[EM] Condorcet How?

[Abd ul-Rahman Lomax]

At 05:57 AM 4/8/2010, Raph Frank wrote:
>On Wed, Apr 7, 2010 at 11:25 PM, Dave Ketchum 
><davek at clarityconnect.com> wrote:
> > Write-ins permitted (if few write-ins expected,
> > counters may lump all such as if a single candidate - if assumption correct
> > the count verifies it; if incorrect, must recount).
>
>How do you handle write-ins.  Are write-ins assumed to be equal last
>on all ballots which don't mention them?

Yes. Average Range will treat them as abstentions from rating, but as 
votes, they are problematic. Only Asset Voting can truly fix this 
problem. However, there is another solution: require a majority. In 
that case, with good runoff rules, a write-in could get into a runoff 
election by causing majority failure, at some threshold or standard, 
one designed to catch write-ins that might win, given a chance. My 
proposal is to implement Bucklin as a runoff voting system and thus 
start to collect data that could then be used to determine future 
reforms. If the runoff allows write-ins, and the first election 
results show promise, a write-in candidacy at that point would be one 
where other voters were informed. Write-ins in a Bucklin runoff with, 
say, no more than three candidates, and a serious poll preceding it 
as the primary, is very interesting.

And then, if this is put on the table, we will clearly run into the 
fact that established power almost certainly doesn't want write-ins 
to be viable, nor does it want third parties to have a chance. 
Instant runoff voting, almost certainly, confines winning to two 
major parties except in multiwinner elections (where it really isn't 
"instant runoff," it's different.)

In real runoff elections, write-ins sometimes win, even without a 
write-in runoff. All they have to do is make it up to second place. 
In fact, famous pathological elections are based on this, because the 
Condorcet winner got bumped down to third place. Lizard vs. Wizard. 
If the Wizard had been a write-in candidate, this would have been an 
example, but, since it was close between Duke (the Wizard) and 
Roemer, Duke wouldn't have made it into the runoff, but it would have 
been Roemer vs. the Lizard, and Roemer would have won, certainly.

Bucklin would easily fix elections like this; and good runoff rules 
would detect a viable third candidate and include him or her. If it's 
top-two, then, for sure, write-ins should be allowed. With a Bucklin 
runoff, the voters who prefer a write-in (and they would have been in 
the majority, I believe, in Lizard v. Wizard) would have written in 
Roemer. And would have put in bottom approved rank, the Lizard. Duke 
would have ended up in third place in the end, even if he didn't get 
dropped in a Bucklin primary.