[EM] fix to Rivest 3-ballot flaw! very simple
I posted it on http://groups.yahoo.com/group/RangeVoting/ now back in action for plurality voting. (And to reply Juho Laatu, the problem is, plurality votign really cannot be "split" into "cells" because the cell-entries depend on one another - with approval and range they can be split because they do not depend.) wds election-methods mailing list - see http://electorama.com/em for list info
Re: [EM] 3Ballot -- Condorcet version # (or is it #3?) by Mrouse
I think there are ways around this, or at least tricks that can make the problems more manageable. The particular trick I'm thinking at the moment is to allow the voter to give her preference order to the voting machine in whatever nice way. Then the voting machine proposes three ballots. The voter is however able to change the ballots in whatever way she likes (the voting machine will assist her) . When she is satisfied she tells the voting machine to print the ballots. It is possible to manually check some rows that they are as intended (e.g. "A>B, A>B, B>A" is ok since she preferred A over B). SInce many voters may just ask the voting machine to print the ballots right away without changing or checking them it would make sense to first propose a receipt (and corresponding ballot) that has random content that the voting machine has generated independently of the voter's preferences. (The other two ballots would be generated so that they take into account voter's preferences and the content of the first ballot.) Juho Laatu On Oct 3, 2006, at 23:59 , <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> wrote: > That's true, this method would definitely require computers to > generate > the ballots, which is a fairly significant flaw. I was mainly > throwing out > ideas to see if someone would think of something clever and say, > "AHA! If > you just do it this way, people can easily cast a secure Condorcet > vote > they can check later on." > > I didn't realize you had come up with something similar, and by no > means > did I intend to take credit for it. I think the 3Ballot idea is clever > enough that a bit of brainstorming can be fun. > > Michael Rouse > [EMAIL PROTECTED] > >> To Mrouse: >> you've come full circle - that pairwise plan was the one I originally >> thought of before you came to my attention. >> >> However, I think you are going in the wrong direction because the >> goal is to avoid need for govt-owned >> computers that the voter has to trust to >> (a) do the right thing and (b) not remember anything. >> The goal is to get security without computers, without complexity, >> and >> without having to trust. >> >> -- >> Warren D. Smith >> http://RangeVoting.org <-- add your endorsement >> and >> math.temple.edu/~wds/homepage/works.html >> >> election-methods mailing list - see http://electorama.com/em for list >> info > > > > > election-methods mailing list - see http://electorama.com/em for > list info ___ The all-new Yahoo! Mail goes wherever you go - free your email address from your Internet provider. http://uk.docs.yahoo.com/nowyoucan.html election-methods mailing list - see http://electorama.com/em for list info
Re: [EM] 3Ballot -- Condorcet version # (or is it #3?) by Mrouse
That's true, this method would definitely require computers to generate the ballots, which is a fairly significant flaw. I was mainly throwing out ideas to see if someone would think of something clever and say, "AHA! If you just do it this way, people can easily cast a secure Condorcet vote they can check later on." I didn't realize you had come up with something similar, and by no means did I intend to take credit for it. I think the 3Ballot idea is clever enough that a bit of brainstorming can be fun. Michael Rouse [EMAIL PROTECTED] > To Mrouse: > you've come full circle - that pairwise plan was the one I originally > thought of before you came to my attention. > > However, I think you are going in the wrong direction because the > goal is to avoid need for govt-owned > computers that the voter has to trust to > (a) do the right thing and (b) not remember anything. > The goal is to get security without computers, without complexity, and > without having to trust. > > -- > Warren D. Smith > http://RangeVoting.org <-- add your endorsement > and > math.temple.edu/~wds/homepage/works.html > > election-methods mailing list - see http://electorama.com/em for list > info election-methods mailing list - see http://electorama.com/em for list info
Re: [EM] 3ballot / rouse idea
Warren Smith wrote:
> So you see the tremendous devastation this whole 3-pattern attack
> idea can wreak.
> But in range and approval voting, all candidate scoring decision
> are independent
> and can be on seperate "ballots" so there IS NO pattern, so not a
> problem.
> With Condorcet, Plurality, etc, they depend and cannot be seperated
> without making ballot-validity-check impossible.
I didn't quite understand this point.
I assume that even if the three ballots are split in several smaller
ballots, checking them all (against the rules of the voting method in
question) at the same time would be possible. Rivest also discussed
the option of splitting the three plurality ballots in 3r separate
ballots ("cell-based approach").
In the preference order based methods like IRV and Condorcet there
are some additional problems. But in such Condorcet methods that are
based purely on a summable pairwise comparison matrix, separating the
elements of the matrix from each others should be ok (e.g. splitting
the three ballots in the "version #2" that was proposed by Michael
Rouse).
On the other side, also range voting has some problems.
