Vote counting procedure

[jjefferies]

Ok, so there's already a thread on this subject floating around. But it's pretty dated and who knows how long it really is. This should already be on both the USFA and USFFC websites.


VOTE COUNTING PROCEDURES



With input from both presidential candidates the Election Committee has adapted the following method for implementing preferential voting for the offices of Vice President as required by the USFA Bylaws.

For the Vice president positions, the candidates are listed in order based on their vote totals and the three candidates receiving the greatest number of votes are designated as elected. In the case of an absolute tie for one of the elected positions, the following tie breakers will be used in the order indicated:

1. The candidate receiving the greatest number of "1" votes.
2. The candidate receiving the greatest number of "2" votes.
3. The candidate receiving the greatest number of "3" votes.
4. The drawing of lots.

When tallying the tie breakers, only "1", "2" and "3" votes will be counted. Ballots where the voting was indicated by an "X", or checkmark, or other indication are not included.



The Election Committee

[telkanuru]

About damn time.

[mfp]

Originally Posted by jjefferies

VOTE COUNTING PROCEDURES

With input from both presidential candidates the Election Committee has adapted the following method for implementing preferential voting for the offices of Vice President as required by the USFA Bylaws.

For the Vice president positions, the candidates are listed in order based on their vote totals and the three candidates receiving the greatest number of votes are designated as elected. [...]

The Election Committee

This process clearly does NOT follow the bylaws. In fact it pretty much makes a farce of the bylaws.

Is also amazingly sad that both presidential candidates (and perhaps their slates?) seem to be in favor of ignoring the association's bylaws right out of the gate.

Some of the very notable violations of the bylaws.

1) The bylaws state National officers *shall* be elected by a majority of the votes cast, not a plurality:

Originally Posted by bylaws

8.Section Elections by Majority. The National officers of the USFA shall be elected by a majority of the votes cast.

2) The bylaws also state a preferential voting system is to be used, "Where three or more candidates have qualified for any office, the ballot shall be designed according to a system of preferential voting."

Any attempt to claim the Election Committee is following the bylaws on this point because they "designed a preferential ballot" (while not really running a preferential election) would be laughably disingenuous.

3) Furthermore, the bylaws state:

Originally Posted by bylaws

"[p]rocedure relating to contested elections shall be governed by the provisions of Robert*s Rules of Order.

So while the Bylaws require "majority" and "preferential voting" and Robert's defines the terms, the election committee's process fails to comply.

Is it really a good idea for the Election Committee to fail their duties to the Association and its membership and run an "election" of 3 National Officers using a process that flagrantly fails to comply with the bylaws?

Is it really a good idea for the presidential candidates to go along with the bylaw violations? How long after the "election" will the association be at risk for someone to challenge the flawed election?

While the candidates may have signed an arbitration agreement with the Election Committee, the rest of the membership and world certainly did not.

[Philistine]

Originally Posted by mfp

This process clearly does NOT follow the bylaws. In fact it pretty much makes a farce of the bylaws.

{snip}

1) The bylaws state National officers *shall* be elected by a majority of the votes cast, not a plurality:

{snip}


2) The bylaws also state a preferential voting system is to be used, "Where three or more candidates have qualified for any office, the ballot shall be designed according to a system of preferential voting."
{snip}

I'm not convinced that in the case of VP's this actually constitutes 3 or more candidates qualifying for office. I think a good argument can be made that each VP position is a separate office--and that the provision on preferential voting was meant for instances where you'd have 3 or more people running for, say, President.

On the issue of majority--while it's possible there will be a non-majority for 1 or more of the VP's, my WAG is that it is unlikely.

My only complaint with the procedure is the exclusion of X's, checks, etc. Given the way they're counting votes, there doesn't seem to be any reason to exclude them, because they could be fully counted--and if they were to be excluded, I think the ballot should have made that explicit--live and learn for next time, I suppose.

--Philistine

[mfp]

Originally Posted by Philistine

On the issue of majority--while it's possible there will be a non-majority for 1 or more of the VP's, my WAG is that it is unlikely.

It's actually impossible for more than 1 one candidate to achieve a majority of votes cast with the process described in the original post.

