[EM] [CES #3389] SODA strategy

[Warren Smith]

>> Well... good question.  I don't know the answer.   Perhaps the answer
>> is some computer simulation study, but that isn't so easy nor will it
>> necessarily be so clear (people will
>> complain: "why didn't your study include voting method X?").
>>
>>
> Can you expand on how you'd design that study? The fact that it won't
> satisfy everybody, isn't a reason for me not to want to do it. My primary
> reason to want to do this is to honestly test my ideas, not just to create
> propaganda with foregone conclusions.

--well, I don't know exactly, but the general idea, is make a class of
pretty realistic elections, sample random examples from the class...
and then analyse each vis-a-vis
strategy for both SODA and a number of competing voting methods.

The part after the ... is particularly unclear.   You could just get
IEVS, add SODA to the
voting methods inside IEVS, and try to do a BR analysis using whatever voter
strategic behaviors IEVS already has, plus whatever more ones you want
to implement.

Try not to be too horrified by the bugs and ugliness inside IEVS...
but it's a good start
despite imperfectness...  and if somebody else worked on it for a
while they'd hopefully make it less ugly and more complete.

http://RangeVoting.org/IEVS

-- 
Warren D. Smith
http://RangeVoting.org