The content of a range ballot is rich, containing numbers e.g. from 0
to 99. These numbers give more information to a coercer than e.g. in
plurality and approval. What if the coercer gives the voter a set of
9 numbers that she should use when filling the 3 ballots in a 3
candidate election. Let's say that one of the ballots was supposed to
be 12/34/56 but none of the ballots in the bulletin board matches
that. The coercer can now see that the voter didn't vote as told. I
guess one could break the range vote to three separate ballots and
even further e.g. to 9 rows (and separate ballots) per each
candidate, each ballot giving 0..11 points to a candidate (instead of
the original 0..99). But that wouldn't be very practical. Anyway, the
problem is that range ballots should be split in smaller ballots than
e.g. approval to reach the same security level.
Juho Laatu
___
The all-new Yahoo! Mail goes wherever you go - free your email address from
your Internet provider. http://uk.docs.yahoo.com/nowyoucan.html
election-methods mailing list - see http://electorama.com/em for list info
Re: [EM] 3ballot - revolutionary new protocol for secure secret ballot elections
To KPY: I thought you had in mind, computer randomizes pattern and prints out the 3 ballot plus 1 copy (it decides which). That is a disaster since enables trivial vote buying using statistical effects. If same, but YOU decide which to copy, then the scheme still is vulnerable to statistical effects: Example: I support both Clinton and Gore but hate Bush (suppose). I tell you "vote for Clinton and bring me back a ballot-copy that says Clinton=1, Gore=1, and you get a dollar." If you do not vote Clinton but do vote Gore, then your probability of being able to get the dollar, is 2/3. If you do not vote Clinton and do vote Bush, then your probability of being able to get the dollar, is 1/3. If you do vote Clinton, your probability is 2/3. If I do this enough times, I bias the vote a lot. -- Warren D. Smith http://RangeVoting.org <-- add your endorsement and math.temple.edu/~wds/homepage/works.html election-methods mailing list - see http://electorama.com/em for list info
[EM] 3Ballot -- Condorcet version # (or is it #3?) by Mrouse
To Mrouse: you've come full circle - that pairwise plan was the one I originally thought of before you came to my attention. However, I think you are going in the wrong direction because the goal is to avoid need for govt-owned computers that the voter has to trust to (a) do the right thing and (b) not remember anything. The goal is to get security without computers, without complexity, and without having to trust. -- Warren D. Smith http://RangeVoting.org <-- add your endorsement and math.temple.edu/~wds/homepage/works.html election-methods mailing list - see http://electorama.com/em for list info
Re: [EM] 3ballot - revolutionary new protocol for secure secret ballot elections
Warren Smith wds at math.temple.edu wrote: > [The specified pattern] attack appears to be quite devastating to me. > I personally regard Rivest's scheme as therefore dead or anyway on the critical list, > for purpose of applying it to plurality voting. Rivest has a few lame attempts > to rescusitate it, but they do not impress me much. However, for approval & range voting, > this attack (in the de-bundled world of course) is not an attack at all. It > is nothing. So Rivest's idea is strong and healthy for range & approval voting. One solution would be to cut the ballot after/during casting. The 35 person ballot would be kept together until after it has been validated for correct marking. After that it can be split into say 7 5-person ballots or something. There would need to be an ID on each ballot for each set of 5 candidates. The voter would effectively get 7 receipts in the above example. What is still needed is a way to prevent the voter from knowing the ID of votes that are cast that he doesn't get a receipt for. Otherwise, the "buyer" can demand the list of ID numbers. Also, if the ballot is split into 35 parts, then this is equivalent in complexity (and number of receipts) to using approval or range. > >A u-R Lomax: > Vote-buying may shift close elections; but, > frankly, I think it is rare. very rare. > > --Well, that is not clear. May be rare here & now, but historically it has at the right > times and places been not rare, but "the norm." Read LBJ biography & Gumbel book if you > want to learn about that. Also in the right kind of country, it (or coercion) may be common or > an easy way to slide into non-democracy. I guess the issue is that it might be rare because the voting system (tries to) guard against it. The benefit for secret ballots is that if voters have a better than 50% chance of being correct on a question, then asking lots of them will likely give the right answer. If people can sell their votes, this law of large number effect is reduced. This is in addition to the issue of coercion. > STENCILS: > --One more thing. Raphfrk on this forum suggested some > ideas about sliding widgets and 3ballot. > I thought his stuff (assuming raphfrk is a "him") was rather kludgy and confused > and I wasn't even sure it would work It would :). The main issue is getting ID numbers to work, without showing the voter the ID for ballots that are not in the copy slot. Btw, a mechanical system that people can see operating is a hell of alot easier for people to understand than a list of instructions. Using the suggested system, it would be impossible to cast an illegal vote. > , BUT, it had the germ of a very good idea. > Which I call "STENCILS." I think that the two systems can be combined. You put the ballot in a slot. There is one hole over each candidate's name. You use the handle to move a slider to cover one hold for each candidate. There is also a "voting" handle which you use to cover an empty hole for the candidate that you want to vote for. > > I have
