Free State Project Forum

Archive => Which State? => Topic started by: JasonPSorens on October 24, 2002, 12:12:42 pm

Title: experiment
Post by: JasonPSorens on October 24, 2002, 12:12:42 pm
This is an experiment, the purpose of which may not be immediately apparent to you.  However, I'll let you know once I get an adequate number of responses.  You can post your responses here or email them to me at jason.sorens@yale.edu .

The question: how would rank the candidate states 1-10, how would you assign points to them in a cumulative count vote if you had 100 points to give, and how would you "rate" each state on a scale of 0-100 (not considering other states)?

Here's an example of the format to use (rankings, points, and ratings given here are random):

State             Ranking       Points    Rating
Alaska                 1             29           99
Idaho                  2             20            90
Montana             3             18            87
Wyoming             4             11            76
North Dakota      5              8              69
South Dakota      6              7             68
Vermont              7              5              62
New Hampshire   8             1               55
Maine                   9             1              54
Delaware            10            0               47
Title: Re:experiment
Post by: Eddie_Bradford on October 24, 2002, 01:27:02 pm
Hey Jason I think this is a great experiment!!!!
You know that I'm biased towards Rating in fact here is my rating vote for the the different voting systems!

Ranking   20
Points      51
Rating    100

okay and here is the one you've asked for this is how I would actually vote if we were voting today.

State            Ranking      Points    Rating
New Hampshire  1            75             100
Delaware            2            25              98  
Alaska                3               0              50
Wyoming            4               0              30
Vermont              5              0              20
Idaho                  6              0              10
Montana             7               0                6
Maine                  8              0                 5
South Dakota      9              0                 0
North Dakota      10             0                 0

 
 
 
 

 

Title: Re:experiment
Post by: craft_6 on October 24, 2002, 01:31:33 pm
State  Ranking  Points  Rating
-------  ----------  -------   --------
AK        1           40        90
MT        2           26        85
ND        3           24        80
WY        4             9         65
SD        5             1         50
DL        6             0         45
VT        7             0         40
ID        8             0         35
NH        9             0         30
ME      10            0          25
Title: Re:experiment
Post by: Stumpy on October 24, 2002, 03:23:40 pm
State             Ranking       Points    Rating
Montana             1             30            85
North Dakota      2            25             75
Idaho                  3            20            70
Wyoming            4            18             60
New Hampshire   5            5              45
South Dakota      6             2             40
Maine                   7            0             20
Vermont              8             0             20
Alaska                 9             0             15
Delaware            10           0              0
Title: Re:experiment
Post by: wolf_tracker on October 24, 2002, 04:02:17 pm
Alaska                1              40            90
Montana             2              20            75
Wyoming            3              10            60
Vermont             4               10            50
Idaho                  5              10            50
New Hampshire  6               10            40
Maine                  7                0               5
South Dakota      8                0             0
North Dakota      9                0             0
Delaware            10              0             0
Title: Re:experiment
Post by: varrin on October 24, 2002, 05:09:05 pm
Idaho                   1               50         90
Montana              2               30          65
Delaware              3              20         40
New Hampshire      4               0          11
Maine                    5              0          9
Vermont                6              0          7
Alaska                    7             0          5
South Dakota        8              0           0
Wyominig               9             0           0
North Dakota         10           0           0

In my analysis, inadequate air service is a veto (SD, WY, and ND fail the test).  Cold and snow are heavily weighted.  Western, spacious, dry, and liberty oriented are important.  
Title: Re:experiment
Post by: cathleeninsc on October 24, 2002, 07:23:48 pm
Montana        1       20     85
N. Dakota      2       15     75
Alaska            3       15     75
New Hampshire 4    15    75
Maine             5       12     73
Idaho             6       10     70
Delaware       7        8      50
Vermont         8        5       40
Wyoming        9        0      20
S. Dakota       10      0      20

Cathleen in SC
Title: Re:experiment
Post by: Condon on October 24, 2002, 07:56:07 pm
North Dakota            1          30            85
Wyoming                   2          20            80
Montana                    3          15            80
South Dakota            4            9             65
Delaware                   5            9             60
Vermont                     6            6             65
Alaska                        7             5            60
Idaho                         8            3              65
Maine                         9            2              60
New Hampshire         10           1              60
Title: Re:experiment
Post by: phylinidaho on October 24, 2002, 08:12:58 pm
State            Ranking      Points    Rating
Montana           1            75           95
North Dakota     2            10            85
Idaho               3              8           80
Maine               4              6              60
New Hampshire   5             1            40
South Dakota       6              0            40
Wyoming            7              0             20
Vermont               8              0                5
Alaska                 9              0                0
Delaware          10            0                0
Title: Re:experiment
Post by: Dakotabound on October 24, 2002, 10:32:11 pm
State            Ranking      Points    Rating
North Dakota     1              50         100
Montana             2             10           80
Wyoming            3             10           75
South Dakota     4               5           65
Alaska                5               5           50
Idaho                 6               5           50
Vermont             7               5           45
New Hampshire  8              5            45
Maine                  9              3            30
Delaware            10             2            30
Title: Re:experiment
Post by: Robert H. on October 25, 2002, 01:10:38 am
State            Ranking      Points    Rating
Montana            1             30        100
North Dakota     2             20          90
Alaska                3            10          75
Maine                 4             10          65
Wyoming            5             10          60
Idaho                 6              10         40
South Dakota     7              5           30
New Hampshire  8              5            25
Vermont              9              0            0
Delaware            10            0            0
Title: Re:experiment
Post by: Michelle on October 25, 2002, 11:36:36 am
State            Ranking      Points    Rating
New Hampshire  1            75            100
Maine                  2            10              75  
Idaho                  3              5              70
Montana             4              5               70
Alaska                 5              5               60
Wyoming             6              0               50
Vermont              7              0              50
South Dakota      8              0                40
North Dakota      9               0               40
Delaware            10            0                30
Title: Re:experiment
Post by: mlilback on October 25, 2002, 06:28:04 pm

State         Ranking  Points  Rating
New Hampshire    1       50      90
Delaware         2       20      75
Maine            3       15      80
Vermont          4       10      80
Montana          5        5      60
Idaho            6        0      50
Alaska           7        0      30
North Dakota     8        0      15
South Dakota     9        0      10
Wyoming         10        0      10

Title: Re:experiment
Post by: JasonPSorens on October 25, 2002, 07:52:57 pm
All right, 12 responses are sufficient for me to do the analysis, so here it is...

