[EM] Modified Overall Preferences

classic Classic list List threaded Threaded
23 messages Options
12
Reply | Threaded
Open this post in threaded view
|

[EM] Modified Overall Preferences

Juho Laatu-4
The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.

You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).

In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).

Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).

(I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)

I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.

The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)

Here's one example set of votes.

45 A>B>>C --> A>>C>>B
15 B>A>>C
40 C>B>>A

B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).

Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.

17 A>>B>>C>>D
17 B>>C>>A>>D
17 C>>A>>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.

17 A>B>C>>D
17 B>C>A>>D
17 C>A>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.


----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences

Juho Laatu-4
Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.

> Forest Simmons fsimmons at pcc.edu
> Thu May 30
>
>> In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??
>>
>> I am interested in simple methods that always ...
>>
>> (1) elect candidate A given the following profile:
>> P: A
>> Q: B>>C
>> R: C,
>>
>> and
>> (2) elect candidate C given
>> P: A
>> Q: B>C>>
>> R: C,
>>
>> and
>> (3) elect candidate B given
>> P: A
>> Q: B>>C?? (or B>C)
>> R: C>>B. (or C>B)


I copy my answers to those questions below.

Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).

> I tried to see how it relates to Forest's requests.
>
> Case (2)
> Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.
>
> Case (1)
> The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.
>
> Case (3)
> Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.
>
> In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.

I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.

P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typical Condorcet me
 thods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.

P.P.S. In the original mail, ignore words "which means".

Juho


> On 16 Jun 2019, at 16:51, Juho Laatu <[hidden email]> wrote:
>
> The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.
>
> You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).
>
> In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).
>
> Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).
>
> (I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)
>
> I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.
>
> The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)
>
> Here's one example set of votes.
>
> 45 A>B>>C --> A>>C>>B
> 15 B>A>>C
> 40 C>B>>A
>
> B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).
>
> Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.
>
> 17 A>>B>>C>>D
> 17 B>>C>>A>>D
> 17 C>>A>>B>>D
> 16 D>>A>>B>>C
> 16 D>>B>>C>>A
> 16 D>>C>>A>>B
>
> A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.
>
> 17 A>B>C>>D
> 17 B>C>A>>D
> 17 C>A>B>>D
> 16 D>>A>>B>>C
> 16 D>>B>>C>>A
> 16 D>>C>>A>>B
>
> A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.
>
>
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences

Juho Laatu-4
I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.

The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.

The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).

The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.

I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial strengths) to create
  space also for the weakened preferences.

Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.

Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.

Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.

45: A > B >> C
15: B > A >> C
40: C > B > A

A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).

40: A (no "clones")
35: B>>C>A (A and C indicated as "clones")
25: C (no "clones")

This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.

40: A
35: B>C>>A
25: C

This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.

40: A
35: B>>C>A
25: C>>B>A

This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.

If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.



> On 18 Jun 2019, at 09:01, Juho Laatu <[hidden email]> wrote:
>
> Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.
>
>> Forest Simmons fsimmons at pcc.edu
>> Thu May 30
>>
>>> In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??
>>>
>>> I am interested in simple methods that always ...
>>>
>>> (1) elect candidate A given the following profile:
>>> P: A
>>> Q: B>>C
>>> R: C,
>>>
>>> and
>>> (2) elect candidate C given
>>> P: A
>>> Q: B>C>>
>>> R: C,
>>>
>>> and
>>> (3) elect candidate B given
>>> P: A
>>> Q: B>>C?? (or B>C)
>>> R: C>>B. (or C>B)
>
>
> I copy my answers to those questions below.
>
> Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).
>
>> I tried to see how it relates to Forest's requests.
>>
>> Case (2)
>> Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.
>>
>> Case (1)
>> The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.
>>
>> Case (3)
>> Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.
>>
>> In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
>
> I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.
>
> P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typical Condorcet
 methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.

>
> P.P.S. In the original mail, ignore words "which means".
>
> Juho
>
>
>> On 16 Jun 2019, at 16:51, Juho Laatu <[hidden email]> wrote:
>>
>> The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.
>>
>> You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).
>>
>> In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).
>>
>> Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).
>>
>> (I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)
>>
>> I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.
>>
>> The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)
>>
>> Here's one example set of votes.
>>
>> 45 A>B>>C --> A>>C>>B
>> 15 B>A>>C
>> 40 C>B>>A
>>
>> B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).
>>
>> Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.
>>
>> 17 A>>B>>C>>D
>> 17 B>>C>>A>>D
>> 17 C>>A>>B>>D
>> 16 D>>A>>B>>C
>> 16 D>>B>>C>>A
>> 16 D>>C>>A>>B
>>
>> A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.
>>
>> 17 A>B>C>>D
>> 17 B>C>A>>D
>> 17 C>A>B>>D
>> 16 D>>A>>B>>C
>> 16 D>>B>>C>>A
>> 16 D>>C>>A>>B
>>
>> A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.
>>
>>
>> ----
>> Election-Methods mailing list - see https://electorama.com/em for list info
>

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences

Juho Laatu-4
James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.

Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.

Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.

- AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.

- In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.

In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates that one wants
 to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".

In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.

I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably not practical,
 since they would be just "weakened opinions".)


> On 26 Jun 2019, at 10:55, Juho Laatu <[hidden email]> wrote:
>
> I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.
>
> The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.
>
> The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).
>
> The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.
>
> I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial strengths) to crea
 te

> space also for the weakened preferences.
>
> Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.
>
> Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.
>
> Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.
>
> 45: A > B >> C
> 15: B > A >> C
> 40: C > B > A
>
> A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).
>
> 40: A (no "clones")
> 35: B>>C>A (A and C indicated as "clones")
> 25: C (no "clones")
>
> This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.
>
> 40: A
> 35: B>C>>A
> 25: C
>
> This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.
>
> 40: A
> 35: B>>C>A
> 25: C>>B>A
>
> This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.
>
> If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.
>
>
>
>> On 18 Jun 2019, at 09:01, Juho Laatu <[hidden email]> wrote:
>>
>> Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.
>>
>>> Forest Simmons fsimmons at pcc.edu
>>> Thu May 30
>>>
>>>> In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??
>>>>
>>>> I am interested in simple methods that always ...
>>>>
>>>> (1) elect candidate A given the following profile:
>>>> P: A
>>>> Q: B>>C
>>>> R: C,
>>>>
>>>> and
>>>> (2) elect candidate C given
>>>> P: A
>>>> Q: B>C>>
>>>> R: C,
>>>>
>>>> and
>>>> (3) elect candidate B given
>>>> P: A
>>>> Q: B>>C?? (or B>C)
>>>> R: C>>B. (or C>B)
>>
>>
>> I copy my answers to those questions below.
>>
>> Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).
>>
>>> I tried to see how it relates to Forest's requests.
>>>
>>> Case (2)
>>> Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.
>>>
>>> Case (1)
>>> The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.
>>>
>>> Case (3)
>>> Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.
>>>
>>> In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
>>
>> I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.
>>
>> P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typical Condorcet
 

> methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
>>
>> P.P.S. In the original mail, ignore words "which means".
>>
>> Juho
>>
>>
>>> On 16 Jun 2019, at 16:51, Juho Laatu <[hidden email]> wrote:
>>>
>>> The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.
>>>
>>> You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).
>>>
>>> In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).
>>>
>>> Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).
>>>
>>> (I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)
>>>
>>> I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.
>>>
>>> The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)
>>>
>>> Here's one example set of votes.
>>>
>>> 45 A>B>>C --> A>>C>>B
>>> 15 B>A>>C
>>> 40 C>B>>A
>>>
>>> B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).
>>>
>>> Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.
>>>
>>> 17 A>>B>>C>>D
>>> 17 B>>C>>A>>D
>>> 17 C>>A>>B>>D
>>> 16 D>>A>>B>>C
>>> 16 D>>B>>C>>A
>>> 16 D>>C>>A>>B
>>>
>>> A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.
>>>
>>> 17 A>B>C>>D
>>> 17 B>C>A>>D
>>> 17 C>A>B>>D
>>> 16 D>>A>>B>>C
>>> 16 D>>B>>C>>A
>>> 16 D>>C>>A>>B
>>>
>>> A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.
>>>
>>>
>>> ----
>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>
>
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / burying

Juho Laatu-4
The MOP approach is quite good in thwarting possible strategic burying attempts. I'll address briefly the basic case of having three candidates that will be looped strategically.

I'll use one particular MOP method that we might call MOP-F2 (identify Favourites using a cutoff, stretch factor = 2). It will be symmetric (both A>B and B>A will be weakened if A and B are marked as favourites), and it will be simply based on use of margins and Minmax. The modification function will be f(m) = m + sign(m) * (1-P/N) * 2 * N, where N is the number of voters, and P is the number of voters that identified the two candidates as their (Protected) favourites (= candidates above the cutoff). Parameter m is the original margin, and the output of the function will be the modified value to be fed to Minmax. Note that with 100 voters the original margin values are in range [-100, 100], and the modified values are in range [-300, 300].

I chose stretch factor to be 2, since that gives us some quite natural defensive properties. it seems that when certain number of of voters try to bury the sincere CW to make their favourite the winner (or vote sincerely that way), it is enough if a "corresponding" number of the other voters indicate the other two candidates as their favourites, to thwart the strategic attempt. In the example below 100% of the A supporters are strategic, and 100% of those voters that can easily defend against the strategy will do so. Stretch factor 1 would not give us such "full protection", and stretch factor 3 might be considered already an overkill.

49: A>B>>C --> A>C>B large number of strategists that try to bury B
15: B>A>>C --> B>A>C sincere CW, and a member of the AB group
36: C>B>A --> C>B>>A the other group that decides to protect the CW

The opinions on the left are sincere. The actual votes are on the right. Note that all defensive votes (B and C supporters) can maintain the original ranking order. There is thus no need to falsify their rankings. The only required change is that they change the position of their favourite cutoff. B supporters stop protecting A, once they hear about the strategic plans of the A supporters. C supporters extend their protection to cover B (the CW), once they hear about the strategic plans of the A supporters.

The worst defeats with the original (non-modified) margins are A:-2, B:2 (CW), and C:-28. If A voters apply their strategy, the worst defeat margins become A:-2, B:-70, and C:-28. If the votes are modified, and B supporters still protect A, and C supporters do not protect B, the modified worst defeats become A:-172, B:-270, and C:-228. If B supporters stop protecting A, the modified worst defeats become A:-202, B:-270, and C:-228. And if C supporters protect B, the modified worst defeats become A:-202, B:-198, and C:-228. CW (B) wins again. It seems that one can effectively defend against any such (single) strategic burying attempt. And if you can, maybe nobody ever even tries such strategic attempts. This would make use of the cutoffs even more sincere, and more unnecessary, but maybe still good sincere information for the election analysts.

I like the idea that rankings of the non-starategic voters may stay the same, and also that the favourite cutoffs may stay quite natural, even when used defensively. The non-strategic voters may use the protective cutoff in a quite natural way, while it is of no use to the strategists. (If the A supporters would vote A>C>>B, they would just support the election of C that is the worst alternative to them.) There is also no need to resort to use of winning votes or something similar to provide means to defend against strategies (margins are a more natural way to measure sincere preferences).




P.S. I note that in an earlier mail I forgot to include the sign function in one of the the equations. That equation should have been sign(initial)*(1-mod)*k*max + initial. I'm sorry if someone tried to calculate the results and got confused, and didn't dare to ask why that function seemed strange. I hope at least the function I gave above is correct. I think that one (MOP-F2) is a good one to start with, if someone wants to test the MOP methods in practice.

> On 08 Jul 2019, at 10:06, Juho Laatu <[hidden email]> wrote:
>
> James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.
>
> Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.
>
> Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.
>
> - AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.
>
> - In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.
>
> In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates that one want
 s
> to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".
>
> In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.
>
> I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably not practical
 ,

> since they would be just "weakened opinions".)
>
>
>> On 26 Jun 2019, at 10:55, Juho Laatu <[hidden email]> wrote:
>>
>> I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.
>>
>> The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.
>>
>> The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).
>>
>> The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.
>>
>> I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial strengths) to cre
 a

> te
>> space also for the weakened preferences.
>>
>> Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.
>>
>> Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.
>>
>> Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.
>>
>> 45: A > B >> C
>> 15: B > A >> C
>> 40: C > B > A
>>
>> A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).
>>
>> 40: A (no "clones")
>> 35: B>>C>A (A and C indicated as "clones")
>> 25: C (no "clones")
>>
>> This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.
>>
>> 40: A
>> 35: B>C>>A
>> 25: C
>>
>> This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.
>>
>> 40: A
>> 35: B>>C>A
>> 25: C>>B>A
>>
>> This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.
>>
>> If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.
>>
>>
>>
>>> On 18 Jun 2019, at 09:01, Juho Laatu <[hidden email]> wrote:
>>>
>>> Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.
>>>
>>>> Forest Simmons fsimmons at pcc.edu
>>>> Thu May 30
>>>>
>>>>> In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??
>>>>>
>>>>> I am interested in simple methods that always ...
>>>>>
>>>>> (1) elect candidate A given the following profile:
>>>>> P: A
>>>>> Q: B>>C
>>>>> R: C,
>>>>>
>>>>> and
>>>>> (2) elect candidate C given
>>>>> P: A
>>>>> Q: B>C>>
>>>>> R: C,
>>>>>
>>>>> and
>>>>> (3) elect candidate B given
>>>>> P: A
>>>>> Q: B>>C?? (or B>C)
>>>>> R: C>>B. (or C>B)
>>>
>>>
>>> I copy my answers to those questions below.
>>>
>>> Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).
>>>
>>>> I tried to see how it relates to Forest's requests.
>>>>
>>>> Case (2)
>>>> Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.
>>>>
>>>> Case (1)
>>>> The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.
>>>>
>>>> Case (3)
>>>> Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.
>>>>
>>>> In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
>>>
>>> I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.
>>>
>>> P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typical Condorce
 t

>
>> methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
>>>
>>> P.P.S. In the original mail, ignore words "which means".
>>>
>>> Juho
>>>
>>>
>>>> On 16 Jun 2019, at 16:51, Juho Laatu <[hidden email]> wrote:
>>>>
>>>> The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.
>>>>
>>>> You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).
>>>>
>>>> In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).
>>>>
>>>> Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).
>>>>
>>>> (I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)
>>>>
>>>> I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.
>>>>
>>>> The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)
>>>>
>>>> Here's one example set of votes.
>>>>
>>>> 45 A>B>>C --> A>>C>>B
>>>> 15 B>A>>C
>>>> 40 C>B>>A
>>>>
>>>> B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).
>>>>
>>>> Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.
>>>>
>>>> 17 A>>B>>C>>D
>>>> 17 B>>C>>A>>D
>>>> 17 C>>A>>B>>D
>>>> 16 D>>A>>B>>C
>>>> 16 D>>B>>C>>A
>>>> 16 D>>C>>A>>B
>>>>
>>>> A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.
>>>>
>>>> 17 A>B>C>>D
>>>> 17 B>C>A>>D
>>>> 17 C>A>B>>D
>>>> 16 D>>A>>B>>C
>>>> 16 D>>B>>C>>A
>>>> 16 D>>C>>A>>B
>>>>
>>>> A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.
>>>>
>>>>
>>>> ----
>>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>>
>>
>> ----
>> Election-Methods mailing list - see https://electorama.com/em for list info
>
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

Juho Laatu-4
Few more words on using margins and Minmax in MOP-F2. These words are intended for readers who normally do not like margins, and who prefer methods that always elect the winner from the Smith set.

The previous mail already demonstrated that margins can be used, and still provide protection against some strategic attempts. Margins provide good results with sincere votes, so why not use margins, and leave the strategic concerns to the cutoff part of the method.

The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed). There are however also cases where the top looped candidates are indeed clones in the sense that all voters consider them to be closely related. In such case their defeats to each others could be considered meaningless, and the winner should be one of them, even if there are candidates outside the Smith set with smaller defeats than the looped candidates have between themselves.

MOP-F2 can separate these different cases from each other by using the cutoff. This makes Smith set criterion compatibility a matter of the voters to decide. If voters vote 17: A>B>C>>D, 17: B>C>A>>D, 17: C>A>B>>D, 16: D>A>B>C, 16: D>B>C>A, 16: D>C>A>B, one of A, B and C will win. If A, B and C supporters will not use cutoffs, D wins in MOP-F2.

The point here is that use of Minmax as part of MOP-F2 makes it possible to voters to express their opinions more extensively (rankings + favourites) than in methods that use only rankings, and that this approach allows voters to indicate which candidates should be treated as clones and which ones not. Explicitly indicated clones will be treated as one fixed group of candidates (clones), without paying too much attention to their defeats to each others, in the final results.

One typical case where looped candidates would not be treated as if they were one candidate (group of clones) is 17: A>B>D>C, 17: B>C>D>A, 17: C>A>D>B, 16: A>D>B>C, 16: B>D>C>A, 16: C>D>A>B. In these votes A, B and C are clearly not clones since they are not next to each others in any of the ballots. D wins because its worst defeat is the smallest. With these votes, voters can not identify A, B and C as one uniform group on favourites (since they are not next to each others). Some voters might however use the cutoffs in some way, and change the result in some direction.

