[EM] Why I think IRV isn't a serious alternative 2

[Abd ul-Rahman Lomax]

At 05:31 PM 12/25/2008, James Gilmour wrote:

>It is not a question of my thinking in terms of plurality  -  that 
>is where our electors (UK and USA) are coming from.  It is my
>experience (nearly five decades of campaigning) that UK electors 
>attach great importance to their first preference.

*Of course they do.* At least the majority do; more accurately, some 
do and some don't, with the majority having a strong preference for 
their first preference, over all others. There is a feedback between 
single-winner plurality, or other strong two-party system, and the 
strength of preference for the favorite:

>   You may say
>that's the result of bad conditioning, but if we want to achieve 
>real reform of the voting systems used in public elections, these
>are the political inconveniences we have to accommodate.

Absolutely. This is why I concluded that Bucklin was the place to 
start. The only argument for Approval that might prevail in some 
places is that it's cheap. The strong preference for the first 
preference will result in more disuse of additional rankings with 
Open Voting -- Approval -- than with Bucklin. But Bucklin provides 
sufficient protection for the first preference, in my opinion. And 
this is the question that you have not been asking, you have been 
asking within the assumptions of other methods and the presentation 
of Later No Harm within those assumptions.

And you need to ask yourself, first, you seem to be quite ambivalent, 
confusing your own position with political expedience. That's a form 
of strategic voting, isn't it?

A 5% Condorcet winner could possibly be a disaster, or could possibly 
be a great relief. Which is more likely? Doesn't it depend on the 
conditions that led to it? If a condorcet winner only gets 5% first 
preference votes, what was the system? What was the overall voting 
pattern? It's quite possible that *no* outcome of this election would 
be other than a disaster!

Looking at this in isolation is, for you, projecting present 
experience onto a situation where present assumptions and conditions 
don't apply. Pretty easy to make a drastic mistake, doing this.

Want to consider election scenarios? You *must* consider sincere 
preference strengths, which is the same as saying that you must 
consider underlying utilities. 5% Condorcet winner tells us almost 
nothing about this. So you are taking a situation where we know 
almost nothing, and confidently predicting chaos. If it's 5% first 
preference, with twenty candidates, similarly to what was noted 
originally, the Condorcet winner might *unanimously* be considered an 
excellent compromise. The voters could be *very* happy with the result.

Or it might be very different. It depends on underlying utilities; 
and to be accurate, it depends on underlying *absolute* utilities, 
not merely relative ones.