The purpose of this experiment was twofold: 1) to see whether the presentation of the vote increased the "rationality" of cumulative count ballots; 2) to see whether cumulative count stacked up well against the two alternative methods most often promoted.

I'll deal with the second first.  It appears that cumulative count performs much better than its critics expect.  

Instant runoff voting (also called "alternative vote" and "single transferable vote") uses the rankings in the first column.  It works by eliminating choices with the least amount of first-place votes and then allocating second preferences among those candidates left.  Then the process is repeated, until a winner is found.  When this done to the above ballots, New Hampshire and Montana are the two left standing at the end, and Montana wins handily, 8-4.

Under cumulative count, New Hampshire and Montana are again the top two, and Montana wins, 281-257.  The third place choice is North Dakota with 179; under IRV North Dakota and Alaska are tied for third.  Alaska is a close fifth under cumulative count, getting beaten by Idaho 121-120.  So the cumulative count result closely matches the IRV result.

Under approval vote, which simply adds up the absolute ratings in the third column, Montana and New Hampshire are again the top two candidates, and again Montana wins, 860-731.  However, Idaho sneaks all the way into third place with 715, not far behind New Hampshire.  North Dakota and Alaska are fourth and fifth, with 685 and 625 respectively.

Apart from the anomaly of Idaho, cumulative count and approval vote are very, very similar.  The Pearson r correlation between the two methods for all 10 candidate states was 0.89, which is quite good: 1.0 is a perfect correlation, and 0.0 means no correlation.

My conclusion: the hullabuloo over voting method is much ado about nothing.  All 3 methods yield essentially the same result.  (By the way, pure plurality voting yields a tie between Montana and New Hampshire for first-place votes.  Plurality vote is perhaps the worst voting method we could possibly choose. ;))

When it comes to rationality of the cumulative count ballots, however, there are some real puzzles, the solutions to which may have important implications for how the Research Committee designs the ultimate state ballot, and for how we go about educating people about cumulative count.

The main potential problem with just about any voting system is strategic voting.  Strategic voting is the phenomenon whereby people refrain from voting for candidates that they think won't win, and instead vote for the lesser of the evils that they think can win.

[In what follows, please recognize that I am not criticizing how anyone voted.  Your ballot is inviolate, and neither I nor anyone else has the right to criticize the way you choose to cast it.  However, paradoxes in the way certain people voted are instructive for us in learning how to describe cumulative count and how to design the "real" ballot.]

Now then...the most obvious case of strategic voting was Eddie's ballot.  He gave all his CC votes to 2 states, and 50 more CC votes to New Hampshire than to Delaware, even though Delaware for him was almost as good a candidate as New Hampshire (rating 98 versus 100).  Alaska, Idaho, Vermont, Montana, Wyoming, and Maine did not get any votes from him even though he considered them to have some value as candidates (ratings>0).  Phyllis' ballot was also apparently strategic: even though North Dakota and Idaho were almost as good for her as Montana, she gave 65 more CC points to Montana than to either of them.  Two very instructive cases!  Still, Eddie's and Phyllis' strategic voting was a lot less than what people have been doing on the practice poll on the State Data page.  A few days ago I examined the individual votes in the practice poll - fully 42% of them gave all 10 of their points to a single candidate!  This, of course, defeats the whole point of cumulative count.  I'm heartened to see that no one here gave all 100 votes to a single state.  There are two potential reasons for this: having people rank states made them think about candidates that were good but not their first choice, and giving people 100 points rather than 10 made them feel as if they could spread them out more.  If this is true, we should do the same thing on the final ballot.  Matt Cheselka is designing a new practice poll for the website to determine this more scientifically, over a bigger sample.  But the results from this experiment are encouraging.

There were some minor paradoxes in almost everyone's ballot.  Mike Lilback gave 15 CC points to Maine and 10 to Vermont even though they are equally good candidates according to his ratings.  Michelle gave 5 points to 3 states even though 1 fell below the others in terms of rating.  Robert gives 10 points to a state he rates 75 and 10 points to a state he rates 40 - wow!  There seems to be a strong tendency to give "round numbers" in the CC vote (and in the ratings for that matter), even when people are clearly not indifferent among two or more candidates they may "round them both off" to the nearest 5.

Something everyone except Tim and Kim did was to give 0 points to more than 1 state.  Theoretically, this shouldn't happen: you should want to give at least 1 point to your #9 choice in order to help it out, even if just a very little bit, over your #10 choice.  (Tim's and Kim's ballots have their own paradox, in that they gave votes to their worst choices! ;))

So I'm interested in hearing from people about some of these paradoxes, especially why you didn't give any CC votes to states that were low on your list but still rated much higher than your worst choice.  Again, I'm not criticising the fact that you did this - far from it.  But getting to know people's logic will help us figure out whether we need to put more explicit instructions in the final ballot, or undertake more education about how best to cast a cumulative count ballot given any array of preferences.
Title: Re:experiment
Post by: Jim1 on October 25, 2002, 08:43:28 pm
A few days ago I examined the individual votes in the practice poll - fully 42% of them gave all 10 of their points to a single candidate!  This, of course, defeats the whole point of cumulative count.  I'm heartened to see that no one here gave all 100 votes to a single state.  There are two potential reasons for this: having people rank states made them think about candidates that were good but not their first choice, and giving people 100 points rather than 10 made them feel as if they could spread them out more.