Margins and Minmax can thus be used, and the additional cutoffs may provide additional properties that plain margins and plain Minmax do not provide.



> On 08 Jul 2019, at 13:36, Juho Laatu <[hidden email]> wrote:
>
> The MOP approach is quite good in thwarting possible strategic burying attempts. I'll address briefly the basic case of having three candidates that will be looped strategically.
>
> I'll use one particular MOP method that we might call MOP-F2 (identify Favourites using a cutoff, stretch factor = 2). It will be symmetric (both A>B and B>A will be weakened if A and B are marked as favourites), and it will be simply based on use of margins and Minmax. The modification function will be f(m) = m + sign(m) * (1-P/N) * 2 * N, where N is the number of voters, and P is the number of voters that identified the two candidates as their (Protected) favourites (= candidates above the cutoff). Parameter m is the original margin, and the output of the function will be the modified value to be fed to Minmax. Note that with 100 voters the original margin values are in range [-100, 100], and the modified values are in range [-300, 300].
>
> I chose stretch factor to be 2, since that gives us some quite natural defensive properties. it seems that when certain number of of voters try to bury the sincere CW to make their favourite the winner (or vote sincerely that way), it is enough if a "corresponding" number of the other voters indicate the other two candidates as their favourites, to thwart the strategic attempt. In the example below 100% of the A supporters are strategic, and 100% of those voters that can easily defend against the strategy will do so. Stretch factor 1 would not give us such "full protection", and stretch factor 3 might be considered already an overkill.
>
> 49: A>B>>C --> A>C>B large number of strategists that try to bury B
> 15: B>A>>C --> B>A>C sincere CW, and a member of the AB group
> 36: C>B>A --> C>B>>A the other group that decides to protect the CW
>
> The opinions on the left are sincere. The actual votes are on the right. Note that all defensive votes (B and C supporters) can maintain the original ranking order. There is thus no need to falsify their rankings. The only required change is that they change the position of their favourite cutoff. B supporters stop protecting A, once they hear about the strategic plans of the A supporters. C supporters extend their protection to cover B (the CW), once they hear about the strategic plans of the A supporters.
>
> The worst defeats with the original (non-modified) margins are A:-2, B:2 (CW), and C:-28. If A voters apply their strategy, the worst defeat margins become A:-2, B:-70, and C:-28. If the votes are modified, and B supporters still protect A, and C supporters do not protect B, the modified worst defeats become A:-172, B:-270, and C:-228. If B supporters stop protecting A, the modified worst defeats become A:-202, B:-270, and C:-228. And if C supporters protect B, the modified worst defeats become A:-202, B:-198, and C:-228. CW (B) wins again. It seems that one can effectively defend against any such (single) strategic burying attempt. And if you can, maybe nobody ever even tries such strategic attempts. This would make use of the cutoffs even more sincere, and more unnecessary, but maybe still good sincere information for the election analysts.
>
> I like the idea that rankings of the non-starategic voters may stay the same, and also that the favourite cutoffs may stay quite natural, even when used defensively. The non-strategic voters may use the protective cutoff in a quite natural way, while it is of no use to the strategists. (If the A supporters would vote A>C>>B, they would just support the election of C that is the worst alternative to them.) There is also no need to resort to use of winning votes or something similar to provide means to defend against strategies (margins are a more natural way to measure sincere preferences).
>
>
>
>
> P.S. I note that in an earlier mail I forgot to include the sign function in one of the the equations. That equation should have been sign(initial)*(1-mod)*k*max + initial. I'm sorry if someone tried to calculate the results and got confused, and didn't dare to ask why that function seemed strange. I hope at least the function I gave above is correct. I think that one (MOP-F2) is a good one to start with, if someone wants to test the MOP methods in practice.
>
>> On 08 Jul 2019, at 10:06, Juho Laatu <[hidden email]> wrote:
>>
>> James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.
>>
>> Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.
>>
>> Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.
>>
>> - AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.
>>
>> - In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.
>>
>> In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates that one wan
 t
> s
>> to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".
>>
>> In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.
>>
>> I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably not practica
 l

> ,
>> since they would be just "weakened opinions".)
>>
>>
>>> On 26 Jun 2019, at 10:55, Juho Laatu <[hidden email]> wrote:
>>>
>>> I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.
>>>
>>> The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.
>>>
>>> The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).
>>>
>>> The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.
>>>
>>> I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial strengths) to cr
 e

> a
>> te
>>> space also for the weakened preferences.
>>>
>>> Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.
>>>
>>> Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.
>>>
>>> Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.
>>>
>>> 45: A > B >> C
>>> 15: B > A >> C
>>> 40: C > B > A
>>>
>>> A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).
>>>
>>> 40: A (no "clones")
>>> 35: B>>C>A (A and C indicated as "clones")
>>> 25: C (no "clones")
>>>
>>> This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.
>>>
>>> 40: A
>>> 35: B>C>>A
>>> 25: C
>>>
>>> This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.
>>>
>>> 40: A
>>> 35: B>>C>A
>>> 25: C>>B>A
>>>
>>> This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.
>>>
>>> If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.
>>>
>>>
>>>
>>>> On 18 Jun 2019, at 09:01, Juho Laatu <[hidden email]> wrote:
>>>>
>>>> Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.
>>>>
>>>>> Forest Simmons fsimmons at pcc.edu
>>>>> Thu May 30
>>>>>
>>>>>> In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??
>>>>>>
>>>>>> I am interested in simple methods that always ...
>>>>>>
>>>>>> (1) elect candidate A given the following profile:
>>>>>> P: A
>>>>>> Q: B>>C
>>>>>> R: C,
>>>>>>
>>>>>> and
>>>>>> (2) elect candidate C given
>>>>>> P: A
>>>>>> Q: B>C>>
>>>>>> R: C,
>>>>>>
>>>>>> and
>>>>>> (3) elect candidate B given
>>>>>> P: A
>>>>>> Q: B>>C?? (or B>C)
>>>>>> R: C>>B. (or C>B)
>>>>
>>>>
>>>> I copy my answers to those questions below.
>>>>
>>>> Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).
>>>>
>>>>> I tried to see how it relates to Forest's requests.
>>>>>
>>>>> Case (2)
>>>>> Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.
>>>>>
>>>>> Case (1)
>>>>> The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.
>>>>>
>>>>> Case (3)
>>>>> Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.
>>>>>
>>>>> In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
>>>>
>>>> I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.
>>>>
>>>> P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typical Condorc
 e

> t
>>
>>> methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
>>>>
>>>> P.P.S. In the original mail, ignore words "which means".
>>>>
>>>> Juho
>>>>
>>>>
>>>>> On 16 Jun 2019, at 16:51, Juho Laatu <[hidden email]> wrote:
>>>>>
>>>>> The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.
>>>>>
>>>>> You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).
>>>>>
>>>>> In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).
>>>>>
>>>>> Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).
>>>>>
>>>>> (I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)
>>>>>
>>>>> I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.
>>>>>
>>>>> The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)
>>>>>
>>>>> Here's one example set of votes.
>>>>>
>>>>> 45 A>B>>C --> A>>C>>B
>>>>> 15 B>A>>C
>>>>> 40 C>B>>A
>>>>>
>>>>> B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).
>>>>>
>>>>> Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.
>>>>>
>>>>> 17 A>>B>>C>>D
>>>>> 17 B>>C>>A>>D
>>>>> 17 C>>A>>B>>D
>>>>> 16 D>>A>>B>>C
>>>>> 16 D>>B>>C>>A
>>>>> 16 D>>C>>A>>B
>>>>>
>>>>> A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.
>>>>>
>>>>> 17 A>B>C>>D
>>>>> 17 B>C>A>>D
>>>>> 17 C>A>B>>D
>>>>> 16 D>>A>>B>>C
>>>>> 16 D>>B>>C>>A
>>>>> 16 D>>C>>A>>B
>>>>>
>>>>> A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.
>>>>>
>>>>>
>>>>> ----
>>>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>>>
>>>
>>> ----
>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>
>> ----
>> Election-Methods mailing list - see https://electorama.com/em for list info
>
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

C.Benham
Juho,

I gather that you are proposing that the voters should sometimes be
allowed to give more than one "cutoff" in their rankings.
I consider that to be far too clumsy and also too much looking like an
explicit strategy device.

If you want to do something like that I suggest 0-99 Score ballots be
used, and cutoffs and/or pairwise preference strengths can be
inferred from them.

You mention the Approval-Weighted Pairwise method. That method is
unacceptable to me because it can elect a candidate that
is doubly-defeated  (i.e. both pairwise and approval-wise) by another
candidate.

> Margins provide good results with sincere votes, so why not use margins...
I don't see how egregious failures of the Plurality  and Later-no-Help
(and even Non-Drastic Defense) criteria constitute "good results"
irrespective of whether the votes are "sincere" or not.

Chris Benham

On 8/07/2019 9:09 pm, Juho Laatu wrote:

> Few more words on using margins and Minmax in MOP-F2. These words are intended for readers who normally do not like margins, and who prefer methods that always elect the winner from the Smith set.
>
> The previous mail already demonstrated that margins can be used, and still provide protection against some strategic attempts. Margins provide good results with sincere votes, so why not use margins, and leave the strategic concerns to the cutoff part of the method.
>
> The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed). There are however also cases where the top looped candidates are indeed clones in the sense that all voters consider them to be closely related. In such case their defeats to each others could be considered meaningless, and the winner should be one of them, even if there are candidates outside the Smith set with smaller defeats than the looped candidates have between themselves.
>
> MOP-F2 can separate these different cases from each other by using the cutoff. This makes Smith set criterion compatibility a matter of the voters to decide. If voters vote 17: A>B>C>>D, 17: B>C>A>>D, 17: C>A>B>>D, 16: D>A>B>C, 16: D>B>C>A, 16: D>C>A>B, one of A, B and C will win. If A, B and C supporters will not use cutoffs, D wins in MOP-F2.
>
> The point here is that use of Minmax as part of MOP-F2 makes it possible to voters to express their opinions more extensively (rankings + favourites) than in methods that use only rankings, and that this approach allows voters to indicate which candidates should be treated as clones and which ones not. Explicitly indicated clones will be treated as one fixed group of candidates (clones), without paying too much attention to their defeats to each others, in the final results.
>
> One typical case where looped candidates would not be treated as if they were one candidate (group of clones) is 17: A>B>D>C, 17: B>C>D>A, 17: C>A>D>B, 16: A>D>B>C, 16: B>D>C>A, 16: C>D>A>B. In these votes A, B and C are clearly not clones since they are not next to each others in any of the ballots. D wins because its worst defeat is the smallest. With these votes, voters can not identify A, B and C as one uniform group on favourites (since they are not next to each others). Some voters might however use the cutoffs in some way, and change the result in some direction.
>
> Margins and Minmax can thus be used, and the additional cutoffs may provide additional properties that plain margins and plain Minmax do not provide.
>
>
>
>> On 08 Jul 2019, at 13:36, Juho Laatu <[hidden email]> wrote:
>>
>> The MOP approach is quite good in thwarting possible strategic burying attempts. I'll address briefly the basic case of having three candidates that will be looped strategically.
>>
>> I'll use one particular MOP method that we might call MOP-F2 (identify Favourites using a cutoff, stretch factor = 2). It will be symmetric (both A>B and B>A will be weakened if A and B are marked as favourites), and it will be simply based on use of margins and Minmax. The modification function will be f(m) = m + sign(m) * (1-P/N) * 2 * N, where N is the number of voters, and P is the number of voters that identified the two candidates as their (Protected) favourites (= candidates above the cutoff). Parameter m is the original margin, and the output of the function will be the modified value to be fed to Minmax. Note that with 100 voters the original margin values are in range [-100, 100], and the modified values are in range [-300, 300].
>>
>> I chose stretch factor to be 2, since that gives us some quite natural defensive properties. it seems that when certain number of of voters try to bury the sincere CW to make their favourite the winner (or vote sincerely that way), it is enough if a "corresponding" number of the other voters indicate the other two candidates as their favourites, to thwart the strategic attempt. In the example below 100% of the A supporters are strategic, and 100% of those voters that can easily defend against the strategy will do so. Stretch factor 1 would not give us such "full protection", and stretch factor 3 might be considered already an overkill.
>>
>> 49: A>B>>C --> A>C>B large number of strategists that try to bury B
>> 15: B>A>>C --> B>A>C sincere CW, and a member of the AB group
>> 36: C>B>A --> C>B>>A the other group that decides to protect the CW
>>
>> The opinions on the left are sincere. The actual votes are on the right. Note that all defensive votes (B and C supporters) can maintain the original ranking order. There is thus no need to falsify their rankings. The only required change is that they change the position of their favourite cutoff. B supporters stop protecting A, once they hear about the strategic plans of the A supporters. C supporters extend their protection to cover B (the CW), once they hear about the strategic plans of the A supporters.
>>
>> The worst defeats with the original (non-modified) margins are A:-2, B:2 (CW), and C:-28. If A voters apply their strategy, the worst defeat margins become A:-2, B:-70, and C:-28. If the votes are modified, and B supporters still protect A, and C supporters do not protect B, the modified worst defeats become A:-172, B:-270, and C:-228. If B supporters stop protecting A, the modified worst defeats become A:-202, B:-270, and C:-228. And if C supporters protect B, the modified worst defeats become A:-202, B:-198, and C:-228. CW (B) wins again. It seems that one can effectively defend against any such (single) strategic burying attempt. And if you can, maybe nobody ever even tries such strategic attempts. This would make use of the cutoffs even more sincere, and more unnecessary, but maybe still good sincere information for the election analysts.
>>
>> I like the idea that rankings of the non-starategic voters may stay the same, and also that the favourite cutoffs may stay quite natural, even when used defensively. The non-strategic voters may use the protective cutoff in a quite natural way, while it is of no use to the strategists. (If the A supporters would vote A>C>>B, they would just support the election of C that is the worst alternative to them.) There is also no need to resort to use of winning votes or something similar to provide means to defend against strategies (margins are a more natural way to measure sincere preferences).
>>
>>
>>
>>
>> P.S. I note that in an earlier mail I forgot to include the sign function in one of the the equations. That equation should have been sign(initial)*(1-mod)*k*max + initial. I'm sorry if someone tried to calculate the results and got confused, and didn't dare to ask why that function seemed strange. I hope at least the function I gave above is correct. I think that one (MOP-F2) is a good one to start with, if someone wants to test the MOP methods in practice.
>>
>>> On 08 Jul 2019, at 10:06, Juho Laatu <[hidden email]> wrote:
>>>
>>> James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.
>>>
>>> Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.
>>>
>>> Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.
>>>
>>> - AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.
>>>
>>> - In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.
>>>
>>> In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates that
>    one wan
>   t
>> s
>>> to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".
>>>
>>> In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.
>>>
>>> I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably not
>   practica
>   l
>> ,
>>> since they would be just "weakened opinions".)
>>>
>>>
>>>> On 26 Jun 2019, at 10:55, Juho Laatu <[hidden email]> wrote:
>>>>
>>>> I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.
>>>>
>>>> The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.
>>>>
>>>> The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).
>>>>
>>>> The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.
>>>>
>>>> I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial strength
>   s) to cr
>   e
>> a
>>> te
>>>> space also for the weakened preferences.
>>>>
>>>> Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.
>>>>
>>>> Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.
>>>>
>>>> Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.
>>>>
>>>> 45: A > B >> C
>>>> 15: B > A >> C
>>>> 40: C > B > A
>>>>
>>>> A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).
>>>>
>>>> 40: A (no "clones")
>>>> 35: B>>C>A (A and C indicated as "clones")
>>>> 25: C (no "clones")
>>>>
>>>> This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.
>>>>
>>>> 40: A
>>>> 35: B>C>>A
>>>> 25: C
>>>>
>>>> This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.
>>>>
>>>> 40: A
>>>> 35: B>>C>A
>>>> 25: C>>B>A
>>>>
>>>> This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.
>>>>
>>>> If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.
>>>>
>>>>
>>>>
>>>>> On 18 Jun 2019, at 09:01, Juho Laatu <[hidden email]> wrote:
>>>>>
>>>>> Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.
>>>>>
>>>>>> Forest Simmons fsimmons at pcc.edu
>>>>>> Thu May 30
>>>>>>
>>>>>>> In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??
>>>>>>>
>>>>>>> I am interested in simple methods that always ...
>>>>>>>
>>>>>>> (1) elect candidate A given the following profile:
>>>>>>> P: A
>>>>>>> Q: B>>C
>>>>>>> R: C,
>>>>>>>
>>>>>>> and
>>>>>>> (2) elect candidate C given
>>>>>>> P: A
>>>>>>> Q: B>C>>
>>>>>>> R: C,
>>>>>>>
>>>>>>> and
>>>>>>> (3) elect candidate B given
>>>>>>> P: A
>>>>>>> Q: B>>C?? (or B>C)
>>>>>>> R: C>>B. (or C>B)
>>>>>
>>>>> I copy my answers to those questions below.
>>>>>
>>>>> Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).
>>>>>
>>>>>> I tried to see how it relates to Forest's requests.
>>>>>>
>>>>>> Case (2)
>>>>>> Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.
>>>>>>
>>>>>> Case (1)
>>>>>> The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.
>>>>>>
>>>>>> Case (3)
>>>>>> Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.
>>>>>>
>>>>>> In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
>>>>> I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.
>>>>>
>>>>> P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typical
>    Condorc
>   e
>> t
>>>> methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
>>>>> P.P.S. In the original mail, ignore words "which means".
>>>>>
>>>>> Juho
>>>>>
>>>>>
>>>>>> On 16 Jun 2019, at 16:51, Juho Laatu <[hidden email]> wrote:
>>>>>>
>>>>>> The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.
>>>>>>
>>>>>> You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).
>>>>>>
>>>>>> In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).
>>>>>>
>>>>>> Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).
>>>>>>
>>>>>> (I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)
>>>>>>
>>>>>> I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.
>>>>>>
>>>>>> The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)
>>>>>>
>>>>>> Here's one example set of votes.
>>>>>>
>>>>>> 45 A>B>>C --> A>>C>>B
>>>>>> 15 B>A>>C
>>>>>> 40 C>B>>A
>>>>>>
>>>>>> B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).
>>>>>>
>>>>>> Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.
>>>>>>
>>>>>> 17 A>>B>>C>>D
>>>>>> 17 B>>C>>A>>D
>>>>>> 17 C>>A>>B>>D
>>>>>> 16 D>>A>>B>>C
>>>>>> 16 D>>B>>C>>A
>>>>>> 16 D>>C>>A>>B
>>>>>>
>>>>>> A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.
>>>>>>
>>>>>> 17 A>B>C>>D
>>>>>> 17 B>C>A>>D
>>>>>> 17 C>A>B>>D
>>>>>> 16 D>>A>>B>>C
>>>>>> 16 D>>B>>C>>A
>>>>>> 16 D>>C>>A>>B
>>>>>>
>>>>>> A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.
>>>>>>
>>>>>>
>>>>>> ----
>>>>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>>> ----
>>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>> ----
>>> Election-Methods mailing list - see https://electorama.com/em for list info
>> ----
>> Election-Methods mailing list - see https://electorama.com/em for list info
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info


