?

Log in

No account? Create an account
Vote early -- and often - Many a mickle maks a muckle

> Recent Entries
> Archive
> Friends
> Profile

May 23rd, 2008


Previous Entry Share Next Entry
12:19 am - Vote early -- and often
(Quote source.)

So the voting has begun for the LiveJournal Advisory Board election. Three quick observations:

1) The rules of the election have changed at the last minute with no notice. For the better, admittedly, and changing only the wording of the counting process, not the spirit of what was intended. (Have they been reading Iain? I hope so, but based on Iain's sequel post, I fear not.)

2) You can look at who's winning in real time, before and after you cast your vote.

3) Aaaaaand you can even change your vote should you want to. You can probably even change it after the poll closes, come to think of it. (ETA: apparently not. Thanks, imc. There are also rumours that they may remove the ability to change your vote mid-election, which - while it would make sense - would make this kooky election even more irregular still. Introducing new rule, and all that.)

What a strange way to run an election.

I have to say that property 3 above has an interesting side-effect which I consider beneficial, if possibly unintended; should no candidate reach 51% of votes after the Instant Runoff process - which is something I don't think we'll be able to tally from the real-time results - then it'll be down to whoever gets the most first preferences. It occurs to me that should a candidate conclude that they are unlikely to win, then by posting in their journal, they probably have a good chance of convincing those who voted for them as first preference to change their vote in favour of some other candidate of their choice.

One might expect that a candidate with, say, 4% of first preferences might be able to convince many of the people who voted for them to change their first preference in some other direction and thus add, say, another 3% of first preferences in favour of some other candidate, should it look like the "who has most first preferences" metric might be pointing in an undesirable direction.

If - if - you happened to be standing on a platform emphasising some sense of fun, it would appear a desirable property to me that your sense of fun were not abhorrent and malicious.

Non-Friends' comments screened. No, I'm making no pretence at impartiality.

(15 comments | Leave a comment)

Comments:


From:songmonk
Date:May 22nd, 2008 11:57 pm (UTC)
(Link)
If you can change your vote even after polls close, how can one ever call the election?
[User Picture]
From:imc
Date:May 23rd, 2008 12:11 am (UTC)
(Link)
Aha, but actually LiveJournal has now added the ability to close a poll. It was probably implemented for the specific purpose of this election, but as a bonus the ordinary users get to do it too (just go to one of your polls and click on the poll number, then look at the list of options at the top).
From:songmonk
Date:May 23rd, 2008 12:15 am (UTC)
(Link)
I'm confused. I realize that jiggery_pokery said "probably" in "You can probably even change it after the poll closes", but still:

If you can change your vote after the poll closes, how does the ability to close a poll resolve the issue?

Or is there a difference between closing a poll and closing a poll?
[User Picture]
From:bopeepsheep
Date:May 23rd, 2008 05:55 am (UTC)
(Link)
Normally the closing date simply means "the date at which we grab the results and count them" - votes can still be added or changed after that. This is what jiggery_pokery means, the change in poll functionality having sneaked in very quietly.

If you actually close a poll, you can no longer vote in it, or change your vote. So whenever you revisit it, it will be a true picture of the results at the closing date.
[User Picture]
From:imc
Date:May 23rd, 2008 12:16 am (UTC)
(Link)
Have they been reading Iain?

I actually doubt that, but I do have a current developer and a former developer on my friends-of list, and it's also possible that whoever drew up the formal list of candidates had a quick look at my journal too (since I am on that list). Not that this means much, but I'm just sayin'. ;-)
[User Picture]
From:bopeepsheep
Date:May 23rd, 2008 05:56 am (UTC)
(Link)
Different Iaiaiaiain. Look at the links. :D
[User Picture]
From:imc
Date:May 23rd, 2008 09:25 am (UTC)
(Link)
I'm aware of that — I've even had a (short) correspondence with Iain about my analysis of the election algorithm. So the "but" in my comment was meant to imply more separation than you read into it. :-)
[User Picture]
From:bopeepsheep
Date:May 23rd, 2008 09:31 am (UTC)
(Link)
Well you're not going to win any elections with that level of subtlety...
[User Picture]
From:imc
Date:May 23rd, 2008 10:35 am (UTC)
(Link)
I'm not going to win the election anyway
[User Picture]
From:andrewducker
Date:May 23rd, 2008 09:16 am (UTC)
(Link)
Sadly, you can't really tell (2) - because you can't see where the votes will end up...