I think the real reason for less "strategic voting" was that the vote was public (not a secret ballot). Nobody wants to look like a cheater.
Title: Re:experiment
Post by: phylinidaho on October 25, 2002, 08:55:01 pm

Phyllis' ballot was also apparently strategic: even though North Dakota and Idaho were almost as good for her as Montana, she gave 65 more CC points to Montana than to either of them.  


Perhaps my voting was strategic, although I didn't think of it in that context. The reason I gave 75 points to Montana was that I feel that strongly that it is the state most likely to be successful as a free state. Having given the 75 points, I didn't have much left to allocate to other states.  I rated North Dakota and Idaho quite high on the other list  as an indication of how I would feel if they were chosen, as compared with the other 7 states.
Title: Re:experiment
Post by: JasonPSorens on October 25, 2002, 09:09:04 pm


I think the real reason for less "strategic voting" was that the vote was public (not a secret ballot). Nobody wants to look like a cheater.



That is a possibility that concerns me.  Hopefully the cumulative count practice vote on the website will show a change.  The thing is, strategic voting would probably hurt you in the final vote, because it really isn't clear which state will be the frontrunner by then.  NH was an early frontrunner, but now Montana appears to be the top candidate... Tim has made a powerful case for Wyoming, and my own analysis indicates that Wyoming, Delaware, and Alaska are the top 3.  So it's all up in the air at this point!
Title: Re:experiment
Post by: Robert H. on October 25, 2002, 10:34:47 pm
I'll try to shed some light on the Hawes paradox... ;D

One problem may be that I'm just very new to the whole cummulative count scenario and just haven't figured out how to get it right yet.  For instance, your suggestion about giving at least one point to your #9 position state to simply give it an edge over #10 is well taken.

I gave Alaska, Maine, Wyoming, and Idaho all ten points even though they rank from 75 to 40 on my rating scale due to what I considered our overall chances of success in those states.  Idaho ended up getting the same points as Alaska, even though I considered Alaska a better choice, because I wanted to arrange my votes in such a way as to substantially cut them off from what I considered the worst choices.

After Montana and North Dakota, I think our chances of overall success decline sharply, plateau, and then decline sharply again.  The block of states that all got "10's" did so because I consider them to be less viable candidates (in their rated order), but I consider our overall chances of success as approximately the same in them for various reasons.

I dont' know if that helps explain any method in my madness.   ;D
Title: Re:experiment
Post by: Eddie_Bradford on October 26, 2002, 06:08:50 am
Hey Jason, I agree it will be pretty close no matter which voting system is used.  So in reality it doesn't matter all that much.  But just because it didn't make a big difference here doesn't mean it won't later.

#1  The main problem and reason people don't give points to their #9 choice is because it hurts them to do so.  Look at Robert's he gave his top choice only 30 points and he probably filled his ballot out honestly.  And let's say he like Montana every bit as much as I like New Hampshire.  Well if I give 100 points to New Hampshire (which I probably will do in the final vote) my vote will be 3 times as powerful as his in the end.  He wanted to say "HEY I REALLY LIKE THIS ONE" but in reality because he only gave 30 points his vote didn't really help Montana much.  Okay so let's say we say "Gee Eddie don't do that you shouldn't give all you your points to one state" but that's like if we were doing the rating system and we say "Gee don't give your #1 choice a ranking of 100 unless you REALLY think it's the absolutely perfect place" well that doesn't make sense because you want it to win!  So basically if you vote honestly your favorite choice get hurt.  It's important to note that my choices were responcible for 30% of New Hampshire's total CC points and Robert was only responcilbe for 10% of Montana's points.  In fact if I had voted all 100 point for New Hampshire (which I probably will) it would have won!  Despite the unfortunate fact that Montana seemed much more popular than NH when looked at from any other angle.  

#2 There is no way to "Anti-vote" in CC voting.  Say there are 6 choices I want NH really bad and say you don't care except that you REALLY DON'T want NH.  The correct result should be a tie because our opinions are equally strong.  But instead NH get 100 points and all the other states get 20 points.  

#3 Finnaly I think people round alot because it is difficult to go back and add up all of your points and make sure it's 100 "...no wait it's only 85 so I can but maybe 5 more here 3 more there 2 and 1 and... gee I guess 3 more to the top and second one no wait now I'm at 102 so I take one off the second and third or fourth choice?..."  etc.
(speaking of which whatever we use to count the votes will have to independently verify that only 100 points were used hopefully that won't be much of an issue though since there is already the 10 point poll that does that)

Anyway that my speal again, as you can see I prefer the ranking system.  I think Idaho sneaking in 3rd is an indication that this way of voting is better at finding 'compromise' states (ie states with a broad base of moderately high opinions) than the CC voting is.  If we don't switch systems I still think we should REALLY consider letting people give state negative points so the issue in #2 can be resolved.

-Eddie
Title: Re:experiment
Post by: Eddie_Bradford on October 26, 2002, 06:27:08 am
The problem is strategic voting, but it is IMPOSSIBLE to strategically vote in the rating system.  So what is the downside of the ranking system I just don't see one.  There are plenty of people who appoligise for the CC voting but I just want to know what is the downside of ranking?  Please make a post.
Thanks,
      Eddie
Title: Re:experiment
Post by: wolf_tracker on October 26, 2002, 07:41:42 am
Jason,

You asked why ppl gave states no votes.