---
This email has been checked for viruses by AVG.
https://www.avg.com

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

Juho Laatu-4
One key reason to present the MOP-F2 method was that I had been earlier so vague, just describing possible properties and general principles, and not giving anything very concrete to study and comment. A concrete well defined method was thus needed. I tried to make MOP-F2 such that it would contain most of the useful features, but still be as simple as possible. It can be seen as a "basic MOP method". I think it is simple enough to serve also as a practical election method in real elections.

In order to keep MOP-F2 simple, I also assumed only one cutoff to be used. That should be simple enough, and even more importantly, voters can understand and use that "favourite/defended candidates" cutoff very easily. In principle you could use multiple cutoffs or mixture of ">" and ">>" preferences in the vote (in other MOP methods), but for practical purposes, the one cutoff approach is probably a local optimum in the sense that it offers a lot, while being simple and easy to use. I thus agree that having multiple cutoffs would probably be a too complex approach.

James Green-Armytage's CWP uses "0-99 Score ballots" (or something similar). In that method the "strength" of that range is divided between multiple preferences (e.g. between A>B preference and B>C preference of a A>B>C vote). MOP has a different philosophy. The basic approach is that voters are allowed to weaken whatever number of preferences with full force, or to keep them as they are (= "keep them strong"). In MOP-F2 this ability is limited to indicating some of the highest ranked candidates (= above the cutoff) as favourite candidates whose defeats to each other should be weakened. The modification function of MOP-F2 uses the proportion of votes that proposed to weaken each preference (to modify the defeat strengths).

> You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
> is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.

I have't yet given much thought on how MOP-F2 would behave here. I'll think that a bit, and come back if I find something interesting to say.

> I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
> irrespective of whether the votes are "sincere" or not.


Non-Drastic Defense refers to performance with strategic votes. I referred to good properties of margins with sincere votes, i.e. in the absence of strategies. If we study the performance of MOP-F2 with strategic votes, then one should see it as modifying the regular properties of margins by adding the cutoff in the ballot. I think using the cutoff to defend against burying strategy, without modifying the rankings (I wrote about this in an earlier mail), can be said to be an "easy defense" technique. Also here I must say that I must study the relationship of MOP-F2 and the criteria that you mentioned a bit more.

(Plurality is a criterion that I did not even want to put too much weight on because of its emphasis on the implicit cutoff. We discussed this already earlier. I'm however interested in any related criteria that are formulated so that they do not assume that implicit cutoff at the truncation point (or above the last ranked candidate in the case that the voter ranks all the candidates explicitly).)

My emphasis in MOP-F2 is to allow voters to cast fully sincere votes, with the exception that sometimes one might adjust the cutoff point "semi-sincerely" so that one defends those candidates that one needs to defend, even if they would not be ones "true favourites". In practice that means C supporters defending B in the example in the example below ("36: C>B>A --> C>B>>A").

I'll try to think if I can find some vulnerabilities (or strengths) in MOP-F2 on the areas that you mentioned. Please tell me if you can identify some specific problem points.



> On 08 Jul 2019, at 17:36, C.Benham <[hidden email]> wrote:
>
> Juho,
>
> I gather that you are proposing that the voters should sometimes be allowed to give more than one "cutoff" in their rankings.
> I consider that to be far too clumsy and also too much looking like an explicit strategy device.
>
> If you want to do something like that I suggest 0-99 Score ballots be used, and cutoffs and/or pairwise preference strengths can be
> inferred from them.
>
> You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
> is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.
>
>> Margins provide good results with sincere votes, so why not use margins...
> I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
> irrespective of whether the votes are "sincere" or not.
>
> Chris Benham
>
> On 8/07/2019 9:09 pm, Juho Laatu wrote:
>> Few more words on using margins and Minmax in MOP-F2. These words are intended for readers who normally do not like margins, and who prefer methods that always elect the winner from the Smith set.
>>
>> The previous mail already demonstrated that margins can be used, and still provide protection against some strategic attempts. Margins provide good results with sincere votes, so why not use margins, and leave the strategic concerns to the cutoff part of the method.
>>
>> The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed). There are however also cases where the top looped candidates are indeed clones in the sense that all voters consider them to be closely related. In such case their defeats to each others could be considered meaningless, and the winner should be one of them, even if there are candidates outside the Smith set with smaller defeats than the looped candidates have between themselves.
>>
>> MOP-F2 can separate these different cases from each other by using the cutoff. This makes Smith set criterion compatibility a matter of the voters to decide. If voters vote 17: A>B>C>>D, 17: B>C>A>>D, 17: C>A>B>>D, 16: D>A>B>C, 16: D>B>C>A, 16: D>C>A>B, one of A, B and C will win. If A, B and C supporters will not use cutoffs, D wins in MOP-F2.
>>
>> The point here is that use of Minmax as part of MOP-F2 makes it possible to voters to express their opinions more extensively (rankings + favourites) than in methods that use only rankings, and that this approach allows voters to indicate which candidates should be treated as clones and which ones not. Explicitly indicated clones will be treated as one fixed group of candidates (clones), without paying too much attention to their defeats to each others, in the final results.
>>
>> One typical case where looped candidates would not be treated as if they were one candidate (group of clones) is 17: A>B>D>C, 17: B>C>D>A, 17: C>A>D>B, 16: A>D>B>C, 16: B>D>C>A, 16: C>D>A>B. In these votes A, B and C are clearly not clones since they are not next to each others in any of the ballots. D wins because its worst defeat is the smallest. With these votes, voters can not identify A, B and C as one uniform group on favourites (since they are not next to each others). Some voters might however use the cutoffs in some way, and change the result in some direction.
>>
>> Margins and Minmax can thus be used, and the additional cutoffs may provide additional properties that plain margins and plain Minmax do not provide.
>>
>>
>>
>>> On 08 Jul 2019, at 13:36, Juho Laatu <[hidden email]> wrote:
>>>
>>> The MOP approach is quite good in thwarting possible strategic burying attempts. I'll address briefly the basic case of having three candidates that will be looped strategically.
>>>
>>> I'll use one particular MOP method that we might call MOP-F2 (identify Favourites using a cutoff, stretch factor = 2). It will be symmetric (both A>B and B>A will be weakened if A and B are marked as favourites), and it will be simply based on use of margins and Minmax. The modification function will be f(m) = m + sign(m) * (1-P/N) * 2 * N, where N is the number of voters, and P is the number of voters that identified the two candidates as their (Protected) favourites (= candidates above the cutoff). Parameter m is the original margin, and the output of the function will be the modified value to be fed to Minmax. Note that with 100 voters the original margin values are in range [-100, 100], and the modified values are in range [-300, 300].
>>>
>>> I chose stretch factor to be 2, since that gives us some quite natural defensive properties. it seems that when certain number of of voters try to bury the sincere CW to make their favourite the winner (or vote sincerely that way), it is enough if a "corresponding" number of the other voters indicate the other two candidates as their favourites, to thwart the strategic attempt. In the example below 100% of the A supporters are strategic, and 100% of those voters that can easily defend against the strategy will do so. Stretch factor 1 would not give us such "full protection", and stretch factor 3 might be considered already an overkill.
>>>
>>> 49: A>B>>C --> A>C>B large number of strategists that try to bury B
>>> 15: B>A>>C --> B>A>C sincere CW, and a member of the AB group
>>> 36: C>B>A --> C>B>>A the other group that decides to protect the CW
>>>
>>> The opinions on the left are sincere. The actual votes are on the right. Note that all defensive votes (B and C supporters) can maintain the original ranking order. There is thus no need to falsify their rankings. The only required change is that they change the position of their favourite cutoff. B supporters stop protecting A, once they hear about the strategic plans of the A supporters. C supporters extend their protection to cover B (the CW), once they hear about the strategic plans of the A supporters.
>>>
>>> The worst defeats with the original (non-modified) margins are A:-2, B:2 (CW), and C:-28. If A voters apply their strategy, the worst defeat margins become A:-2, B:-70, and C:-28. If the votes are modified, and B supporters still protect A, and C supporters do not protect B, the modified worst defeats become A:-172, B:-270, and C:-228. If B supporters stop protecting A, the modified worst defeats become A:-202, B:-270, and C:-228. And if C supporters protect B, the modified worst defeats become A:-202, B:-198, and C:-228. CW (B) wins again. It seems that one can effectively defend against any such (single) strategic burying attempt. And if you can, maybe nobody ever even tries such strategic attempts. This would make use of the cutoffs even more sincere, and more unnecessary, but maybe still good sincere information for the election analysts.
>>>
>>> I like the idea that rankings of the non-starategic voters may stay the same, and also that the favourite cutoffs may stay quite natural, even when used defensively. The non-strategic voters may use the protective cutoff in a quite natural way, while it is of no use to the strategists. (If the A supporters would vote A>C>>B, they would just support the election of C that is the worst alternative to them.) There is also no need to resort to use of winning votes or something similar to provide means to defend against strategies (margins are a more natural way to measure sincere preferences).
>>>
>>>
>>>
>>>
>>> P.S. I note that in an earlier mail I forgot to include the sign function in one of the the equations. That equation should have been sign(initial)*(1-mod)*k*max + initial. I'm sorry if someone tried to calculate the results and got confused, and didn't dare to ask why that function seemed strange. I hope at least the function I gave above is correct. I think that one (MOP-F2) is a good one to start with, if someone wants to test the MOP methods in practice.
>>>
>>>> On 08 Jul 2019, at 10:06, Juho Laatu <[hidden email]> wrote:
>>>>
>>>> James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.
>>>>
>>>> Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.
>>>>
>>>> Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.
>>>>
>>>> - AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.
>>>>
>>>> - In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.
>>>>
>>>> In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates that
>>   one wan
>>  t
>>> s
>>>> to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".
>>>>
>>>> In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.
>>>>
>>>> I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably not
>>  practica
>>  l
>>> ,
>>>> since they would be just "weakened opinions".)
>>>>
>>>>
>>>>> On 26 Jun 2019, at 10:55, Juho Laatu <[hidden email]> wrote:
>>>>>
>>>>> I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.
>>>>>
>>>>> The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.
>>>>>
>>>>> The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).
>>>>>
>>>>> The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.
>>>>>
>>>>> I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial strength
>>  s) to cr
>>  e
>>> a
>>>> te
>>>>> space also for the weakened preferences.
>>>>>
>>>>> Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.
>>>>>
>>>>> Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.
>>>>>
>>>>> Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.
>>>>>
>>>>> 45: A > B >> C
>>>>> 15: B > A >> C
>>>>> 40: C > B > A
>>>>>
>>>>> A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).
>>>>>
>>>>> 40: A (no "clones")
>>>>> 35: B>>C>A (A and C indicated as "clones")
>>>>> 25: C (no "clones")
>>>>>
>>>>> This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.
>>>>>
>>>>> 40: A
>>>>> 35: B>C>>A
>>>>> 25: C
>>>>>
>>>>> This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.
>>>>>
>>>>> 40: A
>>>>> 35: B>>C>A
>>>>> 25: C>>B>A
>>>>>
>>>>> This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.
>>>>>
>>>>> If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.
>>>>>
>>>>>
>>>>>
>>>>>> On 18 Jun 2019, at 09:01, Juho Laatu <[hidden email]> wrote:
>>>>>>
>>>>>> Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.
>>>>>>
>>>>>>> Forest Simmons fsimmons at pcc.edu
>>>>>>> Thu May 30
>>>>>>>
>>>>>>>> In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??
>>>>>>>>
>>>>>>>> I am interested in simple methods that always ...
>>>>>>>>
>>>>>>>> (1) elect candidate A given the following profile:
>>>>>>>> P: A
>>>>>>>> Q: B>>C
>>>>>>>> R: C,
>>>>>>>>
>>>>>>>> and
>>>>>>>> (2) elect candidate C given
>>>>>>>> P: A
>>>>>>>> Q: B>C>>
>>>>>>>> R: C,
>>>>>>>>
>>>>>>>> and
>>>>>>>> (3) elect candidate B given
>>>>>>>> P: A
>>>>>>>> Q: B>>C?? (or B>C)
>>>>>>>> R: C>>B. (or C>B)
>>>>>>
>>>>>> I copy my answers to those questions below.
>>>>>>
>>>>>> Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).
>>>>>>
>>>>>>> I tried to see how it relates to Forest's requests.
>>>>>>>
>>>>>>> Case (2)
>>>>>>> Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.
>>>>>>>
>>>>>>> Case (1)
>>>>>>> The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.
>>>>>>>
>>>>>>> Case (3)
>>>>>>> Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.
>>>>>>>
>>>>>>> In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
>>>>>> I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.
>>>>>>
>>>>>> P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typical
>>   Condorc
>>  e
>>> t
>>>>> methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
>>>>>> P.P.S. In the original mail, ignore words "which means".
>>>>>>
>>>>>> Juho
>>>>>>
>>>>>>
>>>>>>> On 16 Jun 2019, at 16:51, Juho Laatu <[hidden email]> wrote:
>>>>>>>
>>>>>>> The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.
>>>>>>>
>>>>>>> You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).
>>>>>>>
>>>>>>> In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).
>>>>>>>
>>>>>>> Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).
>>>>>>>
>>>>>>> (I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)
>>>>>>>
>>>>>>> I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.
>>>>>>>
>>>>>>> The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)
>>>>>>>
>>>>>>> Here's one example set of votes.
>>>>>>>
>>>>>>> 45 A>B>>C --> A>>C>>B
>>>>>>> 15 B>A>>C
>>>>>>> 40 C>B>>A
>>>>>>>
>>>>>>> B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).
>>>>>>>
>>>>>>> Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.
>>>>>>>
>>>>>>> 17 A>>B>>C>>D
>>>>>>> 17 B>>C>>A>>D
>>>>>>> 17 C>>A>>B>>D
>>>>>>> 16 D>>A>>B>>C
>>>>>>> 16 D>>B>>C>>A
>>>>>>> 16 D>>C>>A>>B
>>>>>>>
>>>>>>> A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.
>>>>>>>
>>>>>>> 17 A>B>C>>D
>>>>>>> 17 B>C>A>>D
>>>>>>> 17 C>A>B>>D
>>>>>>> 16 D>>A>>B>>C
>>>>>>> 16 D>>B>>C>>A
>>>>>>> 16 D>>C>>A>>B
>>>>>>>
>>>>>>> A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.
>>>>>>>
>>>>>>>
>>>>>>> ----
>>>>>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>>>> ----
>>>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>>> ----
>>>> Election-Methods mailing list - see https://electorama.com/em for list info
>>> ----
>>> Election-Methods mailing list - see https://electorama.com/em for list info
>> ----
>> Election-Methods mailing list - see https://electorama.com/em for list info
>
>
> ---
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] MOP-F2 / Non-Drastic Defense

Juho Laatu-4
In reply to this post by C.Benham
> On 08 Jul 2019, at 17:36, C.Benham <[hidden email]> wrote:

>> Margins provide good results with sincere votes, so why not use margins...
> I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
> irrespective of whether the votes are "sincere" or not.

