A few election statistics

[Jon Seattle]

Hi All, this is a basic breakdown by frequency of the votes. I am in a rush this morning, but I will have a bit more to say later.

Code: Select all

Where did the votes go?

Of 16 voters who put DPU first:
	10 put SP as their second choice
	 6 put CSDF as their second choice
	 
Of 15 voters who put SP first:
	12 put CSDF as their second choice
	 3 put DPU as their second choice

Of 12 voters who put CSDF first:
	 6 put DPU as their second choice
	 6 put SP as their second choice
	 
Where did the Borda counts come form?

SP = 46 borda points
	30 from 15 first place votes
	10 from voters who put DPU first, SP second
	 6 from voters who put CSDF first, SP second

CSDF = 42 borda points
	24 from 12 first place votes
	12 from voters who put SP first, CSDF second
	 6 from voters who put DPU first, CSDF second

DPU = 41 borda points
	32 from 16 first place votes
	 6 from voters who put CSDF first, DPU Second
	 3 from voters who put SP first, DPU Second

I suspect the big story here is that CSDF lost some of the people who put it first in past elections, and lost those mainly to SP.

[Sonja Strom]

It is interesting that in the election the party receiving the most 1st choice votes came in 3rd place in the borda count, and the party receiving the most 3rd choice votes came in 1st place in the borda count.
Another way of describing this outcome is:
Most 1st choice votes (DPU) = 3rd place in count
Most 2nd choice votes (CSDF) = 2nd place in count
Most 3rd choice votes (SP) = 1st place in count

The result is posted here: http://forums.slcds.info/viewtopic.php?f=3&t=2282

Does this reflect a vote-counting structure that could continuously arrive at the same results in the future?

[Cindy Ecksol]

It is interesting that in the election the party receiving the most 1st choice votes came in 3rd place in the borda count, and the party receiving the most 3rd choice votes came in 1st place in the borda count.
Another way of describing this outcome is:
Most 1st choice votes (DPU) = 3rd place in count
Most 2nd choice votes (CSDF) = 2nd place in count
Most 3rd choice votes (SP) = 1st place in count

The result is posted here: http://forums.slcds.info/viewtopic.php?f=3&t=2282

Does this reflect a vote-counting structure that could continuously arrive at the same results in the future?

Well, yes....and no. The problem with this kind of analysis is that it's a very limited view of a much more complex process. What you have to do is look first at the total of 1st and 2nd place votes (3rd place votes get 0 points in a 3-way race), and then at the number of first place vs. 2nd place votes.

SP (first in count) 31 total 1st+2nd
CSDF (2nd in count) 29 total 1st+2nd
DPU (3rd in count) 25 total 1st+2nd

So DPU came in 3rd because they had fewer votes that counted than the other two factions, despite the fact that 16 of their 25 votes were 1st place votes. I'll also note that SP was only one first place vote behind DPU, with 15 to DPU's 16. It would have taken a much more definitive first place preference for DPU for that faction to rank higher since they had significantly fewer 2nd place votes than the other two factions.

This isn't really a complete explanation either. I haven't worked out the math, but my intuition tells me that there are probably some mathematical possibilities that might give a result with one faction getting a whole lot of first place votes and placing first even though they had no second place votes at all. But most results will have the faction with the most 1st+2nd votes as the winner I think.

Cindy

[Sonja Strom]

Thanks for the further explanation Cindy.

Maybe my view is a simple one, but I simply want to be sure the choices of the voters are accurately reflected in the outcome of the elections. If they are then they are, but I still have a concern they might possibly not be.

[Jon Seattle]

The election results reflect the constitution's procedures. We have a system that combines Borda counts with proportional representation. (Proportional, that is, to the Borda count totals.) There are some issues with this approach:

1. Unless there are really very widely different voting results, it tends to allocate the same number of seats to each faction. I call this the super-sandpaper problem. Since the number of representatives is not evenly divisible by the number of factions (and point allocation is not greatly influenced by the number of people who vote for a faction), what you really are competing for is just the one extra (or one fewer) representative that is left after the even division.

The current system gives much more influence to a policy position if it is supported in a greater number of parties, irrespective of how many actual citizens support that position. Some citizens get much more of a say than others.

2. Its really impossible to figure out how to assign a weight to second-place votes. How many second place votes equal a first place vote? Who can ever tell. Nevertheless the system says that two second place votes are exactly equal to one first place vote. But there are only three choices so some people may make a second place selection because they have to.

At one point in CDS history two factions, say A and B made an outright deal to put each other in second place. Faction C, because of our system, was forced to give their second place votes to A or B. This strategy can lead to very unfair results.

The CDS ended up with this odd system because of a historic accident -- I understand that it was originally an error in the interpretation of the programming specifications that later got written in to the constitution. Personally, I think it should be changed. I and a few others have been saying that for a long time now.