I gave no votes to states that I do not want to live
in.  Why hurt the states I want to live in by wasting
votes states I dont like.  If this was a final vote, my
first thought is that ppl will give 100 percent of their
votes to the state they want the most.  In your shaking
up of the number it appears Alaska was one place down
by 1 vote.  So if I has given Alaska 45 instead of 40 then
Alaska would have been higher up by one position.

Ranking was just an order after the first 4 or 5 position.
After the first 3 or 4 positions I would not care because
I did not like that places in the first place. Positon 8 - 10
could be in any order.  I didnt care.  To be they are equilly
bad.  You asked to order so I put then in order.

I like the idea of negative votes but it would not work
because person one loves alaska and gives then high
numbers, person 2 hates alaska and gives them mag
neg numbers, kind of balancing it out.

The state that gets choisen would then be the one
that brings out the least emotions, because the love
and hate would kind of balance it out.

I would probably give my most favorite votes to Alaska
and Delware would be my most disliked.  Then someone
will knock my Alaska vote out with their dont like vote
and I would knock their Delware vote out with my
dont like vote.

Thus we are down to the least liked and dont liked and
I feel a large group of ppl would not move to the least of
the good and bad.

my .02
Title: Re:experiment
Post by: JasonPSorens on October 26, 2002, 09:25:44 am
Eddie - It's true that voting honestly may reduce the probability of your #1 choice winning.  But voting honestly will increase the probability of your #2 or #3 or #4 choice winning if your #1 choice doesn't!  In the end, I think almost all the states will have an equal chance of winning.  People have talked about N.H. and Montana a lot, but both are sinking in popular perception as new data and new analyses come in.  N.D. appears to be slightly superior to MT, and WY may be as well.  NH because of its population and the new election data coming in will probably fall below VT and DE in my quantitative analyses.  Since voting dishonestly only works if you know your top choice is also one of the top two choices of the group, I don't think it will make sense to vote dishonestly when the final ballot comes out.
Title: Re:experiment
Post by: Eddie_Bradford on October 29, 2002, 08:15:25 pm
Jason I dissagree :)
First off I don't think we should call this voting "disshonestly" because I really don't think it is.  I voted the way I wanted to.  IT's not disshonest to say vote for Bush instead of Brown because you really don't want Gore.  People make these decisions and it is impossible to vote "disshonestly" unless you are voting randomly.

This is my "proof" of the sillyness of an "Honest" CC vote.  I am going to translate this into an economic issue so we Libertarians will feel at home.  

You have $100 Million dollars to spend on a presidential campaign.  This amount can SIGNIFICANTLY influence the outcome of the campaign and assume all 4 candidate had at least an okay chance of winning (Bush, Gore, Brown, Nader).  Negative ads are not allowed.  
Would you spend $100 Million on Harry Brown?  Or would you spend $50 Million on Brown, $35 Million on Bush and $15 Million on Gore?
Seriously would you spend money on either Gore or Nader????  Jason thinks you should! :)
Personally I think I might spend $50 million on Brown and $50 Million on negative campaigns against Gore but negative ads are NOT allowed.
The point is limited resources and the $15 million you spend on Gore could be helping Brown.

I hope this helps people to put this in perspective.
-Eddie
Title: Re:experiment
Post by: JasonPSorens on October 29, 2002, 09:54:30 pm
But spending on Bush is the same as spending negative ads on Gore and Nader!  Also, the example doesn't work quite as well, because for us Bush, Gore, and Nader are all virtually just as bad, while that situation doesn't apply to the states.  If you think that all states except 2 are terrible, worthless choices, then certainly don't give points to them.  But if some of them are pretty good choices, just not as good as your top picks, then they should receive points.

Allowing negative points in CC wouldn't make sense - it would allow infinite negative and infinite positive: if I give 50 million and 100 votes to one state and negative 50 million to another state, that would fit within the rules of CC, if negative votes were allowed.
Title: Re:experiment
Post by: Eddie_Bradford on October 30, 2002, 02:22:05 am
Quote
But spending on Bush is the same as spending negative ads on Gore and Nader!

I assume you mean Spending on one candidate is the same as spending negative ads on all the other candidates.  This is true except it costs alot more!  Spending $50 million on someone is like spend $50 million in negative ads on EACH of the 3 other candidates!  So it's 50 M verses 150 M  Just like in CC voting.  Say you REALLY don't want New Hampshire you should be able to spend -10 on it.  Instead you have to put like 2 point in each of the 5 other choices.  Again back to my other example I love NH you hate it I vote 10 points for it and you vote 2 point for each of the 5 other choices NH wins in a landslide 10 points to 2 points.  We both love/hate it equally but I win because of the voting system bias.  Just like our system now is biased.

Quote
Also, the example doesn't work quite as well, because for us Bush, Gore, and Nader are all virtually just as bad
This is true but you also use "Virtually" and according to your logic you should still spend money on the one you like even slightly better.  The reason most people wouldn't is because they know that the money can be spent elsewhere more productivly.

Okay still trying to get my point across...
You are going to win a car.  It will either be a pinto, a ford, or the James Bond Rolls Royce.  There is an equal chance of any three.  You are allowed to spend 100 points towards any of them that will improve the probablity of wining that type.  Everyone would spend all 100 poins of it on the Rolls Royce.  But according to the logic used here we should spend some on the ford because it is better than the pinto.  We say "no dood!  Improving the probablility of the ford is not worth the sacrafice being made by not spending the points on the Rolls Royce"

Quote
Allowing negative points in CC wouldn't make sense - it would allow infinite negative and infinite positive: if I give 50 million and 100 votes to one state and negative 50 million to another state, that would fit within the rules of CC, if negative votes were allowed.