I googled the Non-Drastic Defense criterion and this example.

46 A>C
10 B>A
10 B>C
34 C=B

Non-Drastic Defense criterion:    if on more than half the ballots X is voted both above Y and below no other candidate (i.e. no lower than equal-top) then Y must not win

The Non-Drastic Defense criterion says that A should not be elected.

If no cutoffs are used, MOP-F2 is the same as the base method, i.e. Minmax(margins), and elects A. But if those 34 voters that seem to consider C and B to be clones indicate this in their votes (as they probably should) by voting C=B>>, C>B>>, or B>C>>, then B wins.

I guess it would be ok in MOP-F2 not to support use of "=" in the vote (explicitly) since in this case since they can vote also C>B>>, or B>C>>, and by doing this, indicate that these two candidates should be seen as clones / their mutual defeats should be seen as weak defeats.
----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

C.Benham
In reply to this post by Juho Laatu-4

Non-Drastic Defense refers to performance with strategic votes.
http://mam-docs.000webhostapp.com/
non-drastic defense:  If more than half of the voters prefer alternative y 
        over alternative x, then that majority must have some way of voting 
        that ensures x will not be elected and does not require any of them to
        rank y over any more-preferred alternatives. (This has been promoted
        by Mike Ossipoff under the name Weak Defensive Strategy Criterion.  
        Non-satisfaction means some members of the majority may need to 
        misrepresent their preferences by voting a compromise alternative 
        over favored alternatives if they want to ensure the defeat of less-
        preferred alternatives.) 

Juho, I can see how you might get that impression but it doesn't follow that there is anything desirable about
failing it if the votes aren't  strategic.

That hyperlink is now dead for me. There used to be an "alternative wording" that was specific about what was
on the ballots. That was the one that I had in mind. It said that if on more than half the the ballots Y is voted over
X and not below equal-top, then the winner can't be X.

That strikes me as a very reasonable requirement whether the votes are sincere or not, and any failure would look
very odd and unjustified.

46: A>C
10: B>A
10: B>C
34: C=B

B>A  54-46 (margin=8),   A>C 56-44 (margin=12),  C>B 46-20 (margin=26).

Here on more than half the ballots B is voted both above A and no lower than equal-top, and yet Margins elects A.

No other candidate is voted at least equal-top on more than half the ballots. Fans of Bucklin or any version of Median
Ratings would be particularly incensed. 

The version of Losing Votes that I advocate elects B because it has the 34C=B ballots give a whole vote to each of C and B
in their pairwise comparison, so "C>B 46-20" becomes C>B 80-54, giving B a winning MM Losing Votes score of 54. 

(With this LV(erw) measure of defeat-strength I am happy with any of  Smith//MinMax, Schulze, Ranked Pairs, River.)

I referred to good properties of margins with sincere votes, i.e. in the absence of strategies.
What "good properties" ?

Plurality is a criterion that I did not even want to put too much weight on because of its emphasis on the implicit cutoff. We discussed this already earlier.
Yes, so I don't know why you keep repeating this vague, stupid and (at best) misleading objection.

The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed).
I find the idea that electing the Condorcet winner is essential but if the Smith set has more than one member then maybe
it's a good idea to (without some very strong excuse) elect someone outside outside of it to be a philosophical absurdity.

Comparing Smith//MinMax (Margins) with the plain MMM that you prefer, the Smith version loses compliance with Mono-add-Top
but I don't recall you ever espousing that criterion as being particularly valuable or important.

My emphasis in MOP-F2 is to allow voters to cast fully sincere votes, with the exception that sometimes one might adjust the cutoff point "semi-sincerely" so that one defends those candidates that one needs to defend, even if they would not be ones "true favourites".
I agree that is a desirable goal.  What do you think is wrong with Approval Margins in this respect? 

Chris Benham

One key reason to present the MOP-F2 method was that I had been earlier so vague, just describing possible properties and general principles, and not giving anything very concrete to study and comment. A concrete well defined method was thus needed. I tried to make MOP-F2 such that it would contain most of the useful features, but still be as simple as possible. It can be seen as a "basic MOP method". I think it is simple enough to serve also as a practical election method in real elections.

In order to keep MOP-F2 simple, I also assumed only one cutoff to be used. That should be simple enough, and even more importantly, voters can understand and use that "favourite/defended candidates" cutoff very easily. In principle you could use multiple cutoffs or mixture of ">" and ">>" preferences in the vote (in other MOP methods), but for practical purposes, the one cutoff approach is probably a local optimum in the sense that it offers a lot, while being simple and easy to use. I thus agree that having multiple cutoffs would probably be a too complex approach.

James Green-Armytage's CWP uses "0-99 Score ballots" (or something similar). In that method the "strength" of that range is divided between multiple preferences (e.g. between A>B preference and B>C preference of a A>B>C vote). MOP has a different philosophy. The basic approach is that voters are allowed to weaken whatever number of preferences with full force, or to keep them as they are (= "keep them strong"). In MOP-F2 this ability is limited to indicating some of the highest ranked candidates (= above the cutoff) as favourite candidates whose defeats to each other should be weakened. The modification function of MOP-F2 uses the proportion of votes that proposed to weaken each preference (to modify the defeat strengths).

You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.
I have't yet given much thought on how MOP-F2 would behave here. I'll think that a bit, and come back if I find something interesting to say.

I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
irrespective of whether the votes are "sincere" or not.

Non-Drastic Defense refers to performance with strategic votes. I referred to good properties of margins with sincere votes, i.e. in the absence of strategies. If we study the performance of MOP-F2 with strategic votes, then one should see it as modifying the regular properties of margins by adding the cutoff in the ballot. I think using the cutoff to defend against burying strategy, without modifying the rankings (I wrote about this in an earlier mail), can be said to be an "easy defense" technique. Also here I must say that I must study the relationship of MOP-F2 and the criteria that you mentioned a bit more.

(Plurality is a criterion that I did not even want to put too much weight on because of its emphasis on the implicit cutoff. We discussed this already earlier. I'm however interested in any related criteria that are formulated so that they do not assume that implicit cutoff at the truncation point (or above the last ranked candidate in the case that the voter ranks all the candidates explicitly).)

My emphasis in MOP-F2 is to allow voters to cast fully sincere votes, with the exception that sometimes one might adjust the cutoff point "semi-sincerely" so that one defends those candidates that one needs to defend, even if they would not be ones "true favourites". In practice that means C supporters defending B in the example in the example below ("36: C>B>A --> C>B>>A").

I'll try to think if I can find some vulnerabilities (or strengths) in MOP-F2 on the areas that you mentioned. Please tell me if you can identify some specific problem points.



On 08 Jul 2019, at 17:36, C.Benham [hidden email] wrote:

Juho,

I gather that you are proposing that the voters should sometimes be allowed to give more than one "cutoff" in their rankings.
I consider that to be far too clumsy and also too much looking like an explicit strategy device.

If you want to do something like that I suggest 0-99 Score ballots be used, and cutoffs and/or pairwise preference strengths can be
inferred from them.

You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.

Margins provide good results with sincere votes, so why not use margins...
I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
irrespective of whether the votes are "sincere" or not.

Chris Benham

On 8/07/2019 9:09 pm, Juho Laatu wrote:
Few more words on using margins and Minmax in MOP-F2. These words are intended for readers who normally do not like margins, and who prefer methods that always elect the winner from the Smith set.

The previous mail already demonstrated that margins can be used, and still provide protection against some strategic attempts. Margins provide good results with sincere votes, so why not use margins, and leave the strategic concerns to the cutoff part of the method.

The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed). There are however also cases where the top looped candidates are indeed clones in the sense that all voters consider them to be closely related. In such case their defeats to each others could be considered meaningless, and the winner should be one of them, even if there are candidates outside the Smith set with smaller defeats than the looped candidates have between themselves.

MOP-F2 can separate these different cases from each other by using the cutoff. This makes Smith set criterion compatibility a matter of the voters to decide. If voters vote 17: A>B>C>>D, 17: B>C>A>>D, 17: C>A>B>>D, 16: D>A>B>C, 16: D>B>C>A, 16: D>C>A>B, one of A, B and C will win. If A, B and C supporters will not use cutoffs, D wins in MOP-F2.

The point here is that use of Minmax as part of MOP-F2 makes it possible to voters to express their opinions more extensively (rankings + favourites) than in methods that use only rankings, and that this approach allows voters to indicate which candidates should be treated as clones and which ones not. Explicitly indicated clones will be treated as one fixed group of candidates (clones), without paying too much attention to their defeats to each others, in the final results.

One typical case where looped candidates would not be treated as if they were one candidate (group of clones) is 17: A>B>D>C, 17: B>C>D>A, 17: C>A>D>B, 16: A>D>B>C, 16: B>D>C>A, 16: C>D>A>B. In these votes A, B and C are clearly not clones since they are not next to each others in any of the ballots. D wins because its worst defeat is the smallest. With these votes, voters can not identify A, B and C as one uniform group on favourites (since they are not next to each others). Some voters might however use the cutoffs in some way, and change the result in some direction.

Margins and Minmax can thus be used, and the additional cutoffs may provide additional properties that plain margins and plain Minmax do not provide.



On 08 Jul 2019, at 13:36, Juho Laatu [hidden email] wrote:

The MOP approach is quite good in thwarting possible strategic burying attempts. I'll address briefly the basic case of having three candidates that will be looped strategically.

I'll use one particular MOP method that we might call MOP-F2 (identify Favourites using a cutoff, stretch factor = 2). It will be symmetric (both A>B and B>A will be weakened if A and B are marked as favourites), and it will be simply based on use of margins and Minmax. The modification function will be f(m) = m + sign(m) * (1-P/N) * 2 * N, where N is the number of voters, and P is the number of voters that identified the two candidates as their (Protected) favourites (= candidates above the cutoff). Parameter m is the original margin, and the output of the function will be the modified value to be fed to Minmax. Note that with 100 voters the original margin values are in range [-100, 100], and the modified values are in range [-300, 300].

I chose stretch factor to be 2, since that gives us some quite natural defensive properties. it seems that when certain number of of voters try to bury the sincere CW to make their favourite the winner (or vote sincerely that way), it is enough if a "corresponding" number of the other voters indicate the other two candidates as their favourites, to thwart the strategic attempt. In the example below 100% of the A supporters are strategic, and 100% of those voters that can easily defend against the strategy will do so. Stretch factor 1 would not give us such "full protection", and stretch factor 3 might be considered already an overkill.

49: A>B>>C --> A>C>B		large number of strategists that try to bury B
15: B>A>>C --> B>A>C		sincere CW, and a member of the AB group
36: C>B>A --> C>B>>A		the other group that decides to protect the CW

The opinions on the left are sincere. The actual votes are on the right. Note that all defensive votes (B and C supporters) can maintain the original ranking order. There is thus no need to falsify their rankings. The only required change is that they change the position of their favourite cutoff. B supporters stop protecting A, once they hear about the strategic plans of the A supporters. C supporters extend their protection to cover B (the CW), once they hear about the strategic plans of the A supporters.

The worst defeats with the original (non-modified) margins are A:-2, B:2 (CW), and C:-28. If A voters apply their strategy, the worst defeat margins become A:-2, B:-70, and C:-28. If the votes are modified, and B supporters still protect A, and C supporters do not protect B, the modified worst defeats become A:-172, B:-270, and C:-228. If B supporters stop protecting A, the modified worst defeats become A:-202, B:-270, and C:-228. And if C supporters protect B, the modified worst defeats become A:-202, B:-198, and C:-228. CW (B) wins again. It seems that one can effectively defend against any such (single) strategic burying attempt. And if you can, maybe nobody ever even tries such strategic attempts. This would make use of the cutoffs even more sincere, and more unnecessary, but maybe still good sincere information for the election analysts.

I like the idea that rankings of the non-starategic voters may stay the same, and also that the favourite cutoffs may stay quite natural, even when used defensively. The non-strategic voters may use the protective cutoff in a quite natural way, while it is of no use to the strategists. (If the A supporters would vote A>C>>B, they would just support the election of C that is the worst alternative to them.) There is also no need to resort to use of winning votes or something similar to provide means to defend against strategies (margins are a more natural way to measure sincere preferences).




P.S. I note that in an earlier mail I forgot to include the sign function in one of the the equations. That equation should have been sign(initial)*(1-mod)*k*max + initial. I'm sorry if someone tried to calculate the results and got confused, and didn't dare to ask why that function seemed strange. I hope at least the function I gave above is correct. I think that one (MOP-F2) is a good one to start with, if someone wants to test the MOP methods in practice.

On 08 Jul 2019, at 10:06, Juho Laatu [hidden email] wrote:

James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.

Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.

Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.

- AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.

- In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.

In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates th
 at
  one wan
 t
s
to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".

In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.

I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably no
 t
 practica
 l
,
since they would be just "weakened opinions".)


On 26 Jun 2019, at 10:55, Juho Laatu [hidden email] wrote:

I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.

The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.

The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).

The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.

I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial streng
 th
 s) to cr
 e
a
te
space also for the weakened preferences.

Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.

Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.

Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.

45: A > B >> C
15: B > A >> C
40: C > B > A

A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).

40: A		(no "clones")
35: B>>C>A	(A and C indicated as "clones")
25: C		(no "clones")

This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.

40: A
35: B>C>>A
25: C

This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.

40: A
35: B>>C>A
25: C>>B>A

This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.

If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.



On 18 Jun 2019, at 09:01, Juho Laatu [hidden email] wrote:

Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.

Forest Simmons fsimmons at pcc.edu
Thu May 30

In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??

I am interested in simple methods that always ...

(1) elect candidate A given the following profile:
P: A
Q: B>>C
R: C,

and
(2) elect candidate C given
P: A
Q: B>C>>
R: C,

and
(3) elect candidate B given
P: A
Q: B>>C?? (or B>C)
R: C>>B. (or C>B)
I copy my answers to those questions below.

Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).

I tried to see how it relates to Forest's requests.

Case (2)
Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.

Case (1)
The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.

Case (3)
Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.

In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.

P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typic
 al
  Condorc
 e
t
methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
P.P.S. In the original mail, ignore words "which means".

Juho


On 16 Jun 2019, at 16:51, Juho Laatu [hidden email] wrote:

The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.

You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).

In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).

Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).

(I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)

I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.

The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)

Here's one example set of votes.

45 A>B>>C --> A>>C>>B
15 B>A>>C
40 C>B>>A

B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).

Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.

17 A>>B>>C>>D
17 B>>C>>A>>D
17 C>>A>>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.

17 A>B>C>>D
17 B>C>A>>D
17 C>A>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.


----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info

---
This email has been checked for viruses by AVG.
https://www.avg.com

----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info

Virus-free. www.avg.com

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

Juho Laatu-4
On 08 Jul 2019, at 22:05, C.Benham <[hidden email]> wrote:


Non-Drastic Defense refers to performance with strategic votes.
http://mam-docs.000webhostapp.com/
non-drastic defense:  If more than half of the voters prefer alternative y 
        over alternative x, then that majority must have some way of voting 
        that ensures x will not be elected and does not require any of them to
        rank y over any more-preferred alternatives. (This has been promoted
        by Mike Ossipoff under the name Weak Defensive Strategy Criterion.  
        Non-satisfaction means some members of the majority may need to 
        misrepresent their preferences by voting a compromise alternative 
        over favored alternatives if they want to ensure the defeat of less-
        preferred alternatives.) 

Juho, I can see how you might get that impression but it doesn't follow that there is anything desirable about
failing it if the votes aren't  strategic.

Yes, I can see that there are also some sincere voting related aspects in the definition. The version that I used in my other mail ("[EM] MOP-F2 / Non-Drastic Defense") did not mention strategies explicitly at all. This version (see above) seems to be quite different (btw, I didn't like the special role of first position there, just like I don't like the special role of the truncation point or last position). The electowiki definition is also slightly different (https://electowiki.org/wiki/Weak_Defensive_Strategy_criterion).


That hyperlink is now dead for me. There used to be an "alternative wording" that was specific about what was
on the ballots. That was the one that I had in mind. It said that if on more than half the the ballots Y is voted over
X and not below equal-top, then the winner can't be X.



That strikes me as a very reasonable requirement whether the votes are sincere or not, and any failure would look
very odd and unjustified.

As already noted, I don't like the special role of first position preferences very much, when talking about (pure) ranked methods. I'd like to talk about majority opinions (that would be independent of whether some candidates have been ranked higher). If one takes the first position aspect out of that definition, it makes less sense. To me later preferences are as important as first position related preferences (in "pure ranked methods").


46: A>C
10: B>A
10: B>C
34: C=B

B>A  54-46 (margin=8),   A>C 56-44 (margin=12),  C>B 46-20 (margin=26).

Here on more than half the ballots B is voted both above A and no lower than equal-top, and yet Margins elects A.

No other candidate is voted at least equal-top on more than half the ballots. Fans of Bucklin or any version of Median
Ratings would be particularly incensed. 

The version of Losing Votes that I advocate elects B because it has the 34C=B ballots give a whole vote to each of C and B
in their pairwise comparison, so "C>B 46-20" becomes C>B 80-54, giving B a winning MM Losing Votes score of 54. 

