the 'drool warren based on original vote' with the 'drool warren with matias excluded' links in the bottom of the main article.

    The *only* difference between the 2 vote sources is that Matias is excluded.

    In the primary votes, Matias has 1453 and Mcblueshooter 355

    In the re-run, on the first round, Mcblueshooter jumps up to 831, or gets 33% of now available Matias' vote. Not a landslide, but still significantly more than random.

    At the end of the day, it is clear (to me at least) that the 'by fiat' approach of CCP is the same as a recount option. It is less clear that Mcblueshooter is on a different party ticket to Matias. While the flows are weak, they are not random.

    I need to have a look at who else got a better than random preference flows, but that is a task for tomorrow

    ReplyDelete
  4. I seems to remember from looking at some corner cases that the code has the ability to withdraw candidates by adding the list of withdrawn candidates. So it's even more trivial than Gevlon points out. There is no reason not to do it and plenty of reasons it is the right thing to do. I kind of wish somebody who knows this shit had educated leelo about the voting system before this happened.

    ReplyDelete
  5. The ability to exclude the candidate is provided by adding a single line near the top of the file.

    In this instance, inserting a new line 2; with the value of -27 (Matias' candidate number)

    ReplyDelete
  6. Sorry, a recount still suffers from the same problem as last excluded candidate.
    The votes were cast under a specific scenario (with all the candidates in the running). There is no guarantee that any voter would vote the same way if you re-ran the election (complete with time machine and the removal of the candidate stepping down). Votes were cast under specific circumstances and are invalid if the circumstances change.

    Further, I see no evidence that doing a recount in this manner would produce an outcome "better" than a random number generator.

    ReplyDelete
  7. @Ael : your evidence is at http://en.wikipedia.org/wiki/List_of_Tasmanian_House_of_Assembly_casual_vacancies

    Using a recount method, 39 vacancies filled since 1965. 38 to to same party; 1 moving from one 'independent' (Australian Democrats) to another 'independent' (Greens)

    Either 38 out of 39 'like for like' members being elected on a recount, in a real world does not appear random to me.

    Of course selection by recount assumes that people voting now will vote the way they did during the original election. This is a polite fiction.

    I use both Tasmania and Australia as a point of reference, as it is the first life example I have followed most. It has 3 main factions; not as many as Eve does but a very rough approximation.

    You will find that STV usually polarises between candidates of different 'sets'.

    For example following the scrutiny sheets available at http://www.electoral.tas.gov.au/StateElection/Results.html (clicking on 'Division of xxx results at the bottom). I understand and accept that the scrutineer sheets are hard to read at first. It is very unusual to have 'take the next candidate' give what could be decided as 'fair'.

    I have provided references, and real world examples of how STV systems have worked. These are not the only ways for the system to work.

    There are some (both here and on EN24) that don't like my analysis. From what I understand it is because it doesn't fit with their preconceptions.

    I also accept that STV will occasionally turn up a result for the last seat that appears somewhat bizzare. <a href ="http://blogs.abc.net.au/antonygreen/2014/08/ricky-muirs-strange-path-to-the-senate.html#more>http://blogs.abc.net.au/antonygreen/2014/08/ricky-muirs-strange-path-to-the-senate.html#more</a> (TLDR: micro parties between them get 17% of the vote; one of the micro party members gets elected on 6th/last spot)

    Stating that you disagree is fine but won't change anyone's mind unless you provide examples and what others have done. Even then, with the best of examples and logic, you might not change people's minds.

    My own summary:
    * Appointment by recount is as likely as anything to provide 'like with like' candidates.
    * Appointment by last excluded is as likely as anything to provide 'like with opposite' candidate.
    * Likely is not the same as always
    * All elections can provide weird results; STV is no exception

    ReplyDelete

Posts older than 14 days are subject to moderation before being published. I do so sporadically. If you have a question regarding older posts, also evemail dotoo foo.

Blogger comments supports basic html. You can make a link 'clicky' by <a href="http://yoursite/yourpage">yoursite/yourpage</a>

While I currently accept anonymous users, please include a pseudonym. I get confused answering anonymous.

If the word verification is preventing you from adding a comment, please evemail DoToo Foo for alternative methods