[EM] Re: Approval Strategy A- Question for Rob LeGrand

[Dgamble997 at aol.com]

Kevin Venzke wrote:

>Are you looking to show that Plurality, for example, is more likely to be
>proportional
>than Condorcet? Random Ballot is easily more proportional than that. Better
>yet,
>put a PR method in your model.

The one thing the model has demonstrated clearly than anything else is the 
truth of what myself and a number of other people on the list have been saying 
for a long time - an assembly made up of single seats can be proportional only 
by chance. 

Which single seat method you use can have a considerable effect on the 
make-up of the assembly though. Plurality is neutral as regards where parties are 
positioned on a left-right spectrum, IRV is favourable towards centrists, 
Approval ( with all voters using strategy A once) and Condorcet are more favourable 
to centrist parties and also extremely similar. Borda (which tends to be 
unpredictable and throw up a minority of odd results) is most favourable to  
centrists.

Adding a PR method is good idea.


>> I wanted to use in my model an Approval strategy which Approval supporters 
(
>> of which I am not one) say will give a result that is satisfactory to the
>> voters. I decided to use Rob LeGrand's strategy A. Rob said the following
about
>> strategy A:
>>
>> > Strategy A: Approve all candidates I prefer to the current CRAB
>> > first-placer; also approve the first-placer if I prefer him to the
>> > second-placer.
> >>
>> > [S]trategy A always homes in on the Condorcet winner when one exists
>> > and all voters use the same strategy.

>I have a hard time calling this an Approval strategy, since it seems to 
require
>that CRAB be implemented...

Actually ( and bear in mind I am not an Approval supporter) strategy A is an 
extremely effective method of finding the Condorcet winner if there is one. I 
recently did a test comparing the result using strategy A based on a single 
Approval poll with the Condorcet result. Out of 250 seats 240 had a clear 
Condorcet winner and strategy A found the Condorcet winner in 233 of them (97%).


>I suspect you won't be able to model Approval convincingly, if the only
>information to begin with is rankings.

It is certainly complicated to model Approval from ranked ballot information, 
my original idea was to have logical/strategic and illogical/non-strategic 
voters. A certain proportion of the voter could be 'set' to use an Approval 
strategy that Approval supporters consider good (the logical/strategic voter) with 
the remaining voters (the illogical/non-strategic voter) voting for every 
candidate they approve.

David Gamble




    
    

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20031122/747f8768/attachment-0003.htm>