[EM] Strong methods (was Re: 3 or more choices - Condorcet)

[Kristofer Munsterhjelm]

On 09/30/2012 11:47 AM, Juho Laatu wrote:
> On 30.9.2012, at 11.56, Kristofer Munsterhjelm wrote:
>
>>
>> In practice, that means: is cloneproof, passes independence of as
>> much as possible (independence of Smith-dominated alternatives,
>> say), and is monotone.
>
> These criteria could be one set of definitions of a good (sincere)
> winner. I usually do not assume the first two ones since there may be
> good (sincere) winners also outside those criteria. Monotonicity is
> maybe more natural in the Condorcet category.

There might be, but then again, there may also be better outcomes when 
the method does not get confused by vote-splitting problems (e.g. the 
Korean election).

Maybe one could say:
- Good outcome under honesty to start with something reasonable,
- protection against the method getting lost on its own or being 
deceived, and
- assurance that the method doesn't behave in a way that robs the result 
of legitimacy,

which would be the categories in reverse order.

>> I've put strong in quote marks because I know others may disagree
>> with my priorities. FairVote obviously doesn't consider the "having
>> weird result dynamics" part important as long as the strangeness
>> can't be exploited by deliberate strategy.
>
> If one looks positively at their criteria, maybe they put strong
> emphasis on the marketability of the method. That marketability may
> include some tendency to favour the large parties.

That, or they play a long game that involves getting the voters used to 
IRV so they can then slip in STV.

In my opinion, even if that works, it won't have the desired effect. 
Australia shows this.

>>
>> The second is resistance to noise and strategy. Independence of
>> clones fit here, as well as independence of X (Smith-dominated
>> alternatives, Pareto-dominated alternatives, weak IIA). The
>> resistance may protect against strategy - cloneproof methods keep
>> parties from running an army of identical candidates - or improve
>> the outcome when there is no strategy - e.g. by not being affected
>> by the liberal parties' vote-splitting in a replay of the 1988
>> South Korean presidential election.
>
> (I just note that independence of clones can be an interesting topic
> both when discussing behaviour with sincere votes and strategies.)

Right. I don't think Tae-woo plotted to have Young Sam and Dae-jung 
split the liberal vote. Nor do I think the Republicans funneled funds to 
Nader in 2000. Thus I'd call that noise rather than strategy, though 
perhaps "noise" doesn't really fit, either.

>> The third is quality of the outcome under honesty, according to
>> some metric or desired logic. It's hard to say which metric one
>> should pick, unfortunately, and for Ranked Pairs (and Schulze),
>> there's probably no simple metric that the method optimizes.
>> Furthermore, the logic one uses for rated methods probably wouldn't
>> directly fit onto rank methods (because utilities are either
>> unknown or not applicable).
>
> It seems that I assumed above that this category and the first
> category are related. Maybe this category implies also the first
> category. I.e. there is no such good logic of what we desire that
> would break against the first caregory. (Or maybe, if we step outside
> the Condorcet domain and think about IRV, then maybe the idea of
> kicking the weakest candidate out at every round makes sense in some
> setup??)

Say you had a method that was independent of covered alternatives, and 
this method otherwise optimized similarly to Range - i.e. would tend to 
pick the uncovered candidate with best mean Range score on all possible 
Range elections that you could transform the current ranked election 
into. This might satisfy the third category from a utilitarian point of 
view, and the second (by its independence of covered alternatives). 
However, it would fail the first because ICA implies nonmonotonicity.

So the difference between the third and first category is, I think, that 
the third is about what's good for society in general, while the first 
is about what makes the voters (and candidates) accept the outcome. The 
more democracy is about having the losers accept that they've lost, the 
more important the first category becomes with respect to the third, for 
instance.