LOL! What the heck Jason?  You're smoking the crack on this one! :) Obviously Your total magnitude cannot exceed 100.   By this logic I could speed down the highway at 200 mph by driving backwards!  When I get pulled over I would say "but officer I was going -200 miles per hour which is clearly less than the posted speed limit!"
-Eddie
Title: Re:experiment
Post by: JasonPSorens on October 30, 2002, 10:21:29 am

I assume you mean Spending on one candidate is the same as spending negative ads on all the other candidates.  This is true except it costs alot more!  Spending $50 million on someone is like spend $50 million in negative ads on EACH of the 3 other candidates!  So it's 50 M verses 150 M  Just like in CC voting.  Say you REALLY don't want New Hampshire you should be able to spend -10 on it.  Instead you have to put like 2 point in each of the 5 other choices.  Again back to my other example I love NH you hate it I vote 10 points for it and you vote 2 point for each of the 5 other choices NH wins in a landslide 10 points to 2 points.  We both love/hate it equally but I win because of the voting system bias.  Just like our system now is biased.


But this happens only if the other person doesn't particularly like any other candidate.  If he sort of likes Montana and gives it 6 points but hates New Hampshire, and the other person hates Montana but really loves New Hampshire and gives it 10 points, then NH wins, but it's hard to see that as unfair.

Quote

Okay still trying to get my point across...
You are going to win a car.  It will either be a pinto, a ford, or the James Bond Rolls Royce.  There is an equal chance of any three.  You are allowed to spend 100 points towards any of them that will improve the probablity of wining that type.  Everyone would spend all 100 poins of it on the Rolls Royce.  But according to the logic used here we should spend some on the ford because it is better than the pinto.  We say "no dood!  Improving the probablility of the ford is not worth the sacrafice being made by not spending the points on the Rolls Royce"


Well, I disagree that state choice is that way.  Almost all the states are good candidates; there certainly is no Rolls Royce-Ford relationship among them.

Quote

LOL! What the heck Jason?  You're smoking the crack on this one! :) Obviously Your total magnitude cannot exceed 100.  


So you mean people would be allowed to distribute 100 positive votes and -100 negative votes?  But that's what we have now!  All you're doing is changing the scale.  Let's say we use your system, and you assign points to five candidates this way:

State 1               +40 points
State 2               +35 points
State 3               +25 points
State 4                -15 points
State 5                -85 points

You can send the same message if negative points are not allowed:

State 1               +30 points
State 2               +28 points
State 3               +25 points
State 4               +17 pointsState 5                0 points

Allowing negative points just spreads the values out more, but that's true for everyone's vote, so the difference in systems doesn't affect the result systematically.  There may be some circumstances where different systems would yield a different result, but it wouldn't be clear which one was superior in that case.
Title: Re:experiment
Post by: Eddie_Bradford on October 30, 2002, 05:30:11 pm
No no no.  Jason you can spend only 100 points whether they are positive or negative.  So you can spend 100 points on New Hampshire OR you can spend -100 points on Montana OR you can spend 50 points on New Hampshire and -50 points on Montana.  The sum of the magnitudes (ie absolute value) cannot exceed 100.  You can spend +90 points and -10 points OR +40 and -60 point but after you take the absolute value of each number and then sum them together the total cannot exceed 100.
And yes this DOES matter.  I have already given the most obvious situation ubove.  I put 100 points in for New Hampshire, Bob puts in -100 points for New Hampshire and then it would be tied as it should be.  This is instead of spreading out 20 points for all states except NH.  You admit that this is a problem in this quote

Quote
But this happens only if the other person doesn't particularly like any other candidate.


I think it should be okay if somene holds this opinion.  But Jason even that is beside the point because I have choosen an extreme example to more clearly illustrate the problem.  This problem will exist in many people's vote to a less degree.  Okay watch as I dassle you with yet another numerical example.
If I vote 50 for New Hampshire and -50 for Montana you say that I can do the same thing with  the current system.  This is NOT so the Mathamatical equivalent in non-negative numbers is this: 4 state choices
100 for New Hampshire   50 for state A,B   0 for Montana
This maintains the 100 point preferance of NH over Montana and also maintains the 50 point spread of NH over the other states.  In order to do this I need 200 points which I don't have!!!!  
The mainly fallacy I see here is that somehow in the ranking system people think that when you give 2 candidates a score of 100 and the rest 0 you are somehow "voting twice" the most obvious example of this was when Elizabeth said "But Eddie if we do your system then everyone would just give each state a score of 100!" what we all need to realise is that if I gave each state a vote of 100 then I have for all intents and purposes not voted at all because I have not affected the outcome.  We cannot have any kind of meaningful discussion about voting until everyone realises this fact.
Title: Re:experiment
Post by: JasonPSorens on October 30, 2002, 05:40:17 pm
Well, those are all valid points, and that's why I did this experiment with the rating method, to see whether in the real world it does differ from cumulative count - and apparently, the results aren't that much different.
Title: Re:experiment
Post by: Eddie_Bradford on October 30, 2002, 06:36:04 pm
No I agree Jason and I hope I didn't say anything that sounded harsh.  I realise that there are many voting system that work very well and in all likelyhood the result will turn out the same.  Again sorry if I sounded harsh I didn't mean to.  Somtimes the points I make are worded poorly and confusing.  Anyway I'm happy just as long as you understand the points I was trying to make.  Thanks for doing this experiment.
-Eddie
Title: Re:experiment
Post by: craft_6 on October 31, 2002, 10:14:40 am
The best way to produce an "honest" vote (one where people vote for what they want, rather than against what others want) would be to immediately stop publishing the practice vote results.  

