[EM] Appreciation for KM.

[Kristofer Munsterhjelm]

On 2024-03-24 22:41, robert bristow-johnson wrote:
> 
>> On 03/24/2024 3:03 PM EDT Kristofer Munsterhjelm <km_elmet at t-online.de> wrote:
>>
>> How would you suggest that such a calibration be done?
>>
>> If it can't be done, then Range's IIA compliance, while nice in a
>> box-ticking way, doesn't really do what it implies it does.
>>
> 
> <clapping>

Thank you :-) Sometimes I feel like I'm just flinging my posts into the 
void because there's so little response. Granted, a lot of my recent 
posts have been quite heavy on the theory. But still. Thanks.

>> On 03/24/2024 3:07 PM EDT Kristofer Munsterhjelm <km_elmet at t-online.de> wrote:
>>
> ...
>>
>> What do you think about the following reasoning? Call it the "fool me
>> twice" problem.
>>
>> Suppose that a jurisdiction is considering switching from FPTP to Borda.
>> The main organization is heavily marketing Borda as the one ranked
>> voting system, equating the ranked ballot format to Borda, the method.
>>
>> Meanwhile, an organization promoting MDDA (majority defeat
>> disqualification approval) is slowly growing. Someone (call him John)
>> favors MDDA and thinks that due to Borda's clone problem, it will
>> quickly be repealed. Then, he reasons, the jurisdiction will think that
>> ranking equals Borda, so that when some other ranked method is proposed
>> (MDDA, say), they will remember the failure of Borda that led to its
>> repeal and say "no; fool me twice, shame on me".
>>
> 
> This reflects my attitude perfectly.  Course corrections are less
> costly early in the voyage.  Before Hare RCV gets entrenched deeper,
> it's best to recognize these unnecessary flaws in IRV demonstrated by
> the Burlington and Alaska RCV elections that may both get repealed
> within a couple of years of an anomalous election with a lotta people
> that wonder if it was an "honest" election.  And then head the problem
> (the consequences IRV failing to elect the CW) off before it occurs again.
> 
> Now unfortunately for Burlington Vermont it's "Fool me once, just
> wait 12 years and we'll forget."  And it's unlikely (because the GOP
> is now so weak in Burlington) to happen again, so these deniers will
> feel validated in time.  But over the U.S. (or the world), we know
> that this Center Squeeze failure will happen again with the
> consequence of a spoiled election and sincere voters getting burned.
> The problem is if this failure isn't repeated at the same
> jurisdiction where some of the same people ask themselves "weren't we
> here before?", then it's unlikely that enough voters will notice the
> flaw (and its consequences) that very rarely happens.  I think
> FairVote is banking on that.

That, I think, is the other major reason to not go for a particular reform.

There's a chance that a method could be strong enough to appear to be a 
significant improvement, yet fail under additional pressure. If it 
doesn't lead to a repeal, the voters could still draw the conclusion 
that strategy is still necessary; or it could reward preexisting 
strategies like FPTP compromising enough that they linger on, or third 
parties could just grow weaker in the meantime, as you mentioned.

In any case, the method could then slow down further reform because it 
would take time before its flaws become apparent.

At least for this kind of failure, it'll become a lot more obvious if, 
say, one state implements IRV and another implements Condorcet, and then 
"coincidentally" the latter has multiple parties while the first 
doesn't. So it's not quite as damaging an outcome as the "fool me twice" 
situation.

That is, unless the not-good-enough methods pull harder than the good 
methods push. E.g. Australia's use of IRV pull harder towards two party 
rule than its use of STV pushes away from it.

-km