[oiuyt]

Originally Posted by Philistine

My only complaint with the procedure is the exclusion of X's, checks, etc. Given the way they're counting votes, there doesn't seem to be any reason to exclude them, because they could be fully counted--and if they were to be excluded, I think the ballot should have made that explicit--live and learn for next time, I suppose.

The procedure, as decided upon by the Election Committee, doesn't ignore "X" votes. They aren't counted during the tie-breaking phase, but they do count towards the initial totals.

-B

[dekko]

Originally Posted by oiuyt

The procedure, as decided upon by the Election Committee, doesn't ignore "X" votes. They aren't counted during the tie-breaking phase, but they do count towards the initial totals.

-B

Not for nothin' B but I usually trust your info but when you and a rep from the elec comm differ on info.....

He says they don't count, you say they do, someone is wrong here. They either count or they don't.

[epeemike81]

Originally Posted by dekko

Not for nothin' B but I usually trust your info but when you and a rep from the elec comm differ on info.....

He says they don't count, you say they do, someone is wrong here. They either count or they don't.

They don't disagree... re-read the procedure. the preferential 1/2/3 don't come into play until the tiebreak phase. The first phase of vote counting is simple number of votes (be they 1, 2, 3, x, check mark, whatever):

Originally Posted by jjefferies

When tallying the tie breakers, only "1", "2" and "3" votes will be counted. Ballots where the voting was indicated by an "X", or checkmark, or other indication are not included.

WHEN TALLYING TIE BREAKERS... that does not state that those ballots won't be counted in the initial process (and the inclusion of that clause actually strongly implies that they will).

-m

[seak]

Originally Posted by mfp

It's actually impossible for more than 1 one candidate to achieve a majority of votes cast with the process described in the original post.

Nah Philistine is right, you're thinking about it wrong, especially as it applies to Roberts.

Think about it this way instead, there are 3 VP's being elected, so I get three votes, one vote for each VP position (and you can't vote twice for the same person under this system). So if three people were voting, person A votes for 1,2,3, person B votes for 4,5,6 and person C votes for 1,2,4 then

1 is elected with 2/3 or 67%, as are 2 and 4.

There would be a problem if there were more then 6 candidates, but we don't have that situation. The RRO model that you reference is designed for one position with more then 2 people running for it, not for 3 positions with 6 people running for it.

[dekko]

So on the first count the Xs and checks are counted. If and only if there is a tie do the 1-2-3 votes count? So someone can vote for someone for VP and there is a good chance it won't matter? Also, is the tie only for 3rd? If 1st and 2nd tie there is no 'second' count?

Now I am more confused. If person A put three checks and person B put 1-2-3, person As vote is counted and person Bs vote is not considered unless there is a tie? If no tie no 'second' count and the 1-2-3 votes are counted, so they never count? Why would I vote with 1-2-3 instead of Xs or checks?

[Philistine]

Originally Posted by oiuyt

The procedure, as decided upon by the Election Committee, doesn't ignore "X" votes. They aren't counted during the tie-breaking phase, but they do count towards the initial totals.

-B

Thanks. I misread it.

Well, then, I withdraw my complaint.

Looks good to me.

--Philistine

[Philistine]

Originally Posted by dekko

So on the first count the Xs and checks are counted. If and only if there is a tie do the 1-2-3 votes count? So someone can vote for someone for VP and there is a good chance it won't matter? Also, is the tie only for 3rd? If 1st and 2nd tie there is no 'second' count?

Now I am more confused. If person A put three checks and person B put 1-2-3, person As vote is counted and person Bs vote is not considered unless there is a tie? If no tie no 'second' count and the 1-2-3 votes are counted, so they never count? Why would I vote with 1-2-3 instead of Xs or checks?

AIUI--the way it works is the 1-2-3 marking as such only comes in place if there is a tie. Otherwise a 1, 2, 3 or X all count equally as 1 vote for the indicated candidate.

So if someone gets 300 1st place votes, 200 2nd place, 50 third, and 5 x's they get 555 votes.

They will be treated exactly the same way as someone who got 555 x's, or 555 2nd place votes or some other combination adding up to 555.

The top 3 vote - getters will be elected, and it is only if there is a tie between one or more vote-getters that you look at the number (1st comparing # of 1st place votes, and the one with the highest breaks the tie, then, if still tied looking at 2nd place and going to 3rd if still tied after 2nd place).