I don't like losing votes too much since they may give strange results already with sincere votes. I don't have any dramatic examples available, but I believe there are such examples. I gave one example on winning votes earlier (50: A>B>C, 50: D>E>F, 1: F>A). Sorry about basing my claim on losing votes on "feelings only".

(With this LV(erw) measure of defeat-strength I am happy with any of  Smith//MinMax, Schulze, Ranked Pairs, River.)

I referred to good properties of margins with sincere votes, i.e. in the absence of strategies.
What "good properties" ?

Natural behaviour when compared e.g. to winning votes (see the example few lines above). The "least additional votes to become a Condorcet winner" criterion / description of the best candidate sounds quite natural to me. There could be also other natural measures of which candidate is best with sincere votes, but this one is certainly one of the obvious and natural ones. (I recently discussed also some other quite natural preference functions in http://lists.electorama.com/pipermail/election-methods-electorama.com/2019-May/002126.html.)


Plurality is a criterion that I did not even want to put too much weight on because of its emphasis on the implicit cutoff. We discussed this already earlier.
Yes, so I don't know why you keep repeating this vague, stupid and (at best) misleading objection.

I like approaches where all preferences are treated the same way, irrespective of if they are given at the beginning or at the end of a single vote. There could be also reasonable methods where different positions in the ballot are treated in different ways. But as long as we talk about (pure) rankings with no additional "positional weights" I think criteria that refer to the positions are in some sense out of scope. I thus don't oppose some methods taking candidate positions in the ballot into account, but when one does so, one should explicitly state that this is how the preferences are to be treated / weighted.

Sometimes criteria that refer to specific positions in the ballot could be useful also for methods that treat all positions in the same way. But in those cases I'd prefer to use alternative criteria that do not refer to those positions in the first place.


The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed).
I find the idea that electing the Condorcet winner is essential but if the Smith set has more than one member then maybe
it's a good idea to (without some very strong excuse) elect someone outside outside of it to be a philosophical absurdity.

I tend to think that group opinions may be cyclic, and that is natural, and the idea of seeing the Smith set as a set of tied candidates is flawed. That unjustified thought would lead to the idea of always electing one of the members of the Smith set. One key problem with that thought is that it ignores the fact that candidates within Smith set may have defeats to each other, and those defeats may be strong defeats. Candidates outside the Smith set may on the other hand have only very weak defeats. Most people agree that the strength of defeats is important. There is no point in giving the candidates in the Smith set a "free ride".

People may also want election methods to find a social order. For single winner methods it is enough to find the winner. But establishing a full social order is possible too. One should however not assume that the visual image of having a Smith set above the other candidates to set a requirement that the social order should be such that the Smith set members will be higher in the order than other candidates. An example of an order that does not respect that idea is the order determined by "least additional votes to become a Condorcet winner". For some societies and elections that criterion might be useful, and therefore methods that follow the "Smith set members first" approach would not be acceptable.

I hope these words made my approach clearer. I thus do not consider Smith set members to be "ahead of others" although the visual image of seeing Smith set members as a set of "clones" that occupies one position in a linear ranking might suggest so. I do not believe in "breaking" such "ties". I do not believe in forcing group opinions to become linear in such a way that all cycles are "broken" and kept in places suggested by the visual image of situation where those cycles are still intact. That approach treats defeats within cycles in a totally different way than defeats that are not part of such cycles. To me that approach looks like an attempt to make (=force) cyclic group opinions linear in a way that is not always valid. The problem of cyclic group opinions is more complex than that. Group opinions are sometimes cyclic, and they should be treated as such. Making them linear, or seeing partly cyclic preferences as being partly linear (partly fixing the social order) is a too strong simplification and an easy way out of a more complex problem.

Here's a classic example of a set of votes where candidate D is outside of the Smith set, but since it is two votes short of being a Condorcet winner, it can be seen to be a very good candidate and a potential winner. The defeats of other candidates are considerably stronger (although they lose to a smaller number of other candidates). It would not make sense to say that no ("pure" ranked) method should elect D with these votes.

17: A>B>D>C
17: B>C>D>A
17: C>A>D>B
16: A>D>B>C
16: B>D>C>A
16: C>D>A>B


Comparing Smith//MinMax (Margins) with the plain MMM that you prefer, the Smith version loses compliance with Mono-add-Top
but I don't recall you ever espousing that criterion as being particularly valuable or important.

I accept the fact that many nice sounding criteria are incompatible with each others. Group opinions are mathematically complex, and we must accept the fact that often our first impressions, that are on most cases based on simpler models, like individual preferences that we can assume to be linear, may be simply wrong when we deal with more complex topics like group opinions. For example the idea that group opinions may have cycles sounds first irrational to us, but when we learn to live with that idea, and understand the underlying principles, we slowly learn to think that such cycles can be natural too. Same with electing outside of the Smith set vs trying to break those "ties" to linear orders.


My emphasis in MOP-F2 is to allow voters to cast fully sincere votes, with the exception that sometimes one might adjust the cutoff point "semi-sincerely" so that one defends those candidates that one needs to defend, even if they would not be ones "true favourites".
I agree that is a desirable goal.  What do you think is wrong with Approval Margins in this respect? 

Ah, I'm not quite prepared to answer that question. My first feeling is that since MOP-F2 looks quite natural to me, other candidates would have to be quite good to meet that level :-). I'm not quite sure what method you refer to. Maybe a link would help.

Juho



Chris Benham


One key reason to present the MOP-F2 method was that I had been earlier so vague, just describing possible properties and general principles, and not giving anything very concrete to study and comment. A concrete well defined method was thus needed. I tried to make MOP-F2 such that it would contain most of the useful features, but still be as simple as possible. It can be seen as a "basic MOP method". I think it is simple enough to serve also as a practical election method in real elections.

In order to keep MOP-F2 simple, I also assumed only one cutoff to be used. That should be simple enough, and even more importantly, voters can understand and use that "favourite/defended candidates" cutoff very easily. In principle you could use multiple cutoffs or mixture of ">" and ">>" preferences in the vote (in other MOP methods), but for practical purposes, the one cutoff approach is probably a local optimum in the sense that it offers a lot, while being simple and easy to use. I thus agree that having multiple cutoffs would probably be a too complex approach.

James Green-Armytage's CWP uses "0-99 Score ballots" (or something similar). In that method the "strength" of that range is divided between multiple preferences (e.g. between A>B preference and B>C preference of a A>B>C vote). MOP has a different philosophy. The basic approach is that voters are allowed to weaken whatever number of preferences with full force, or to keep them as they are (= "keep them strong"). In MOP-F2 this ability is limited to indicating some of the highest ranked candidates (= above the cutoff) as favourite candidates whose defeats to each other should be weakened. The modification function of MOP-F2 uses the proportion of votes that proposed to weaken each preference (to modify the defeat strengths).

You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.
I have't yet given much thought on how MOP-F2 would behave here. I'll think that a bit, and come back if I find something interesting to say.

I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
irrespective of whether the votes are "sincere" or not.
Non-Drastic Defense refers to performance with strategic votes. I referred to good properties of margins with sincere votes, i.e. in the absence of strategies. If we study the performance of MOP-F2 with strategic votes, then one should see it as modifying the regular properties of margins by adding the cutoff in the ballot. I think using the cutoff to defend against burying strategy, without modifying the rankings (I wrote about this in an earlier mail), can be said to be an "easy defense" technique. Also here I must say that I must study the relationship of MOP-F2 and the criteria that you mentioned a bit more.

(Plurality is a criterion that I did not even want to put too much weight on because of its emphasis on the implicit cutoff. We discussed this already earlier. I'm however interested in any related criteria that are formulated so that they do not assume that implicit cutoff at the truncation point (or above the last ranked candidate in the case that the voter ranks all the candidates explicitly).)

My emphasis in MOP-F2 is to allow voters to cast fully sincere votes, with the exception that sometimes one might adjust the cutoff point "semi-sincerely" so that one defends those candidates that one needs to defend, even if they would not be ones "true favourites". In practice that means C supporters defending B in the example in the example below ("36: C>B>A --> C>B>>A").

I'll try to think if I can find some vulnerabilities (or strengths) in MOP-F2 on the areas that you mentioned. Please tell me if you can identify some specific problem points.



On 08 Jul 2019, at 17:36, C.Benham [hidden email] wrote:

Juho,

I gather that you are proposing that the voters should sometimes be allowed to give more than one "cutoff" in their rankings.
I consider that to be far too clumsy and also too much looking like an explicit strategy device.

If you want to do something like that I suggest 0-99 Score ballots be used, and cutoffs and/or pairwise preference strengths can be
inferred from them.

You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.

Margins provide good results with sincere votes, so why not use margins...
I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
irrespective of whether the votes are "sincere" or not.

Chris Benham

On 8/07/2019 9:09 pm, Juho Laatu wrote:
Few more words on using margins and Minmax in MOP-F2. These words are intended for readers who normally do not like margins, and who prefer methods that always elect the winner from the Smith set.

The previous mail already demonstrated that margins can be used, and still provide protection against some strategic attempts. Margins provide good results with sincere votes, so why not use margins, and leave the strategic concerns to the cutoff part of the method.

The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed). There are however also cases where the top looped candidates are indeed clones in the sense that all voters consider them to be closely related. In such case their defeats to each others could be considered meaningless, and the winner should be one of them, even if there are candidates outside the Smith set with smaller defeats than the looped candidates have between themselves.

MOP-F2 can separate these different cases from each other by using the cutoff. This makes Smith set criterion compatibility a matter of the voters to decide. If voters vote 17: A>B>C>>D, 17: B>C>A>>D, 17: C>A>B>>D, 16: D>A>B>C, 16: D>B>C>A, 16: D>C>A>B, one of A, B and C will win. If A, B and C supporters will not use cutoffs, D wins in MOP-F2.

The point here is that use of Minmax as part of MOP-F2 makes it possible to voters to express their opinions more extensively (rankings + favourites) than in methods that use only rankings, and that this approach allows voters to indicate which candidates should be treated as clones and which ones not. Explicitly indicated clones will be treated as one fixed group of candidates (clones), without paying too much attention to their defeats to each others, in the final results.

One typical case where looped candidates would not be treated as if they were one candidate (group of clones) is 17: A>B>D>C, 17: B>C>D>A, 17: C>A>D>B, 16: A>D>B>C, 16: B>D>C>A, 16: C>D>A>B. In these votes A, B and C are clearly not clones since they are not next to each others in any of the ballots. D wins because its worst defeat is the smallest. With these votes, voters can not identify A, B and C as one uniform group on favourites (since they are not next to each others). Some voters might however use the cutoffs in some way, and change the result in some direction.

Margins and Minmax can thus be used, and the additional cutoffs may provide additional properties that plain margins and plain Minmax do not provide.



On 08 Jul 2019, at 13:36, Juho Laatu [hidden email] wrote:

The MOP approach is quite good in thwarting possible strategic burying attempts. I'll address briefly the basic case of having three candidates that will be looped strategically.

I'll use one particular MOP method that we might call MOP-F2 (identify Favourites using a cutoff, stretch factor = 2). It will be symmetric (both A>B and B>A will be weakened if A and B are marked as favourites), and it will be simply based on use of margins and Minmax. The modification function will be f(m) = m + sign(m) * (1-P/N) * 2 * N, where N is the number of voters, and P is the number of voters that identified the two candidates as their (Protected) favourites (= candidates above the cutoff). Parameter m is the original margin, and the output of the function will be the modified value to be fed to Minmax. Note that with 100 voters the original margin values are in range [-100, 100], and the modified values are in range [-300, 300].

I chose stretch factor to be 2, since that gives us some quite natural defensive properties. it seems that when certain number of of voters try to bury the sincere CW to make their favourite the winner (or vote sincerely that way), it is enough if a "corresponding" number of the other voters indicate the other two candidates as their favourites, to thwart the strategic attempt. In the example below 100% of the A supporters are strategic, and 100% of those voters that can easily defend against the strategy will do so. Stretch factor 1 would not give us such "full protection", and stretch factor 3 might be considered already an overkill.

49: A>B>>C --> A>C>B		large number of strategists that try to bury B
15: B>A>>C --> B>A>C		sincere CW, and a member of the AB group
36: C>B>A --> C>B>>A		the other group that decides to protect the CW

The opinions on the left are sincere. The actual votes are on the right. Note that all defensive votes (B and C supporters) can maintain the original ranking order. There is thus no need to falsify their rankings. The only required change is that they change the position of their favourite cutoff. B supporters stop protecting A, once they hear about the strategic plans of the A supporters. C supporters extend their protection to cover B (the CW), once they hear about the strategic plans of the A supporters.

The worst defeats with the original (non-modified) margins are A:-2, B:2 (CW), and C:-28. If A voters apply their strategy, the worst defeat margins become A:-2, B:-70, and C:-28. If the votes are modified, and B supporters still protect A, and C supporters do not protect B, the modified worst defeats become A:-172, B:-270, and C:-228. If B supporters stop protecting A, the modified worst defeats become A:-202, B:-270, and C:-228. And if C supporters protect B, the modified worst defeats become A:-202, B:-198, and C:-228. CW (B) wins again. It seems that one can effectively defend against any such (single) strategic burying attempt. And if you can, maybe nobody ever even tries such strategic attempts. This would make use of the cutoffs even more sincere, and more unnecessary, but maybe still good sincere information for the election analysts.

I like the idea that rankings of the non-starategic voters may stay the same, and also that the favourite cutoffs may stay quite natural, even when used defensively. The non-strategic voters may use the protective cutoff in a quite natural way, while it is of no use to the strategists. (If the A supporters would vote A>C>>B, they would just support the election of C that is the worst alternative to them.) There is also no need to resort to use of winning votes or something similar to provide means to defend against strategies (margins are a more natural way to measure sincere preferences).




P.S. I note that in an earlier mail I forgot to include the sign function in one of the the equations. That equation should have been sign(initial)*(1-mod)*k*max + initial. I'm sorry if someone tried to calculate the results and got confused, and didn't dare to ask why that function seemed strange. I hope at least the function I gave above is correct. I think that one (MOP-F2) is a good one to start with, if someone wants to test the MOP methods in practice.

On 08 Jul 2019, at 10:06, Juho Laatu [hidden email] wrote:

James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.

Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.

Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.

- AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.

- In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.

In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates th
 at
  one wan
 t
s
to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".

In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.

I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably no
 t
 practica
 l
,
since they would be just "weakened opinions".)


On 26 Jun 2019, at 10:55, Juho Laatu [hidden email] wrote:

I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.

The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.

The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).

The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.

I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial streng
 th
 s) to cr
 e
a
te
space also for the weakened preferences.

Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.

Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.

Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.

45: A > B >> C
15: B > A >> C
40: C > B > A

A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).

40: A		(no "clones")
35: B>>C>A	(A and C indicated as "clones")
25: C		(no "clones")

This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.

40: A
35: B>C>>A
25: C

This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.

40: A
35: B>>C>A
25: C>>B>A

This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.

If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.



On 18 Jun 2019, at 09:01, Juho Laatu [hidden email] wrote:

Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.

Forest Simmons fsimmons at pcc.edu
Thu May 30

In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??

I am interested in simple methods that always ...

(1) elect candidate A given the following profile:
P: A
Q: B>>C
R: C,

and
(2) elect candidate C given
P: A
Q: B>C>>
R: C,

and
(3) elect candidate B given
P: A
Q: B>>C?? (or B>C)
R: C>>B. (or C>B)
I copy my answers to those questions below.

Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).

I tried to see how it relates to Forest's requests.

Case (2)
Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.

Case (1)
The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.

Case (3)
Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.

In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.

P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typic
 al
  Condorc
 e
t
methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
P.P.S. In the original mail, ignore words "which means".

Juho


On 16 Jun 2019, at 16:51, Juho Laatu [hidden email] wrote:

The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.

You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).

In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).

Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).

(I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)

I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.

The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)

Here's one example set of votes.

45 A>B>>C --> A>>C>>B
15 B>A>>C
40 C>B>>A

B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).

Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.

17 A>>B>>C>>D
17 B>>C>>A>>D
17 C>>A>>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.

17 A>B>C>>D
17 B>C>A>>D
17 C>A>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.


----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
---
This email has been checked for viruses by AVG.
https://www.avg.com

----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info

Virus-free. www.avg.com
<a href="x-msg://22/#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2" width="1" height="1" class="">
----
Election-Methods mailing list - see https://electorama.com/em for list info


----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

C.Benham

Juho,

"Approval Margins" refers either to Smith//MinMax or River or Schulze or Ranked Pairs on
ballots where voters can indicate an approval cutoff (default is all above-bottom candidates
approved) using the margins between the approval scores as the measure of "defeat strength",
or to the Approval-Sorted Margins method.

They are all very similar to each other and always give the same result when there are three
candidates.

https://wiki.electorama.com/wiki/Approval_Sorted_Margins

More later.