Someone suggested doing this after membership reaches 4,000, but it may be better to stop now, before the preferences of the next 3,000 members are known.  This would leave an element of mystery in the final decision, and help ensure that all 10 states are analyzed objectively.

The message boards might give an indication of the way people are leaning, of course, but there is at least some level of doubt that those who post frequently are representative of the entire membership.  I suspect that many of those who post frequently are the most committed to their choices, while those who merely lurk are more undecided.

Libertarians often lament mainstream voters being herded into their decisions on Election Day by the media releasing poll results showing only two viable candidates.  We shouldn't follow the same practice ourselves.
Title: Re:experiment
Post by: JasonPSorens on October 31, 2002, 10:33:59 am
Yeah, we're going to remove the practice poll ASAP.  Matt Cheselka is working on a new practice poll, which we will test for several weeks to see if strategic voting is lessened, and then it's gone.
Title: Re:experiment
Post by: craft_6 on October 31, 2002, 10:42:36 am
Why I gave 0 points to more than one state, even though I gave all 10 states non-zero ratings:  

I considered the rating as a rough approximation of the FSP's chances of success.  The states which I rated below 50 are states I wouldn't recommend selecting, due to high voting populations or socialist tendencies.  They would not be impossible to succeed in, just very difficult.  Why give any points to a state I consider a poor choice, even if another state is worse?  
Title: Re:experiment
Post by: JasonPSorens on October 31, 2002, 10:52:31 am

Why I gave 0 points to more than one state, even though I gave all 10 states non-zero ratings:  

I considered the rating as a rough approximation of the FSP's chances of success.  The states which I rated below 50 are states I wouldn't recommend selecting, due to high voting populations or socialist tendencies.  They would not be impossible to succeed in, just very difficult.  Why give any points to a state I consider a poor choice, even if another state is worse?  


I can understand that logic, and that's basically what Eddie and I were discussing.  If people are nearly indifferent among some choices, I don't see it as a problem that they give them all the same number of points, and I don't think it will affect the final vote adversely.
Title: Re:experiment
Post by: Eddie_Bradford on October 31, 2002, 10:54:34 pm
Quote

The best way to produce an "honest" vote (one where people vote for what they want, rather than against what others want) would be to immediately stop publishing the practice vote results. 


I don't want to sound like a broken record but again I think this is attacking the symptom and not the cause.  I as always will use an example to illustrate my point.  

1. I have NO knowledge of what anyone else is going to vote.

2. So as far as I can tell there is an equal chance of each of say 5 states to win and I have 10 points to spend.

3.  I like New Hampshire best and Delaware 2nd.

4.Say each point I spend improves the chances of that state winning by 1%

5. if New Hampshire wins I am very happy!  I am as happy as if you gave me $100.  If Delaware wins then I am as happy as if you gave me $80

6.Let's caclulate my the utility of voting 2 different ways.
10 points for NH
9 points for NH and 1 point for Delaware.

by spending a point it basically gives me 1% of the value of my happiness if that state is chosen.
10 point for NH = 10% improved chance NH is selected
$100 * .10 = $10 of value
9 points for NH 1 point for DE
$100 * .09 + $80 * .01  =  $9.80

-Because voting for a candidate you like less than your favorite actually takes votes away from your favorite the economically optimal solution is to vote all of your points to your favorite.  To sarafice even one vote for your favorite to give to your second favorite is not optimal.  Actually the BEST real world expample and exact analogy is if you went to the poll this Nov. except you would be allowed to vote 10 times instead of once.  Are you going to start to vote for demorcats or repulicans on a few of these votes?  No you aren't because obviously you should vote for your favorite choice especially if you know they all have an equal chance of winning.  This is an exact analogy.  If you dissagree please tell me how it is not.
-Eddie
Title: Re:experiment
Post by: Robert H. on November 01, 2002, 10:24:13 am

-Because voting for a candidate you like less than your favorite actually takes votes away from your favorite the economically optimal solution is to vote all of your points to your favorite.  To sarafice even one vote for your favorite to give to your second favorite is not optimal.  


It sounds like you'd be taking a rather large risk here.  If you bet everything on one state, and that state doesn't make the cut, haven't you effectively deprived yourself of a voice in which of the other states might be the next best choice?
Title: Re:experiment
Post by: JasonPSorens on November 01, 2002, 11:16:25 am

by spending a point it basically gives me 1% of the value of my happiness if that state is chosen.


This only works if people are not risk-averse.  However, I think when it comes to a decision like a move people are definitely risk-averse: they want to make sure their worst choices don't win!  Since you don't know the probability of any state's winning, it's best to spread your votes around a little bit if you are risk-averse, in order to prevent one of your worst choices from winning.
Title: Re:experiment
Post by: Eddie_Bradford on November 01, 2002, 02:51:07 pm
That is a good point Jason and I almost talked about it in my previous post but my posts tend to get really long and go over too many different points.  

First by bringing up that this is not optimal if you are risk adverse to me it sounds like you are conceeding that it you get the best expected value when you put all of your votes towards your favorite.  Correct me if I'm misstaken.  

