[Election-Methods] renaming DYN

[Forest W Simmons]

Greg wrote ...


>Ah, I notice now that A and B can spend their proxies by approving
>multiple candidates other than themselves.  This somewhat increases
>the chance that they will cooperate by approving of each other to
>prevent C from winning.  It still seems likely to me that they will
>choose to not cooperate with each other because they are rivals.

After the voters mark the ballots, (but before the proxies perform 
their service) the results are C(49S, 51X), B(24S, 49X, 27?), and 
A(27S, 49X, 24?).

Furthermore, A gets to make the decision on the 27 question marks about 
B, and B gets to make the decision on the 24 question marks about A.  
Keep in mind that A and B like each other more than they like C.  The 
voters knew this from A and B's published rankings prior to the 
election.

In ordinary approval, if B supporters promised A the 24 S's, it might 
still be in their interest (disregarding the blot on their honor) to 
defect and give A 24 X's instead.

But in this version, where A and B sit down together in front of an 
election referee to do their proxy duty, defection is impossible.

Where defection is not possible, the prisoner's dilemma dissolves.  
Prisoner's dilemma cannot exist w/o the possibility of defection.

If A is nice, A might give B two extra S's to bring B up to the 
prestigious position of second place, along with all of the advantages 
this might bring to his future political career.

>The insight that if you trust someone to represent you then you should
>also trust them to act as a proxy for you seems to not always hold.

Your elected representative is your proxy in the legislature for his 
term of service, whether you like it or not.  But in this method the 
candidate of your choice (perhaps your write-in) is your proxy only on 
the choices that you choose not to make for yourself.

This partial proxy method gives the voter much more control than 
(ordinary) Asset voting does.

The method is simpler than any method based on rankings (IRV, CSSD, 
MAM, etc.), as simple as MCA, but (unlike MCA) it completely relieves 
the voters (except the proxies) from strategy considerations.

It beats Approval in that the strategizers (i.e. the proxies) have 
reliable info (as opposed to corporate poll disinfo) by the time they 
have to strategize.  

Also (with extremely rare exception) the voter gets to distinguish her 
favorite candidate with the letter P without any negative consequences. 
 This answers one of the biggest IRVist complaints about Approval.

In other words, this method comes the closest of any simple 
determisistic method to satisfying the strong FBC.

Asset voting comes in second (because of poorer control over proxy).

Most IRV advocates believe that IRV satisfies the strong FBC, so the 
weak FBC compliance of Approval does not impress them.

Ultimately, DYN/PSX Voting's main advantage over other methods is that 
it is a simple, summable, strong FBC compliant method that tens to 
maximize voting power, the probability of an individual voter's ballot 
making a difference in the voter's favor in a typical election.

To see this, consider that Approval tends to maximize voting power 
(with the proper strategy) given accurate polling statistics.  But the 
stats available for use by the approval strategizers in this method are 
committed partial results, therefore not subject to manipulation, 
therefore more accurate.

In my humble opinion, this is the method that we should all be pushing, 
as soon as we get a decent name for it.

Forest