Chris Benham

On 9/07/2019 7:25 am, Juho Laatu wrote:
On 08 Jul 2019, at 22:05, C.Benham <[hidden email]> wrote:


Non-Drastic Defense refers to performance with strategic votes.
http://mam-docs.000webhostapp.com/
non-drastic defense:  If more than half of the voters prefer alternative y 
        over alternative x, then that majority must have some way of voting 
        that ensures x will not be elected and does not require any of them to
        rank y over any more-preferred alternatives. (This has been promoted
        by Mike Ossipoff under the name Weak Defensive Strategy Criterion.  
        Non-satisfaction means some members of the majority may need to 
        misrepresent their preferences by voting a compromise alternative 
        over favored alternatives if they want to ensure the defeat of less-
        preferred alternatives.) 

Juho, I can see how you might get that impression but it doesn't follow that there is anything desirable about
failing it if the votes aren't  strategic.

Yes, I can see that there are also some sincere voting related aspects in the definition. The version that I used in my other mail ("[EM] MOP-F2 / Non-Drastic Defense") did not mention strategies explicitly at all. This version (see above) seems to be quite different (btw, I didn't like the special role of first position there, just like I don't like the special role of the truncation point or last position). The electowiki definition is also slightly different (https://electowiki.org/wiki/Weak_Defensive_Strategy_criterion).


That hyperlink is now dead for me. There used to be an "alternative wording" that was specific about what was
on the ballots. That was the one that I had in mind. It said that if on more than half the the ballots Y is voted over
X and not below equal-top, then the winner can't be X.



That strikes me as a very reasonable requirement whether the votes are sincere or not, and any failure would look
very odd and unjustified.

As already noted, I don't like the special role of first position preferences very much, when talking about (pure) ranked methods. I'd like to talk about majority opinions (that would be independent of whether some candidates have been ranked higher). If one takes the first position aspect out of that definition, it makes less sense. To me later preferences are as important as first position related preferences (in "pure ranked methods").


46: A>C
10: B>A
10: B>C
34: C=B

B>A  54-46 (margin=8),   A>C 56-44 (margin=12),  C>B 46-20 (margin=26).

Here on more than half the ballots B is voted both above A and no lower than equal-top, and yet Margins elects A.

No other candidate is voted at least equal-top on more than half the ballots. Fans of Bucklin or any version of Median
Ratings would be particularly incensed. 

The version of Losing Votes that I advocate elects B because it has the 34C=B ballots give a whole vote to each of C and B
in their pairwise comparison, so "C>B 46-20" becomes C>B 80-54, giving B a winning MM Losing Votes score of 54. 

I don't like losing votes too much since they may give strange results already with sincere votes. I don't have any dramatic examples available, but I believe there are such examples. I gave one example on winning votes earlier (50: A>B>C, 50: D>E>F, 1: F>A). Sorry about basing my claim on losing votes on "feelings only".

(With this LV(erw) measure of defeat-strength I am happy with any of  Smith//MinMax, Schulze, Ranked Pairs, River.)

I referred to good properties of margins with sincere votes, i.e. in the absence of strategies.
What "good properties" ?

Natural behaviour when compared e.g. to winning votes (see the example few lines above). The "least additional votes to become a Condorcet winner" criterion / description of the best candidate sounds quite natural to me. There could be also other natural measures of which candidate is best with sincere votes, but this one is certainly one of the obvious and natural ones. (I recently discussed also some other quite natural preference functions in http://lists.electorama.com/pipermail/election-methods-electorama.com/2019-May/002126.html.)


Plurality is a criterion that I did not even want to put too much weight on because of its emphasis on the implicit cutoff. We discussed this already earlier.
Yes, so I don't know why you keep repeating this vague, stupid and (at best) misleading objection.

I like approaches where all preferences are treated the same way, irrespective of if they are given at the beginning or at the end of a single vote. There could be also reasonable methods where different positions in the ballot are treated in different ways. But as long as we talk about (pure) rankings with no additional "positional weights" I think criteria that refer to the positions are in some sense out of scope. I thus don't oppose some methods taking candidate positions in the ballot into account, but when one does so, one should explicitly state that this is how the preferences are to be treated / weighted.

Sometimes criteria that refer to specific positions in the ballot could be useful also for methods that treat all positions in the same way. But in those cases I'd prefer to use alternative criteria that do not refer to those positions in the first place.


The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed).
I find the idea that electing the Condorcet winner is essential but if the Smith set has more than one member then maybe
it's a good idea to (without some very strong excuse) elect someone outside outside of it to be a philosophical absurdity.

I tend to think that group opinions may be cyclic, and that is natural, and the idea of seeing the Smith set as a set of tied candidates is flawed. That unjustified thought would lead to the idea of always electing one of the members of the Smith set. One key problem with that thought is that it ignores the fact that candidates within Smith set may have defeats to each other, and those defeats may be strong defeats. Candidates outside the Smith set may on the other hand have only very weak defeats. Most people agree that the strength of defeats is important. There is no point in giving the candidates in the Smith set a "free ride".

People may also want election methods to find a social order. For single winner methods it is enough to find the winner. But establishing a full social order is possible too. One should however not assume that the visual image of having a Smith set above the other candidates to set a requirement that the social order should be such that the Smith set members will be higher in the order than other candidates. An example of an order that does not respect that idea is the order determined by "least additional votes to become a Condorcet winner". For some societies and elections that criterion might be useful, and therefore methods that follow the "Smith set members first" approach would not be acceptable.

I hope these words made my approach clearer. I thus do not consider Smith set members to be "ahead of others" although the visual image of seeing Smith set members as a set of "clones" that occupies one position in a linear ranking might suggest so. I do not believe in "breaking" such "ties". I do not believe in forcing group opinions to become linear in such a way that all cycles are "broken" and kept in places suggested by the visual image of situation where those cycles are still intact. That approach treats defeats within cycles in a totally different way than defeats that are not part of such cycles. To me that approach looks like an attempt to make (=force) cyclic group opinions linear in a way that is not always valid. The problem of cyclic group opinions is more complex than that. Group opinions are sometimes cyclic, and they should be treated as such. Making them linear, or seeing partly cyclic preferences as being partly linear (partly fixing the social order) is a too strong simplification and an easy way out of a more complex problem.

Here's a classic example of a set of votes where candidate D is outside of the Smith set, but since it is two votes short of being a Condorcet winner, it can be seen to be a very good candidate and a potential winner. The defeats of other candidates are considerably stronger (although they lose to a smaller number of other candidates). It would not make sense to say that no ("pure" ranked) method should elect D with these votes.

17: A>B>D>C
17: B>C>D>A
17: C>A>D>B
16: A>D>B>C
16: B>D>C>A
16: C>D>A>B


Comparing Smith//MinMax (Margins) with the plain MMM that you prefer, the Smith version loses compliance with Mono-add-Top
but I don't recall you ever espousing that criterion as being particularly valuable or important.

I accept the fact that many nice sounding criteria are incompatible with each others. Group opinions are mathematically complex, and we must accept the fact that often our first impressions, that are on most cases based on simpler models, like individual preferences that we can assume to be linear, may be simply wrong when we deal with more complex topics like group opinions. For example the idea that group opinions may have cycles sounds first irrational to us, but when we learn to live with that idea, and understand the underlying principles, we slowly learn to think that such cycles can be natural too. Same with electing outside of the Smith set vs trying to break those "ties" to linear orders.


My emphasis in MOP-F2 is to allow voters to cast fully sincere votes, with the exception that sometimes one might adjust the cutoff point "semi-sincerely" so that one defends those candidates that one needs to defend, even if they would not be ones "true favourites".
I agree that is a desirable goal.  What do you think is wrong with Approval Margins in this respect? 

Ah, I'm not quite prepared to answer that question. My first feeling is that since MOP-F2 looks quite natural to me, other candidates would have to be quite good to meet that level :-). I'm not quite sure what method you refer to. Maybe a link would help.

Juho



Chris Benham


One key reason to present the MOP-F2 method was that I had been earlier so vague, just describing possible properties and general principles, and not giving anything very concrete to study and comment. A concrete well defined method was thus needed. I tried to make MOP-F2 such that it would contain most of the useful features, but still be as simple as possible. It can be seen as a "basic MOP method". I think it is simple enough to serve also as a practical election method in real elections.

In order to keep MOP-F2 simple, I also assumed only one cutoff to be used. That should be simple enough, and even more importantly, voters can understand and use that "favourite/defended candidates" cutoff very easily. In principle you could use multiple cutoffs or mixture of ">" and ">>" preferences in the vote (in other MOP methods), but for practical purposes, the one cutoff approach is probably a local optimum in the sense that it offers a lot, while being simple and easy to use. I thus agree that having multiple cutoffs would probably be a too complex approach.

James Green-Armytage's CWP uses "0-99 Score ballots" (or something similar). In that method the "strength" of that range is divided between multiple preferences (e.g. between A>B preference and B>C preference of a A>B>C vote). MOP has a different philosophy. The basic approach is that voters are allowed to weaken whatever number of preferences with full force, or to keep them as they are (= "keep them strong"). In MOP-F2 this ability is limited to indicating some of the highest ranked candidates (= above the cutoff) as favourite candidates whose defeats to each other should be weakened. The modification function of MOP-F2 uses the proportion of votes that proposed to weaken each preference (to modify the defeat strengths).

You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.
I have't yet given much thought on how MOP-F2 would behave here. I'll think that a bit, and come back if I find something interesting to say.

I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
irrespective of whether the votes are "sincere" or not.
Non-Drastic Defense refers to performance with strategic votes. I referred to good properties of margins with sincere votes, i.e. in the absence of strategies. If we study the performance of MOP-F2 with strategic votes, then one should see it as modifying the regular properties of margins by adding the cutoff in the ballot. I think using the cutoff to defend against burying strategy, without modifying the rankings (I wrote about this in an earlier mail), can be said to be an "easy defense" technique. Also here I must say that I must study the relationship of MOP-F2 and the criteria that you mentioned a bit more.

(Plurality is a criterion that I did not even want to put too much weight on because of its emphasis on the implicit cutoff. We discussed this already earlier. I'm however interested in any related criteria that are formulated so that they do not assume that implicit cutoff at the truncation point (or above the last ranked candidate in the case that the voter ranks all the candidates explicitly).)

My emphasis in MOP-F2 is to allow voters to cast fully sincere votes, with the exception that sometimes one might adjust the cutoff point "semi-sincerely" so that one defends those candidates that one needs to defend, even if they would not be ones "true favourites". In practice that means C supporters defending B in the example in the example below ("36: C>B>A --> C>B>>A").

I'll try to think if I can find some vulnerabilities (or strengths) in MOP-F2 on the areas that you mentioned. Please tell me if you can identify some specific problem points.



On 08 Jul 2019, at 17:36, C.Benham [hidden email] wrote:

Juho,

I gather that you are proposing that the voters should sometimes be allowed to give more than one "cutoff" in their rankings.
I consider that to be far too clumsy and also too much looking like an explicit strategy device.

If you want to do something like that I suggest 0-99 Score ballots be used, and cutoffs and/or pairwise preference strengths can be
inferred from them.

You mention the Approval-Weighted Pairwise method. That method is unacceptable to me because it can elect a candidate that
is doubly-defeated  (i.e. both pairwise and approval-wise) by another candidate.

Margins provide good results with sincere votes, so why not use margins...
I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
irrespective of whether the votes are "sincere" or not.

Chris Benham

On 8/07/2019 9:09 pm, Juho Laatu wrote:
Few more words on using margins and Minmax in MOP-F2. These words are intended for readers who normally do not like margins, and who prefer methods that always elect the winner from the Smith set.

The previous mail already demonstrated that margins can be used, and still provide protection against some strategic attempts. Margins provide good results with sincere votes, so why not use margins, and leave the strategic concerns to the cutoff part of the method.

The question of electing from the Smith set is another interesting question. I'm not a proponent of Smith as a strict requirement since I think it makes sometimes sense to elect outside of Smith set (sincere votes assumed). There are however also cases where the top looped candidates are indeed clones in the sense that all voters consider them to be closely related. In such case their defeats to each others could be considered meaningless, and the winner should be one of them, even if there are candidates outside the Smith set with smaller defeats than the looped candidates have between themselves.

MOP-F2 can separate these different cases from each other by using the cutoff. This makes Smith set criterion compatibility a matter of the voters to decide. If voters vote 17: A>B>C>>D, 17: B>C>A>>D, 17: C>A>B>>D, 16: D>A>B>C, 16: D>B>C>A, 16: D>C>A>B, one of A, B and C will win. If A, B and C supporters will not use cutoffs, D wins in MOP-F2.

The point here is that use of Minmax as part of MOP-F2 makes it possible to voters to express their opinions more extensively (rankings + favourites) than in methods that use only rankings, and that this approach allows voters to indicate which candidates should be treated as clones and which ones not. Explicitly indicated clones will be treated as one fixed group of candidates (clones), without paying too much attention to their defeats to each others, in the final results.

One typical case where looped candidates would not be treated as if they were one candidate (group of clones) is 17: A>B>D>C, 17: B>C>D>A, 17: C>A>D>B, 16: A>D>B>C, 16: B>D>C>A, 16: C>D>A>B. In these votes A, B and C are clearly not clones since they are not next to each others in any of the ballots. D wins because its worst defeat is the smallest. With these votes, voters can not identify A, B and C as one uniform group on favourites (since they are not next to each others). Some voters might however use the cutoffs in some way, and change the result in some direction.

Margins and Minmax can thus be used, and the additional cutoffs may provide additional properties that plain margins and plain Minmax do not provide.



On 08 Jul 2019, at 13:36, Juho Laatu [hidden email] wrote:

The MOP approach is quite good in thwarting possible strategic burying attempts. I'll address briefly the basic case of having three candidates that will be looped strategically.

I'll use one particular MOP method that we might call MOP-F2 (identify Favourites using a cutoff, stretch factor = 2). It will be symmetric (both A>B and B>A will be weakened if A and B are marked as favourites), and it will be simply based on use of margins and Minmax. The modification function will be f(m) = m + sign(m) * (1-P/N) * 2 * N, where N is the number of voters, and P is the number of voters that identified the two candidates as their (Protected) favourites (= candidates above the cutoff). Parameter m is the original margin, and the output of the function will be the modified value to be fed to Minmax. Note that with 100 voters the original margin values are in range [-100, 100], and the modified values are in range [-300, 300].

I chose stretch factor to be 2, since that gives us some quite natural defensive properties. it seems that when certain number of of voters try to bury the sincere CW to make their favourite the winner (or vote sincerely that way), it is enough if a "corresponding" number of the other voters indicate the other two candidates as their favourites, to thwart the strategic attempt. In the example below 100% of the A supporters are strategic, and 100% of those voters that can easily defend against the strategy will do so. Stretch factor 1 would not give us such "full protection", and stretch factor 3 might be considered already an overkill.

49: A>B>>C --> A>C>B		large number of strategists that try to bury B
15: B>A>>C --> B>A>C		sincere CW, and a member of the AB group
36: C>B>A --> C>B>>A		the other group that decides to protect the CW

The opinions on the left are sincere. The actual votes are on the right. Note that all defensive votes (B and C supporters) can maintain the original ranking order. There is thus no need to falsify their rankings. The only required change is that they change the position of their favourite cutoff. B supporters stop protecting A, once they hear about the strategic plans of the A supporters. C supporters extend their protection to cover B (the CW), once they hear about the strategic plans of the A supporters.

The worst defeats with the original (non-modified) margins are A:-2, B:2 (CW), and C:-28. If A voters apply their strategy, the worst defeat margins become A:-2, B:-70, and C:-28. If the votes are modified, and B supporters still protect A, and C supporters do not protect B, the modified worst defeats become A:-172, B:-270, and C:-228. If B supporters stop protecting A, the modified worst defeats become A:-202, B:-270, and C:-228. And if C supporters protect B, the modified worst defeats become A:-202, B:-198, and C:-228. CW (B) wins again. It seems that one can effectively defend against any such (single) strategic burying attempt. And if you can, maybe nobody ever even tries such strategic attempts. This would make use of the cutoffs even more sincere, and more unnecessary, but maybe still good sincere information for the election analysts.

I like the idea that rankings of the non-starategic voters may stay the same, and also that the favourite cutoffs may stay quite natural, even when used defensively. The non-strategic voters may use the protective cutoff in a quite natural way, while it is of no use to the strategists. (If the A supporters would vote A>C>>B, they would just support the election of C that is the worst alternative to them.) There is also no need to resort to use of winning votes or something similar to provide means to defend against strategies (margins are a more natural way to measure sincere preferences).




P.S. I note that in an earlier mail I forgot to include the sign function in one of the the equations. That equation should have been sign(initial)*(1-mod)*k*max + initial. I'm sorry if someone tried to calculate the results and got confused, and didn't dare to ask why that function seemed strange. I hope at least the function I gave above is correct. I think that one (MOP-F2) is a good one to start with, if someone wants to test the MOP methods in practice.

On 08 Jul 2019, at 10:06, Juho Laatu [hidden email] wrote:

James Green-Armytage's Cardinal-Weighted Pairwise method (CWP) has also a variant, Approval-Weighted Pairwise (AWP), that uses only approvals (instead of full ratings) to determine the strength of the pairwise preferences (https://electowiki.org/wiki/Cardinal_pairwise). AWP comes actually quite close to the Modified Overall Preferences approach (MOP) when one uses only one cutoff to tell which candidates are to be protected.