Next you say that putting all of you eggs in one basket is risky and that you certainly don't want you worst choice to win.  This might be true.  People are most risk averse when there is a potentially large negative result that they want to avoid.  In this case it would be if you really didn't like one state.  So you do as you say and spread out your votes with the goal of avoiding your worst choice.  But this is just a crude appoximation and much less effective way of voting against one state.  Say for 5 states you like one and hate another.  So you vote  4, 2, 2, 2, 0.  If you were allowed to cast negative votes then this would be the same vote as  2, 0, 0, 0, -2.  But you still have 6 more points you can invest so you vote would be 5, 0, 0, 0, -5.  So if you are worried about risk averse people then you should deffinately allow at least negative votes so their voice can be heard as loud as someone who strongly favors one choice.
Point is that still CC voting favors people who put all of their votes into one candidate.
Essentially CC voting is the EXACT same voting system we use now in this country except that instead of a boolean choice you have a 10 point discrete choice.  It's like if I was allowed to give 1/2 vote to bush and 1/2 a vote to Brown.  The RANKING system is the exact same system as approval voting except instead of a boolean choice you have a 100 point discrete distribution.  Doing this allows you to basically "strongly" approve one guy over another guy you just "approve" of.
Title: Re:experiment
Post by: JasonPSorens on November 01, 2002, 03:18:56 pm

First by bringing up that this is not optimal if you are risk adverse to me it sounds like you are conceeding that it you get the best expected value when you put all of your votes towards your favorite.  Correct me if I'm misstaken.  


That appears to be correct, with the crucial assumption that your vote affects the probability that the state will be chosen to an extent that you can estimate.  Your example puts a known probability value on the effect of your vote.  If there were complete uncertainty (probability values were unknowable), then perhaps your analysis would change there as well.  I haven't figured that out yet.  Under complete uncertainty, expected utility analysis breaks down.

Quote

that they want to avoid.  In this case it would be if you really didn't like one state.  So you do as you say and spread out your votes with the goal of avoiding your worst choice.  But this is just a crude appoximation and much less effective way of voting against one state.  


But what if there are, say, four or five states that you want to vote against?  Then giving a lot of points to your favorite(s) and some to the next few good candidates seems rational.

Quote

Say for 5 states you like one and hate another.  So you vote  4, 2, 2, 2, 0.  If you were allowed to cast negative votes then this would be the same vote as  2, 0, 0, 0, -2.  But you still have 6 more points you can invest so you vote would be 5, 0, 0, 0, -5.  So if you are worried about risk averse people then you should deffinately allow at least negative votes so their voice can be heard as loud as someone who strongly favors one choice.


I'm afraid allowing negative votes would increase people's confusion and complicate tabulation.  Most people don't even know what "absolute value" means.

Under approval voting, assuming expected utilities and risk neutrality, wouldn't you still have an incentive to rate one state 100 and the others zero?
Title: Re:experiment
Post by: Charley on November 01, 2002, 05:52:21 pm
Maybe I'm dense but why not have a straight ranking of the remaining candidates?  Lowest vote total wins.  
Title: Re:experiment
Post by: JT on November 01, 2002, 07:36:26 pm
Is there anyway to narrow down the choices before the final vote?
Title: Re:experiment
Post by: Eddie_Bradford on November 01, 2002, 09:56:24 pm
Quote
That appears to be correct, with the crucial assumption that your vote affects the probability that the state will be chosen to an extent that you can estimate.  Your example puts a known probability value on the effect of your vote.  If there were complete uncertainty (probability values were unknowable), then perhaps your analysis would change there as well.  I haven't figured that out yet.  Under complete uncertainty, expected utility analysis breaks down.

I will analyse this in a post in the near furture.  My posts always end up being long. :)

Quote
But what if there are, say, four or five states that you want to vote against?  Then giving a lot of points to your favorite(s) and some to the next few good candidates seems rational.

Yes indeed!  But bringing up one specific example where CC voting does work well doesn't prove that it is a good system.  You have to dissprove my assertions of weakness to show that it always works well.  The voting system used in the US now works well if you like only one candidate and you really don't like the other ones.  Also even though you like say 2 candidates your vote is still diluted by only getting to put 5 points towards each as opposed to 10 points each could get in the rating system.

Quote
I'm afraid allowing negative votes would increase people's confusion and complicate tabulation.  Most people don't even know what "absolute value" means.

Sadly I agree the 'negative' voting idea was supposed to be a compromise idea by me where we could still use CC voting but fix some of it's flaws.

Quote
Under approval voting, assuming expected utilities and risk neutrality, wouldn't you still have an incentive to rate one state 100 and the others zero?

Yes!  the problem isn't voting 100 to 0 the problem is that in CC voting when you vote 100 you then give all the other states a 0 even the ones you like because you have spent your total.  In approval voting with a 100 point scale you could give your favorite state 100 your second state 80 third 50 and the rest 0.  There is no way to take some of the points from your second choice and then improve your first choice i.e. there is no incentive to vote 100, 0, 0, 0, 0 because you can still spend point on whichever ones you want.

It seems like you agree that there might some issues with CC voting although we dissagree about how severe they might be.  The real way to win me over is to convince me that the rating system is bad.  Here I'll start that debate for us in my next post.
-Eddie
Title: Re:experiment
Post by: JasonPSorens on November 01, 2002, 09:58:35 pm
Charley - not sure what you mean.  Do you mean assigning 1 point to first-place vote, 2 to second-place vote, etc.?  That is one possible voting system.  Usually, though, you'd want to "reward" the top few choices a little more, so for example: 0 points to 1st choice, 3 points to 2nd choice, 5 points to 3rd choice, 7 points to 4th choice, 8 points to 5th choice, and so on...lowest point total then winning.  But with a ranking, instant-runoff voting actually has some desirable properties that the above arrangement does not - for one thing, the points you assign to the rankings are somewhat arbitrary.

But I'm afraid the discussion is mostly academic because changing cumulative count is extremely unlikely, unless a massive consensus in favor of an alternative develops. :(

JT - the Research Committee considered narrowing down the choices further, but there was an outcry against that as "undemocratic" and so on, which is a legitimate concern.  So we opted for an absolute population cutoff.  Eliminating any further states would be fraught with controversy.
Title: Re:experiment
Post by: Eddie_Bradford on November 01, 2002, 10:13:37 pm
Jason:  "Eddie the rating system is sooo crappy!"