The only kink is that it is theoretically possible that one of the VP's may not get a majority of votes cast. Again, my understanding is if there were, for example 1,000 votes, you'd have 3000 total votes cast, but not more than 1000 for any one candidate. Thus, you'd need 3 candidates to get at least 501 votes each to elect them with a majority.

--Philistine

[mfp]

Originally Posted by seak

Nah Philistine is right, you're thinking about it wrong, especially as it applies to Roberts.

[...]

The RRO model that you reference is designed for one position with more then 2 people running for it, not for 3 positions with 6 people running for it.

Reread your copy.

The preferential voting described in RRO is also for:

Originally Posted by RRO

If more than one person is to be elected to the same type of office -- for example, if three members of a board are to be chosen -- the voters can indicate their order of preference among the names in a single list of candidates, just as if only one is to be elected. The counting procedure is the same as described above, except that it is continued until all but the necessary number of candidates have been eliminated (that is, in the example, all but three)

It's clear that the RRO procedure is designed for such elections.

Originally Posted by seak

Think about it this way instead, there are 3 VP's being elected, so I get three votes, one vote for each VP position (and you can't vote twice for the same person under this system). So if three people were voting, person A votes for 1,2,3, person B votes for 4,5,6 and person C votes for 1,2,4 then

1 is elected with 2/3 or 67%, as are 2 and 4.

The system described is called a "Plurality at Large" form of election. It is not a preferential type of election. It does not elect by majority. Even its very name should provide a clue.

[PeterGustafsson]

Hi!


A few comments:
1. Does the EC mention how bullet votes should be counted, if at all? Those are votes in which the voter has only designated one (or in this case, two) candidates of his choice.
2. The general design of the voting procedure - bloc voting - resembles Approval Voting, AV. However, in AV the voter has the right to vote for as many candidates as he wishes, but can not indicate any preference among them. (The procedure described in Robert´s Rules of Order is preferential bloc voting, btw)
3. My hunch is that the EC put in the ranking tiebreakers in order to make the voting system preferential, in their understanding of the term. However, I doubt that anyone specializing in the field will agree with them, and many will disagree vehemently. Should anyone wish to challenge the procedure in court (as mfp suggests might happen) then that person will have absolutely no trouble in finding expert witnesses.
4. The preferential input into this vote will in all likelihood be none whatsoever. The chances that two candidates will get the exact same number of X votes from a field of thousands of voters are slim indeed.
5. The result will quite probably be a landslide for either side, unless there is a large proportion of the voters who mix-and-match VP candidates. A landslide will produce cohesion in the elected leadership, but it also shuts out a large proportion of the electorate. Arguments can be made for both, but the USFA should make a conscious, deliberate, decision on whether it values leadership cohesion or proportionality the most.
6. The game plan for both sides will be simple: get their voter base to vote for a triple slate, and at the same time try to woo those voters leaning for the other slate to include one of theirs. This can be done by appeal to proportionality, pity, or whatever.

Examples of outcomes using this system
In the following, I denote VP candidates from one group by A, and from the other group B. It does not matter if A is the NC group or the other way around. The individual candidates are denoted with a number: 1,2, or 3. Let us assume that the election proves to be close one, otherwise nothing interesting comes out of it at all.

Case 1:
51% of the voters cast votes in this order: A1, A2, A3
20% of the voters cast votes in this order: B1, B2, B3
20% of the voters cast votes in this order: B2, B3, B1
9% of the voters cast votes in this order: B3, B1, B2

The tally is like this:
A1: 51% of the vote, elected first based on preferences
A2: 51% of the vote, elected second based on preferences
A3: 51% of the vote, elected third based on preferences
B1: 49% of the vote
B2: 49% of the vote
B3: 49% of the vote

We see that a small difference in voter preferences gives complete landslide with regard to voted officers. It is also interesting to note that A3 gets elected, despite being considered the best by on one, being last on that score.