Those two methods are similar in that they both use ballots that have one approval like cutoff (in addition to the basic rankings). They both determine the (potentially cyclic) ranking order of the candidates based on the rankings of the ballots only (not using the cutoffs). They both use the cutoff only in determining the strength (not direction) of the pairwise preferences. They both then proceed to use some Condorcet method (pick your favourite) to find the winner, based on the newly calculated defeat strengths.

Now to the differences. I'll use abbreviation MOP-F to refer to a MOP method that is uses only one cutoff to indicate a set of "preferred favourites" whose defeats to each other should be treated as "friendly" defeats.

- AWP uses the cutoff to determine the defeat strengths, while MOP (and MOP-F too) uses the cutoff to modify the defeat strengths that were derived from the rankings. The basic philosophy in MOP is to allow voters to tell which defeats should be treated as "friendly" / "weak" defeats. This means that although most voters may agree on the direction of some pairwise preference, that defeat may still be considered weak, if voters want it that way.

- In AWP every vote that approves A and does not approve B, increases the strength of that defeat (A>B). In MOP-F every vote that has A and B above the cutoff, indicates that defeats between A and B should be considered "friendly" (or only A>B if the method is asymmetric). These approaches are quite similar in the sense that having the two candidates above the cutoff leads to weak defeats, and having the candidates at different sides of the cutoff leads to strong defeats. But defeats between two candidates below the cutoff are weak in AWP but strong in MOP-F.

In most elections all well known Condorcet methods may well be good enough as they are, without any additions. If additions are however used, one additional cutoff may not be too complex to use. (Also CWP's ratings may be easy enough in some elections.) In AWP one should place the cutoff so that one's strongest or most important preferences between candidates cross the cutoff point. In MOP-F one should put those candidates that one sees as "all good choices" above the cutoff. The MOP-F approach is good in the sense that the placement of the cutoff is very sincere. In AWP one may have to consider the strategic placement of the cutoff more. In order to make the vote as efficient as possible in MOP-F, one should however consider positioning the cutoff so that at least two potential winners (= candidates that may end up in the top loop) will be above the cutoff. Otherwise one might defend only candidates that have no chances to win. If one doesn't have two such candidates th
 at
  one wan
 t
s
to defend (sincerely, or potentially in some cases to defend against strategic voting), then setting the cutoff higher (and probably having no influence on the outcome), or not using the cutoff at all, makes sense. This adds a little bit of strategic thinking also to MOP-F. A good simple advice to the MOP-F voters could be to "mark those candidates that you want to defend".

In most typical (large public) elections cutoffs are quite irrelevant in both methods, since they will influence the outcome only when there is a top loop. Voters may thus ignore them if they so wish. If voters do not use the cutoff feature, MOP-F uses the original strengths that were derived from the rankings. In AWP placing the cutoff somewhere in the vote is somewhat more important since otherwise all defeats would have the same strength. Getting additional information on the voter preferences may be useful and interesting for statistical purposes, even if they don't have any influence on the final outcome in most elections.

I note that CWP may be easier to the voters to adopt than AWP, since rating the candidates may be more natural than trying to identify the correct approval cutoff point. On the MOP side MOP-F may be the easiest option since naming one's favourites, or those candidates that one wants to defend, seems much easier than trying to determine preference strengths separately to all given preference relations. I mean that although I make a technical comparison between AWP and MOP-F (to point out the differences in their philosophy), in real life the choice might be between using CWP and MOP-F. When comparing CWP to AWP, the main difference is that in CWP the voter will share the strength of the used range (e.g. from 0 to 100) between the given preferences (in fractions of the range). In MOP methods the philosophy is rather to give maximal protection to some pairs (= weaken their defeats to each other) and none to others. (Also intermediate values would be possible but probably no
 t
 practica
 l
,
since they would be just "weakened opinions".)


On 26 Jun 2019, at 10:55, Juho Laatu [hidden email] wrote:

I studied the Modified Overall Preferences approach a bit more. Here's one new simple pairwise preference modification function. I'll explain the MOP approach again below, and describe the new simple modification function too.

The modification process starts from (1) some initial set of preferences. I'll use margins again, but also e.g. winning votes would be possible, although maybe not as natural as margins since the modification process already does quite similar things to the results that also the winning votes approach is supposed to do.

The next step (2) is to modify the strengths of the initial pairwise preferences. The simple modification function is (1-mod)*k*max + initial, where "initial" is the initial pairwise comparison result, "max" is the largest possible defeat size (= number of voters in margins), "k" refers to the easiness of influencing the preferences (e.g. k=4 means that if 1/4 of the voters (25%) think that A and B are "clones", then defeats between them are always weaker than any defeat between two candidates that nobody claims to be "clones"), "mod" refers to the proportion of voters that think that the compared candidates should be considered "clones" (protected from strong defeats to each others), and that the pairwise defeats should be modified / weakened (0 = nobody says so, 1 = all voters say so).

The third and final phase (3) is to find the winner based on the modified pairwise preferences. In this mail I'll use Minmax. Other approaches work too.

I'll describe the philosophy of the modification function briefly. Vote A>B>>C will be read as a normal ranked preference, but in addition to that the voter wants to say that A and B should be seen as "clones" (and defeats between them should be weakened). The presence of ">>" is interpreted so that the ">" preferences will be read as indicating "clones" / protection. In this modification function these opinions are symmetric (i.e. also (possible) B>A defeat should be weakened). If k=4 and we have 100 voters, that means that the modified preference strength values of defeats (that were initially margins in range 1..100) may range from 1 to 400. If mod=0, the strength range of defeats is from 301 to 400. If mod=1/4=1/k, the range is from 201 to 300. If mod=1, the range is from 1 to 100. This modification function can be said to use a "spread spectrum" technique since it expands the range of pairwise preference strengths (k times wider than the range of the initial streng
 th
 s) to cr
 e
a
te
space also for the weakened preferences.

Note that only the strengths of the defeats are modified. The (cyclic and transitive) preferences will stay just as they are in the initial preferences.

Rankings and protection ("clone indications") are in principle two separate opinions that could be given separately, but in practice protection is probably derived from enhanced rankings that may contain e.g. cutoffs, numeric gaps, positional gaps or strong and weak preference relations.

Next, few examples (with numeric results, in case you want to check the operation of the method). The first one describes a burying strategy example.

45: A > B >> C
15: B > A >> C
40: C > B > A

A and B clearly form a uniform group that has 60% support. B would win because it is a Condorcet winner, but then all the 45 A supporters decide to bury B and vote A>C>B. That would make A the winner. If k=4 and 31 C supporters (out of the 40) decide to protect B by voting C>B>>A, then B wins despite of the deceptive A supporters (A:-350, B:-346, C:-420). The method can thus defend against this kind of burying strategy in a quite natural way. Also the 15 B supporters could stop supporting A (as a "clone" of B) and vote B>A>C instead, in which case already 16 C supporters (voting C>B>>A) would make A supporters' strategy void (A:-410, B:-406, C:-420).

40: A		(no "clones")
35: B>>C>A	(A and C indicated as "clones")
25: C		(no "clones")

This is the first challenge of Forest Simmons. A wins if k=4, as it should (A:-280, B:-405, C:-410). A wins also with k=1, k=2 and k=3.

40: A
35: B>C>>A
25: C

This is the second challenge. C wins if k=4, as it should (A:-420, B:-405, C:-270). C wins also with k=1, k=2 and k=3.

40: A
35: B>>C>A
25: C>>B>A

This is the third challenge. B wins if k=4, as it should (A:-320, B:320, C:-410). C wins also with k=0 (no modification), k=1, k=2 and k=3. Any number of C supporters could change their vote to C>B>A, and/or B supporters to B>C>A, and B will still win. That is because B is a Condorcet winner, and the modification function will keep initial defeats (and ties) as defeats (and ties), just with different strengths.

If you have examples where this kind of pairwise overall defeat modifying / weakening approach would not work well, please tell. I'm eager to evaluate MOP methods against such challenges.



On 18 Jun 2019, at 09:01, Juho Laatu [hidden email] wrote:

Chris Benham and I discussed (in private mail) on how well the example method below complies with the challenge that Forest Simmons proposed.

Forest Simmons fsimmons at pcc.edu
Thu May 30

In the example profiles below 100 = P+Q+R, and?? 50>P>Q>R>0.??

I am interested in simple methods that always ...

(1) elect candidate A given the following profile:
P: A
Q: B>>C
R: C,

and
(2) elect candidate C given
P: A
Q: B>C>>
R: C,

and
(3) elect candidate B given
P: A
Q: B>>C?? (or B>C)
R: C>>B. (or C>B)
I copy my answers to those questions below.

Note that I'm struggling a bit with presentation of votes since Forest Simmons' challenge seems to talk about strengthening some preferences, while my mail talked about weakening some preferences (= making the defeats friendlier). When interpreting Forest's presentation of votes I assumed weakened preferences to be present only in places where the stronger (">>") preferences make the presence of weaker preferences obvious. I thus assumed that e.g. vote B>C (= B>C>A) in Case (3) to not contain any weakened preferences. Same e.g. with vote A in all of the cases. Only votes like B>>C (= B>>C>A) were assumed to contain weakened preferences (between C and A).

I tried to see how it relates to Forest's requests.

Case (2)
Since all candidates lose to one of the others, and the only defeat to be weakened is the B>C defeat, and there are always more than 25% of the voters demanding that (Q>25%), then the "4*" version always lowers the strength of that defeat to 0. Therefore C always wins with that method.

Case (1)
The example method didn't have any special "dislike" cutoff (only a "near clones this far" cutoff). I.e. only votes like A>B>C>>D>>E were allowed. The easiest way to introduce richer use of preference strengths would be to allow any preference (in the ranked vote) to be either ">" or ">>". In this case that would lead to votes Q: B>>C>A. That would weaken A's defeat to C. Again, the number of Q voters is higher than 25%, and with the "4*" factor the strength of A's defeat to C would become 0, and A would win.

Case (3)
Also here I must assume that votes Q: B>>C>A and R: C>>B>A are allowed. Since (Q+R)>25%, strength of A's defeat would be 0. Candidate B has however no defeats nor ties, so B will win in any case. Having votes of form B>C>A or C>B>A would make A's defeat more severe (at least once we get below the 25% "clone treatment recommendation" limit). Candidate B (Condorcet winner) wins thus also in the original example method.

In summary, the method seems to meet Forest's requirements if we allow also weakening of the lower preferences.
I note that the method doesn't even have a name yet. Maybe MMM-MOP will do. It refers to the base method (Minmax(margins)) and the modification to it (Modified Overall Preferences). That leaves some space to using different parameters and approaches in modifying the pairwise preferences in the matrix.

P.S. A bit more on weakened preferences / indicated clones (vs strengthened preferences). I'll use square brackets to indicate clones/weakening here. One can vote [A>B]>C (corresponds to A>B>>C), which means that B's possible defeat to A should be just a friendly one. One could vote also [A=B]>C if ties are allowed. This means that also ties can be weak. Or maybe one should rather say that also tied candidates can be considered clones/friendly (maybe even as default). One option would be to allow also truncated candidates (equal last) be indicated as being friendly. As noted in the beginning of my original mail, the friendliness indicators / weak preferences can be seen as a separate opinion that doesn't have much to do with the original ranking of the candidates (those rankings are as strong as ever in determining the (possibly looped) order, even though the final strength of preferences will be weakened, once the (possibly looped) order has been determined). In typic
 al
  Condorc
 e
t
methods weakening of overall pairwise preferences has no meaning except when there are loops, and strengths of different pairwise defeats need to be compared. In the example method of the original mail I stuck with one approval like cutoff only, but in order to respond to the challenge of Forest Simmons we need a bit richer preference structure. If we would go for three preference strengths (>, >>, >>>), maybe that would lead to a "clones among clones" approach in the weakened votes terminology.
P.P.S. In the original mail, ignore words "which means".

Juho


On 16 Jun 2019, at 16:51, Juho Laatu [hidden email] wrote:

The point of this mail is to promote the idea of separating the strength of different pairwise preferences from the ordinary rankings (of ranked or Condorcet methods). Instead of giving voters the ability to adjust the strength of their own vote or pairwise preferences (the more common approach), they can influence the strength of the final preferences of the whole electorate. In this approach the pairwise preferences of the electorate will not be changed from A>B to B>A. Only the strength of that preference (A>B or B>A) will be modified.

You could make many complex tricks with this kind of tools, but I'll concentrate here on one example method that is very simple to the voters. It will consist of adding just one approval like threshold to a basic Condorcet method. I'll use Minmax(margins) as the base method to be modified. Instead of voting A >> B >> C >> D >> E (a normal ranked vote) voters can vote also A > B > C >> D >> E (i.e. one or more of the first preferences may be weakened/moderated). The philosophy of the latter vote is that A, B and C are promoted as "favoured candidates" or "near clones" or "protected candidates" (with the full strength of one vote).

In the counting process, first count the normal pairwise preferences (i.e. the matrix). After that, some of the pairwise preferences are made weaker. And then the final results are counted, (almost) as in Minmax(margins).

Our first approach is to count the number of votes that had preference A>B (moderated preference) (= Mab). Then the A over B preference count (of the whole electorate) will be weakened by multiplying it with 1-Mab/N (where N = number of votes). Moderated preferences are transitive in the sense that A > B > C increases also the Mac count. The idea is that if voters consider A and B to be "near clones", their defeats to each other should be seen and treated as "friendly defeats". Their strength can thus be weakened, although a large number of voters may have preferred one over another (in a friendly way).

(I note that some alternatives to the presented example method could be to make it symmetrical by using factor 1-(Mab+Mba)/N, or one could use the number of votes that preferred either A to B or B to A instead of N.)

I'll add one more trick to the example method. The "friendliness factor" can be stronger than presented above. Let's say that 50% of the voters (maybe one of two parties) think that A and B should be treated as "near clones". Half of those voters have voted A > B, and half B > A. This means that 25% of the voters have given (moderated) preference A > B. Maybe already this 25% is enough to convince us that A and B indeed are to be treated as full clones. If so, we can use factor max(1-4*Mab/N,0) instead of 1-Mab/N. Here reaching strength 0 means that those candidates will be declared as "clones" and treated as such.

The counting process may have some problems with ties if multiple preferences will be equal to 0. I will not discuss this question much more in this mail. Let's just say that also those preferences (although they all seem to be equal to 0) can be seen to have different strength, e.g. based on the initial margins. And that if there are two parties with 50 votes, and there are some "clones" in one of the parties, they could be seen as one alternative when flipping the coin. (Winner among the clones to be decided separately if the clone party wins.)

Here's one example set of votes.

45 A>B>>C --> A>>C>>B
15 B>A>>C
40 C>B>>A

B is the sincere Condorcet winner. Supporters of A will however strategically bury B under C. In Minmax(margins) the worst defeats are A:-10, B:-70, C:-20. The strategic voters will get what they wanted. There were however 40 voters that said Mcb. This means that in the example method the strength of B's defeat to C will be 0 (using the "4*" in the factor). There are also 15 voters that said Mba. This helps A a bit (-10 --> -4), but not too much. If all C supporters (or 17 or more of them) had voted C>>B>>A, A would win. C supporters may thus vote C>B>>A sincerely or as a (still quite sincere) defensive tactic (after hearing about A supporters' plans).

Another question where this ability to moderate the defeats of favourite near clones is interesting is whether to elect from the Smith set or outside of it.

17 A>>B>>C>>D
17 B>>C>>A>>D
17 C>>A>>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are not "clones" in the described sense (which means no weakening of their mutual defeats since not a single voter proposed that). D wins.

17 A>B>C>>D
17 B>C>A>>D
17 C>A>B>>D
16 D>>A>>B>>C
16 D>>B>>C>>A
16 D>>C>>A>>B

A, B and C are "clones" (weight of mutual defeats = 0, with the "4*" moderation). One of them wins.


----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info
---
This email has been checked for viruses by AVG.
https://www.avg.com

----
Election-Methods mailing list - see https://electorama.com/em for list info
----
Election-Methods mailing list - see https://electorama.com/em for list info

Virus-free. www.avg.com
<a href="x-msg://22/#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2" width="1" height="1" class="" moz-do-not-send="true">
----
Election-Methods mailing list - see https://electorama.com/em for list info


----
Election-Methods mailing list - see https://electorama.com/em for list info

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] MOP-F2 / Later-no-Help

Juho Laatu-4
In reply to this post by C.Benham
> On 08 Jul 2019, at 17:36, C.Benham <[hidden email]> wrote:

>> Margins provide good results with sincere votes, so why not use margins...
> I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
> irrespective of whether the votes are "sincere" or not.

Later-no-Help is not one of my favourites either. I would have some sympathy towards "Later-yes-Help". That is because Later-no-Help seems to tell the voters that it is ok to truncate and not give their sincere rankings, while "Later-yes-Help" says that voters would be better off if they would tell the method all their preferences.