Eddie: "Oh really?  What makes you think that?"

Jason: "Well Eddie in the rating system you are allowed to really like 2 states and give them both a score of 100"

Eddie: "So what's wrong with that?"

Jason: "Well if you really like only 1 state then shouldn't you be able to have a stronger positive affect on that one state than someone who like 2 different states?  Because the best score you can give is 100 while the other guy gave 2 100s."

Eddie: "No I dissagree I think just because you like 2 states doesn't mean you can't  like both as much as the person who likes just one 1."

Third party: "I dunno Eddie, I think that the rating system causes voting inflation because many people would probably give all their prefered states scores of 100.  In the other sytem you couldn't give all the state the highest scores."

Eddie: "That fine if you give more than one state a score of 100 but you must realise that if you really like your first choice and your 3rd choice is just okay then giving both a score of 100 is NOT a good idea.  Yeah it will your choice number 3 to beat out the lower ones but then also you are not helping your choice #3 beat out your choice #1 in effect you are saying 'choice one and three you are the same in my mind' when really you like your number one choice alot more.  So you lost your chance to distinguish between something you really like and something that is just okay."

etc.

etc.

etc.

-Eddie
Title: Re:experiment
Post by: Eddie_Bradford on November 01, 2002, 10:20:43 pm
Quote
JT - the Research Committee considered narrowing down the choices further, but there was an outcry against that as "undemocratic" and so on, which is a legitimate concern.  So we opted for an absolute population cutoff.  Eliminating any further states would be fraught with controversy.

Not necissarily Jason!  People just got mad last time because they didn't have a say in it.  Why not let people vote on this kind of thing?  There are several fun ways of cutting down on the number of states in a way that gets everyone involved.
There could be just a vote where you indicate one or two states you'd like eliminated.  Next to it you could have the research comittee's recomendation.
I liked my elimistate idea where the research comittee will pick a state to be eliminated and then the membership does a vote of confidence on it, just indicating whether it should be eliminated or not.  I really think people would enjoy that and it would help keep them involved in the project.

-Eddie
Title: Re:experiment
Post by: JasonPSorens on November 01, 2002, 10:35:12 pm
Yeah - I'm just worried about having some kind of proxy vote on the state before the time comes.  The 2nd method you mention might alleviate this problem...but then a few hundred people have signed up in the last 2 months on the assumption that 10 states are being considered.  I can see maybe South Dakota being eliminated without too much fuss but on the others there would probably be controversy. ;)
Title: Re:experiment
Post by: Charley on November 02, 2002, 08:20:40 pm
My "low vote total wins" proposal was only a little tounge-in-cheek.  The primary reason I proposed it, or something similar, is to get away from the strategic voting schemes that the cumulative vote plan encourages.  Cumulative voting, proportional voting etc are far more popular in some other countries and in poly-sci classes than in most of the US.   Are we going to try to change the voting process in the freestate to this same schema??  If so,  count me out.  If not, then why are we using it now??
Title: Re:experiment
Post by: JasonPSorens on November 03, 2002, 10:06:51 pm
No, we don't have any official position on government elections; I favor something like proportional representation for the legislature, instant runoff voting for governor races.
http://www.fairvote.org
Title: Re:experiment
Post by: Eddie_Bradford on November 04, 2002, 12:16:43 am
Darn I thought of a drawback to the rating system.  Although it's not necissarily the wrong answer.  Should someone get elected even if 50%+ of the population prefers ANYONE else?

49% give bob a rating of 100 out of 100 give sue 0 and Mike 0

25% give Mike 100 and sue 10 give bob 0

26% give Sue 100 and Mike 10 give bob 0

Final score:
Bob 4900
Sue  2850
Mike 2760

Bob wins by lots despite the fact that 51% want him the least.
-Eddie
Title: Re:experiment
Post by: Otosan on November 04, 2002, 06:43:21 am
I have found that you can make numbers do anything ya want, and what I have read pple are trying to figure out how to make their favorite state #1.

My suggestion is to have all ten states listed.
Give 10 pts to your #1 state, 9 pts to your #2 state and 8 pts to your #3 state ....etc....to the #10 state gets 1 pt.

If you give zero pts to any state your vote would not count!

And the state with the most pts win!

Title: Re:experiment
Post by: ZionCurtain on November 04, 2002, 08:56:02 am
Darn I thought of a drawback to the rating system.  Although it's not necissarily the wrong answer.  Should someone get elected even if 50%+ of the population prefers ANYONE else?

49% give bob a rating of 100 out of 100 give sue 0 and Mike 0

25% give Mike 100 and sue 10 give bob 0

26% give Sue 100 and Mike 10 give bob 0

Final score:
Bob 4900
Sue  2850
Mike 2760

Bob wins by lots despite the fact that 51% want him the least.
-Eddie
Isn't that about how Bush got elected? Or something like it.  ;D
Title: Borda Count Results
Post by: JasonPSorens on November 26, 2002, 02:40:15 pm
The Borda Count results of the above rankings are as follows:

Alaska   60
Idaho   59
Montana   89
Wyoming   54
North Dakota   61
South Dakota   36
Vermont   41
New Hampshire   58
Maine   47
Delaware   35

Note that Montana still wins, by a huge margin, so Montana wins under all 4 methods.  The Pearson r correlation between Borda Count and CC results is 0.84 and between Borda Count and Rating results is 0.94.  Interestingly, North Dakota comes in second under Borda Count, while the best it does under any other method is third.  New Hampshire falls to fifth even though it is second under other methods.