I've always thought it would be interesting to see how the vote tallying was going over time - and to extend voting to last longer. It would solve the problem of people saying "I didn't realise that it was going to be so close, or I'd have voted." and might encourage higher turnout on a second day in close constituencies.
[User Picture]
From:jiggery_pokery
Date:May 23rd, 2008 10:02 am (UTC)
(Link)
*nods* Agreed, but see the preantepenultimate* paragraph.

Help me think this through here, please. Suppose the clearly front-running candidate has, say, 18% of first-preference votes, 15% of second-preference votes and 13% of third-preference votes. Can we prove that because 18% + 15% + 13% > 50% (+1 vote), they - and by extension, no other candidate - aren't going to have the 50% + 1 vote required to end the count in the Instant Runoff? Even if that can't be proven, will it work as a reasonable rule of thumb in all but the most pathologically constructed cases?

It would solve the problem of people saying "I didn't realise that it was going to be so close, or I'd have voted." and might encourage higher turnout on a second day in close constituencies.

In real life, I fear this would cause more problems than it solves. Being able to try this out on LJ is both fun and may add to the sum of human electoral knowledge.

* copyright Iain Weaver's Long Words Dictionary.
[User Picture]
From:andrewducker
Date:May 23rd, 2008 10:29 am (UTC)
(Link)
they - and by extension, no other candidate - aren't going to have the 50% + 1 vote required to end the count in the Instant Runoff

Can you rephrase that with less double negatives? Do you mean "every other candidate"?

And someone could be in second place on 17% of first-preference votes, but have 35% of second-preference votes, which would give them the position. I don't know how likely that is though.

What's interesting is that we do have a few people with more second-preference votes than first.

Now, as a voter, I have no interest in reading through 22 documents telling me why I should vote for that person. So I'm going to wait until the end, and read the ones that look likely to win. Especially as I can only have two transfers on my vote, so it's pointless paying attention to anyone further down.

Hmmmmmm. This kind of bias is going to be catching. People are going to see Jameth in the lead, go read their statement, and then vote if they seem like a good choice. This is automatically going to bias things in favour of the people who get a boost up-front - particularly when 99% of the electorate start off ignorant.
[User Picture]
From:jiggery_pokery
Date:May 23rd, 2008 10:53 am (UTC)
(Link)
Do you mean "every other candidate"?

If you undouble the negative in the second half of the sentence: yep. :-)

And someone could be in second place on 17% of first-preference votes, but have 35% of second-preference votes, which would give them the position. I don't know how likely that is though.

Doesn't seem likely to me.

Another thing to realise is that (as I type) there have been o.t.o.o. 5600 first preferences cast, o.t.o.o. 4800 second preferences and o.t.o.o. 4400 third preferences - so even if 35% of second prefrence votes were cast for one candidate and transferred across, that's 35% of a smaller number than the number of first preferences cast, even before taking into account people voting erroneously for one candidate as all three preferences.

I nod sadly in full agreement with the rest of your post.
[User Picture]
From:wmk06
Date:May 23rd, 2008 11:09 am (UTC)
(Link)

If - if - you happened to be standing on a platform emphasising some sense of fun, it would appear a desirable property to me that your sense of fun were not abhorrent and malicious.


That about sums up how I feel about candidate position statements. I didn't really know who to vote for in the run-off slots so I read every last one of them. I was kind of amazed that the "fun" platform people were running and that they...seemed kind of crazy and not especially fun.
[User Picture]
From:jiggery_pokery
Date:May 23rd, 2008 07:08 pm (UTC)
(Link)
I salute you and thank you for going about things exactly the right way. :-)

> Go to Top
LiveJournal.com