On the other hand, it would be good if voters are not punished too much if they truncate in an election with hundreds of candidates. Truncation of candidates that have no chances to win should be harmless. This means that a "Later-Irrelevant-Alternatives-no-Help" could be a better criterion than Later-no-Help. I would at least strongly encourage voters to rank all (hopefully not too many) potential winners (except the last one, whose position can be made clear already by ranking all the others).


----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

[EM] MOP-F2 definition

Juho Laatu-4
In order to keep definitions clear, I wrote a compact definition of the MOP-F2 method.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

MOP-F2

1) Votes are ranked votes with one cutoff. Candidates above the cutoff are called protected candidates. The default position of the cutoff is at the beginning of the vote (i.e. no protected candidates).

2) Count the margins (for each pair of candidates). This is based on the rankings only.

3) Modify the margins using function f, where
- f(m,p) = m + sign(m) * (1-p/N) * 2 * N
- m is the original pairwise margin
- p is the number of votes that identified both candidates as protected candidates
- N is the number of votes
- sign(x) is +1 if x>0, -1 if x<0, and 0 if x=0

4) Use the Minmax algorithm to find the winner (i.e. the candidate whose worst defeat (based on the modified margins) is smallest).



* Note that function f will modify only the strength of the defeats. The direction of the defeats will not change.
* Note that function f can be simplified to (a shorter but maybe less intuitive) form f(m,p) = m + sign(m) * 2 * (N-p)
* Note that function f stretches the range of defeat strength values from [-N, N] to [-3*N, 3*N]
* In the name of the method MOP refers to "Modified Overall Preferences", F refers to word Favourites, and 2 refers to the "stretch factor" 2



----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] MOP-F2 / Weak Defensive Strategy criterion, Strong Defensive Strategy criterion

Juho Laatu-4
In reply to this post by C.Benham
On 08 Jul 2019, at 22:05, C.Benham <[hidden email]> wrote:


Non-Drastic Defense refers to performance with strategic votes.
http://mam-docs.000webhostapp.com/
non-drastic defense:  If more than half of the voters prefer alternative y 
        over alternative x, then that majority must have some way of voting 
        that ensures x will not be elected and does not require any of them to
        rank y over any more-preferred alternatives. (This has been promoted
        by Mike Ossipoff under the name Weak Defensive Strategy Criterion.  
        Non-satisfaction means some members of the majority may need to 
        misrepresent their preferences by voting a compromise alternative 
        over favored alternatives if they want to ensure the defeat of less-
        preferred alternatives.) 

The electowiki definitions of Weak Defensive Strategy criterion and Strong Defensive Strategy criterion seem to play funny word games. Strange things are hidden in the supporting definitions, e.g. on what voting two candidates equal is supposed to mean. Truncation pops up again as a point with a special meaning. Some detective work needed to find out what is allowed and what not.

Let's study an alternative, clearer and a bit stronger defensive strategy criterion.

"If a majority prefers one particular candidate to another, then they should have a way of voting that will ensure that the other cannot win, without any member of that majority changing their rankings."

In MOP-F2 this means that voters in that majority should leave their rankings intact and only (possibly) change the position of the cutoff. In a three candidate election where C is the candidate that the majority doesn't like, those majority members that prefer both A and B to C should put their cutoff after A and B (this means A>B>>C or B>A>>C with fully ranked votes). Other majority members should not use the cutoff. I couldn't identify any scenario in MOP-F2 where C could still win, when majority members follow this strategy.

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

Juho Laatu-4
In reply to this post by Juho Laatu-4
On 09 Jul 2019, at 00:55, Juho Laatu <[hidden email]> wrote:

On 08 Jul 2019, at 22:05, C.Benham <[hidden email]> wrote:

46: A>C
10: B>A
10: B>C
34: C=B

B>A  54-46 (margin=8),   A>C 56-44 (margin=12),  C>B 46-20 (margin=26).

Here on more than half the ballots B is voted both above A and no lower than equal-top, and yet Margins elects A.

No other candidate is voted at least equal-top on more than half the ballots. Fans of Bucklin or any version of Median
Ratings would be particularly incensed. 

The version of Losing Votes that I advocate elects B because it has the 34C=B ballots give a whole vote to each of C and B
in their pairwise comparison, so "C>B 46-20" becomes C>B 80-54, giving B a winning MM Losing Votes score of 54. 

I don't like losing votes too much since they may give strange results already with sincere votes. I don't have any dramatic examples available, but I believe there are such examples. I gave one example on winning votes earlier (50: A>B>C, 50: D>E>F, 1: F>A). Sorry about basing my claim on losing votes on "feelings only".

P.S.

Concerning concrete examples with Losing Votes, similar examples seem to serve as bad examples also with Losing Votes and Pairwise Opposition (in addition to Winning Votes). I mean vote sets like 50: A>B, 50:C>D, 1:D>*, where * refers to either A or B.

Winning Votes, Losing Votes and Pairwise Opposition are quite artificial constructions when compared to margins and other such pairwise comparison functions that aim more at estimating the (society related) seriousness of each defeat. I discussed such more natural pairwise comparison functions (e.g. Relative Margins and Moderated Margins) a bit more in http://lists.electorama.com/pipermail/election-methods-electorama.com/2019-May/002126.html. If one wants to change the behaviour of Margins, one option is not to jump directly to Winning Votes etc but to make some less radical changes (that avoid the worst pathologies). Also the MOP track is interesting in the sense that the modification function can be used to tweak the original (Margins or other) strengths in a similar manner (e.g. based on indicated favourites as in MOP-F2).




----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] MOP-F2 / ties

Juho Laatu-4
In reply to this post by Juho Laatu-4
> On 08 Jul 2019, at 21:01, Juho Laatu <[hidden email]> wrote:
>
>> On 08 Jul 2019, at 17:36, C.Benham <[hidden email]> wrote:
>
>>> Margins provide good results with sincere votes, so why not use margins...
>> I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
>> irrespective of whether the votes are "sincere" or not.
>
> I googled the Non-Drastic Defense criterion and this example.
>
> 46 A>C
> 10 B>A
> 10 B>C
> 34 C=B
>
> Non-Drastic Defense criterion:    if on more than half the ballots X is voted both above Y and below no other candidate (i.e. no lower than equal-top) then Y must not win
>
> The Non-Drastic Defense criterion says that A should not be elected.
>
> If no cutoffs are used, MOP-F2 is the same as the base method, i.e. Minmax(margins), and elects A. But if those 34 voters that seem to consider C and B to be clones indicate this in their votes (as they probably should) by voting C=B>>, C>B>>, or B>C>>, then B wins.
>
> I guess it would be ok in MOP-F2 not to support use of "=" in the vote (explicitly) since in this case since they can vote also C>B>>, or B>C>>, and by doing this, indicate that these two candidates should be seen as clones / their mutual defeats should be seen as weak defeats.

Few more words on different options in handling ties in MOP-F2.

Every vote that ranks A above B makes B's defeat to A one point stronger. But if both A and B are protected (above the cutoff), both B's defeat to A and A's defeat to B will be made two points weaker. As a result ranking one's favourites does not harm them (in the sense that they would be worse off in comparisons if there is a top loop).

Someone might try to maximise the protection of one's favourites by not giving any preference between them, but vote those candidates equal in order to avoid any defeats between one's favourites. This could be achieved by casting a A=B>>... style vote. In principle I don't like any incentives to hide preferences in ranked elections, so let's see what possible alternatives there are.

First alternative could be to not allow use of "=" in the votes (except implicitly between the truncated candidates). That would make sense if the used ballots are such that supporting use of "=" would just introduce additional technical problems and unnecessary complexity in the voting process. Not being able to use "=" is no problem since voters can easily replace it with ">". Usually there is a small preference difference anyway. And if there really is none, then flipping a coin is a good solution. In some other ballot types marking candidates as tied may however be a very natural (e.g. if there is a fixed number of positions/ratings/columns/slots to use), and even unavoidable if e.g. the number of slots is smaller than the number of favourite candidates. This approach is thus quite good for some ballot types. It is not a problem not to support explicit ties ("=" between ranked candidates) if it is easier to implement the voting process without them.

Another approach would be to handle A=B>> style voting so that both candidates will lose to each others (not very pretty). Or alternatively one could give only one point of protection to candidates that are tied or at the winning end of a ">" relation. The calculation process would be a bit more complex than in the original MOP-F2. It may be easier to voters to understand the simple protection rules of the original MOP-F2 than this kind of nuances.
----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] MOP-F2 / compromising

Juho Laatu-4
In reply to this post by Juho Laatu-4
This mail will be short. :-)

Condorcet methods are vulnerable to compromising strategy when there is a top loop. It is not easy to defend against that strategy. This threat may however look more serious on paper than it is in real life, since in real life voters are typically not aware that the group opinion will be circular.

MOP-F2 offers voters an easy and natural way to indicate willingness to compromise, without modifying their rankings. Instead of voting A>B>C, the voter can vote A>B>>C, and thereby indicate that A and B are his favourite candidates, and electing B would not be a bad thing. Without going to the details, I just note that this way of handling the compromising incentive seems to provide quite natural means to pick the winner among the looped candidates in a sincere manner, and without falsified rankings in the ballots.




> MOP-F2
>
> 1) Votes are ranked votes with one cutoff. Candidates above the cutoff are called protected candidates. The default position of the cutoff is at the beginning of the vote (i.e. no protected candidates).
>
> 2) Count the margins (for each pair of candidates). This is based on the rankings only.
>
> 3) Modify the margins using function f, where
> - f(m,p) = m + sign(m) * (1-p/N) * 2 * N
> - m is the original pairwise margin
> - p is the number of votes that identified both candidates as protected candidates
> - N is the number of votes
> - sign(x) is +1 if x>0, -1 if x<0, and 0 if x=0
>
> 4) Use the Minmax algorithm to find the winner (i.e. the candidate whose worst defeat (based on the modified margins) is smallest).
>
>
>
> * Note that function f will modify only the strength of the defeats. The direction of the defeats will not change.
> * Note that function f can be simplified to (a shorter but maybe less intuitive) form f(m,p) = m + sign(m) * 2 * (N-p)
> * Note that function f stretches the range of defeat strength values from [-N, N] to [-3*N, 3*N]
> * In the name of the method MOP refers to "Modified Overall Preferences", F refers to word Favourites, and 2 refers to the "stretch factor" 2



----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] Modified Overall Preferences / margins, Smith set

C.Benham
In reply to this post by Juho Laatu-4


On 10/07/2019 8:48 pm, Juho Laatu wrote:

Concerning concrete examples with Losing Votes, similar examples seem to serve as bad examples also with Losing Votes and Pairwise Opposition (in addition to Winning Votes). I mean vote sets like 50: A>B, 50:C>D, 1:D>*, where * refers to either A or B.
Juho, say * is A. Whereas my examples are simple and plausible yours are not.  Your example would be a bit more "realistic" if none of the pairwise
scores were identical, none of the first-preference scores were identical, none of the pairwise contests were exact ties and every candidate got some support
in every pairwise contest.

In your example three of the pairwise contests are 51-50 (A>C, D>A, D>B), A and C both have 50 first-preference votes, B and C tie 50-50 and B has no pairwise
support against A. Nonetheless  in your example I don't see how all three of Margins, Winning Votes and Margins not elect C (say with Smith//MinMax).

Chris Benham

On 09 Jul 2019, at 00:55, Juho Laatu <[hidden email]> wrote:

On 08 Jul 2019, at 22:05, C.Benham <[hidden email]> wrote:

46: A>C
10: B>A
10: B>C
34: C=B

B>A  54-46 (margin=8),   A>C 56-44 (margin=12),  C>B 46-20 (margin=26).

Here on more than half the ballots B is voted both above A and no lower than equal-top, and yet Margins elects A.

No other candidate is voted at least equal-top on more than half the ballots. Fans of Bucklin or any version of Median
Ratings would be particularly incensed. 

The version of Losing Votes that I advocate elects B because it has the 34C=B ballots give a whole vote to each of C and B
in their pairwise comparison, so "C>B 46-20" becomes C>B 80-54, giving B a winning MM Losing Votes score of 54. 

I don't like losing votes too much since they may give strange results already with sincere votes. I don't have any dramatic examples available, but I believe there are such examples. I gave one example on winning votes earlier (50: A>B>C, 50: D>E>F, 1: F>A). Sorry about basing my claim on losing votes on "feelings only".

P.S.

Concerning concrete examples with Losing Votes, similar examples seem to serve as bad examples also with Losing Votes and Pairwise Opposition (in addition to Winning Votes). I mean vote sets like 50: A>B, 50:C>D, 1:D>*, where * refers to either A or B.

Winning Votes, Losing Votes and Pairwise Opposition are quite artificial constructions when compared to margins and other such pairwise comparison functions that aim more at estimating the (society related) seriousness of each defeat. I discussed such more natural pairwise comparison functions (e.g. Relative Margins and Moderated Margins) a bit more in http://lists.electorama.com/pipermail/election-methods-electorama.com/2019-May/002126.html. If one wants to change the behaviour of Margins, one option is not to jump directly to Winning Votes etc but to make some less radical changes (that avoid the worst pathologies). Also the MOP track is interesting in the sense that the modification function can be used to tweak the original (Margins or other) strengths in a similar manner (e.g. based on indicated favourites as in MOP-F2).




----
Election-Methods mailing list - see https://electorama.com/em for list info

Virus-free. www.avg.com

----
Election-Methods mailing list - see https://electorama.com/em for list info
Reply | Threaded
Open this post in threaded view
|

Re: [EM] MOP-F2 / Later-no-Help

C.Benham
In reply to this post by Juho Laatu-4
On 9/07/2019 3:16 pm, Juho Laatu wrote:

> Later-no-Help is not one of my favourites either. I would have some sympathy towards "Later-yes-Help". That is because Later-no-Help seems to tell the voters that it is ok to truncate and not give their sincere rankings, while "Later-yes-Help" says that voters would be better off if they would tell the method all their preferences.
Since the method is only interested in finding a single-winner, how is
it fair that some voters would be "better off" (than other voters)
because they
"tell the method all their preferences"?

Getting back to your quibbles about the Plurality criterion (which you
say you "don't like"), when I first started thinking about single-winner
voting
methods, I (living in IRV land) thought it was "obviously" highly
desirable that a ballot that has an irrelevant weak candidate ranked
above some
other candidates should have exactly the same effect (and so probably
"significance" in the view of criteria) as a ballot that ignores (or
just lower
ranks) that weak candidate but ranks all the other candidates in the
same order (so has a different candidate ranked below no others).

However on becoming better acquainted with all the problems and issues,
I now think property is far less important.

We get a higher "social utility", more legitimate-looking winner if the
voters are discouraged from expressing their very weak, perhaps light-minded
ill-informed preferences that the method is unable to distinguish from
their serious strong preferences.

Even if the method allows the voter to do nothing except rank the
candidates, we nonetheless logically know which are the voters strongest
most
serious pairwise preferences: their preference for the candidate/s they
vote below no others over the candidate/s they vote above no others.
If I had coined something like the Plurality criterion to take account
of ballots that allow equal-top ranking and not have any possible confusion
or ambiguity about what a "vote" is, it would say something like "If A
is voted above B and below no other candidate on more ballots than B is
voted above any candidate, then B can't win."

But, based on your vague negativity regarding the Plurality criterion
and some of your other statements, apparently you have it as a
"principle" that
this information should be ignored (at least in methods that don't allow
the voters to give any explicit approval cutoff).

We infer from your "some sympathy for later-yes-help" that you like
Later-no-Harm.  Both that and Later-no-Help are incompatible Condorcet,
which is why when attacking Margins (while promoting some other
Condorcet methods) I refer to "egregious" failure of Later-no-Help (going
along with failure of Plurality).

> This means that a "Later-Irrelevant-Alternatives-no-Help" could be a better criterion than Later-no-Help.

Juho, How (for this purpose) would you define an "Irrelevant Alternative"?

Chris Benham

On 9/07/2019 3:16 pm, Juho Laatu wrote:

>> On 08 Jul 2019, at 17:36, C.Benham <[hidden email]> wrote:
>>> Margins provide good results with sincere votes, so why not use margins...
>> I don't see how egregious failures of the Plurality  and Later-no-Help (and even Non-Drastic Defense) criteria constitute "good results"
>> irrespective of whether the votes are "sincere" or not.
> Later-no-Help is not one of my favourites either. I would have some sympathy towards "Later-yes-Help". That is because Later-no-Help seems to tell the voters that it is ok to truncate and not give their sincere rankings, while "Later-yes-Help" says that voters would be better off if they would tell the method all their preferences.
>
> On the other hand, it would be good if voters are not punished too much if they truncate in an election with hundreds of candidates. Truncation of candidates that have no chances to win should be harmless. This means that a "Later-Irrelevant-Alternatives-no-Help" could be a better criterion than Later-no-Help. I would at least strongly encourage voters to rank all (hopefully not too many) potential winners (except the last one, whose position can be made clear already by ranking all the others).
>
>
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info

---
This email has been checked for viruses by AVG.
https://www.avg.com

----
Election-Methods mailing list - see https://electorama.com/em for list info
12