Case 2:
51% of the voters cast votes in this order: B1, B2, B3
20% of the voters cast votes in this order: A1, A2, A3
20% of the voters cast votes in this order: A2, A3, A1
9% of the voters cast votes in this order: A3, A1, A2

The tally is like this:
B1: 51% of the vote, elected first based on preferences
B2: 51% of the vote, elected second based on preferences
B3: 51% of the vote, elected third based on preferences
A1: 49% of the vote
A2: 49% of the vote
A3: 49% of the vote

When one compares cases 1 and 2, one sees that voter ordering is completely irrelevant to the outcome, provided that all voters vote for triple slates. Voting discipline will not help a minority group - even a large minority group as this one - to elect even one of theirs, A1 in case 2.

Case 3:
Assume that the A candidates see case 2 coming, and that they campaign against that unfavorable outcome. Candidate A1 approaches a B1-leaning voter, and uses the following pitch:

"Yes, I know that you support the B slate, and I acknowledge that there are good points in both candidates B1 and B2. However, where you look at perspective (insert specific reason here), I am actually more similar to B1 than B3 is. That in mind, why not vote for me as your third-hand choice?"

Candidates A2 and A3 do the same with the B2- and B3-leaning voters, respectively. All three A candidates are successful in wooing a small part of the electorate with this pitch. Given that, the votes come out like this:

45% of the voters cast votes in this order: B1, B2, B3
2% of the voters cast votes in this order: B1, B2, A1
2% of the voters cast votes in this order: B2, B3, A2
2% of the voters cast votes in this order: B3, B1, A3
20% of the voters cast votes in this order: A1, A2, A3
20% of the voters cast votes in this order: A2, A3, A1
9% of the voters cast votes in this order: A3, A1, A2

The tally is like this:
B1: 49% of the vote
B2: 49% of the vote
B3: 49% of the vote
A1: 51% of the vote, elected tied first based on preferences
A2: 51% of the vote, elected tied first based on preferences
A3: 51% of the vote, elected third based on preferences

One sees that despite that an absolute majority of the voters thought that the B slate was the overall best, the A slate wins a landslide. Also, the candidate which was considered the best of by far the most voters (B1 got 47% of all 1-hand votes, compared to A1 and A2 with 20% each) failed to get elected.

Whether these are acceptable outcomes of an election, that I leave to the USFA members to decide.


Have a nice time!

Peter Gustafsson

[seak]

Originally Posted by mfp

It's clear that the RRO procedure is designed for such elections.

Not really, the RRO procedure rests on the assumption that the voters will have absolute, linear preferences...such that I prefer A to B, B to C, C to D, and that therefore I prefer A to C and D, B to D, etc. In the case of the USFA VP's, I strongly suspect that many people have an absolute preference between 1,2,3 vs. 4,5,6 but not necessarily between 1,2,3.

Additionally, under RRO each person essentially gets "1 vote" (generalization) even though we're electing 3 people, to do 3 different jobs (even if they're all called VP and the duties aren't explicitly defined by the USFA, the VP's have historically each had their own area and that looks poised to continue under either slate).

Essentially, I think Roberts stinks at this type of system, and isn't really designed for the USFA VP case.

Originally Posted by mfp

The system described is called a "Plurality at Large" form of election. It is not a preferential type of election. It does not elect by majority. Even its very name should provide a clue.

Which system are you referring to? The one you described in Roberts, or the one that the USFA is using?

[seak]

Originally Posted by PeterGustafsson

Lots of intelligent stuff said by Peter

This is why I would not argue with a proportional representation slate vote, with a threshold percentage (say 33%), but given that it's not slate voting, you can't implement it in this election, and slate voting is pretty close to the only way to effectively do proportional representation.

That being said the current USFA voting system, as it relates to the VP's is sub-optimal in many ways. IMHO it would be far better for either a proportional system to be used, or the VP candidates to be lined up against each other, and split into separate posts, such that there is one winner for each post, then the current mishmash.

[mfp]

Originally Posted by seak

Which system are you referring to? The one you described in Roberts, or the one that the USFA is using?

The system the USFA is using. It's basically Plurality-at-Large, a (non preferential) form of bloc voting upon which the Election Committee appears to have written "Preferential" in crayon.

Even if someone thought that a preferential vote wasn't required by the bylaws or was deranged and somehow believed the election committee's system is preferential, then there's still the "majority of votes cast" bylaw.