[EM] Approval vs. IRV

[Brian Olson]

On Oct 18, 2004, at 7:09 AM, Bill Clark wrote:

> On Fri, 15 Oct 2004 11:03:28 -0700, Brian Olson <bolson at bolson.org> 
> wrote:
>
>> I think I'm allergic to the use of randomness in election methods, so 
>> I
>> don't plan on implementing such an option.
>
> The unique appealing feature of random methods is that they're the
> only ones that can be completely immune to the the "Tyranny of the
> Majority" problem.  If a particular political party consistently gets
> 10% of the vote, wouldn't it be nice if their candidates were elected
> 10% of the time?  Random methods are the only ones (that I know of, at
> least) capable of accomplishing that for single-seat elections --
> they're basically the single-seat "time-sharing" version of PR.

If they get 10% of whatever PR body, that's fine and there's no need to 
augment that with anything else.

For a single seat, I think the vast majority would be poorly served by 
four years of office holding by a tiny majority. I'm even more scared 
of tyrannies by minorities. (If you consider US non-voters and the 
current absolutist power-politics, we already have that. Rule by a 
quarter or less.) I think in all cases we're trying to move the 
democratic process to be more democratic and better represent as many 
people as possible. That's why we want an election method that can find 
the compromise choice that serves 60% of the people when we might 
otherwise get some faction's 40% or 41% choice.

If all of the choices are good enough, then randomly picking between 
them might be fine since it wouldn't really matter and random might be 
ok. Based on the current highly polarized political climate, a whole 
lot of people think that a whole lot of candidates are far less than 
good enough.

> Random methods are nothing to be concerned about, they've been used in
> mathematical physics and other hard-sciences for quite some time (e.g.
> the Monte-Carlo method).
>
> http://en.wikipedia.org/wiki/Monte_Carlo_method

Oh sure, great for science, bad for my sense of fairness. The coin-toss 
is at the beginning of a football match, not the end.

Obviously what we really want is to stick our candidates in blenders, 
and then mix proportionally and reconstitute a proper representative.

Brian Olson
http://bolson.org/