[EM] Critics and Advocates

[Adam Tarr]

John B. Hodges wrote:

>>Subject: Re: [EM] Cheering for simplicity/Orphan
>>From: Joe Mason
>>
>>On Thu, Aug 21, 2003 at 10:41:39PM -0500, Adam Haas Tarr wrote:
>>>  2) Since it is a Condorcet-compliant method, it shares all the 
>>> weaknesses that
>>>  all Condorcet methods have in the eyes of the IRV advocates (i.e. the weak
>>>  center winner).  I don't think an IRV supporter would see this as a 
>>> compromise.
>>
>>Since it requires multiple rounds, it also shares IRV's non-summability
>>problem, so there a lot of Condorcet supporters who wouldn't see it as a
>>compromise either.
>>
>>Joe
>
>(JBH) I've seen this attitude often on this list.
>
>It is a wonderful thing to be a critic. One's moral purity and social 
>superiority are assured, without ever having to get out of one's armchair.

John,

All I meant by what I wrote was that this method is not, in any real sense, 
a compromise.  A compromise would be something that met halfway (or three 
quarters of the way, or some distance along the way) between two 
positions.  In this case, those positions are IRV (fist place votes are 
sacrosanct, centrists should not have power simply by being a compromise, 
et cetera) and Condorcet (respect majority rule and count all preferences 
together, et cetera).

But here's the thing about that method - it's a Condorcet method.  There's 
no getting around this.  It satisfies the Condorcet criteria.  So, it is 
not in any real sense a concession to IRV-ists.  Just because the method of 
vote counting bears a cosmetic resemblance to IRV, doesn't mean the results 
will.  The entrenched IRV supporters would reject this method for all the 
same reasons they reject ranked pairs or beatpath.  And given the arguments 
they use to reject those methods, they're simply being consistent by 
rejecting this one as well.

So, that's all I was saying in the part of my message quoted above.  As I 
alluded to in other parts of my method, and as Joe noted in the next 
message, there's also some additional weaknesses of this method when 
compared to ranked pairs and beatpath.  The fact that some of these 
weaknesses are shared by IRV does NOT make this method a compromise between 
Condorcet methods and IRV.  It just makes it a bad Condorcet method.  After 
all, nobody supports IRV just because it's non-summable.

-Adam