[EM] Probability of losing a sincere Cond. Winner

[matt at tidalwave.net]

On 25 Nov 2002 at 13:13, Adam H Tarr wrote:

> The more important issue is that it's very easy for these problems to 
> crop up in margins, and it's easy to recognize them when they do, and 
> it's easy to foul up the results just by truncating (which is a very 
> natural thing for the voter to do).  In the winning votes examples, the 
> problems only show up when there's a lot of truncation to start with, and 
> then they only show up in a pretty fractured race where it would probably 
> be hard to tell whether truncation will help or hurt you.

I understand and I agree that reducing incentive for truncation is important.  A 
downside is that if more than one disapproved option is ranked then at least one 
disapproved option will receive more votes then any nonvoted "no opinion" options.  
Thus, voter's who both disapprove and have no opinion of one or more options 
each should either rank just their approved optons and no more than one of the 
disapproved options or fully rank all of the options (with the no opinion options 
ranked equally at the approval cutoff).  The first approach is an emotionally difficult 
decision of how best not to fully express oneself, similar to the situation facing the 
voter in Approval contests, and the latter approach is onerous like Borda contests, 
especially when there are many options.  Faced with these two choices, voters will 
be tempted to take the easier third approach of just voting their full set of approved 
and disapproved and hope for the best even though this gives their disapproved 
options more votes than their no opinion options.

Nevertheless, I agree that the benefit of less manipulativeness with wv done this 
way, if substantial enough, is good reason for favoring wv.  And the benefit for 
undermining truncation strategy does appear to be substantial.  However, as Blake 
Cretney notes, the strategy opportunity, instead of being eliminated by wv, is moved 
elsewhere.  The winning strategy for wv becomes random voting of the lower ranks.  
This is obviously correct since random voting of the lower ranks "turns wv into 
margins" like 1/2 votes does.  Granted, truncation is both an easier and more 
obvious strategy than random voting, but all it takes is some die tosses and coin 
flips behind the curtain or a little time with a home or office computer or calculator 
with a psuedo-random generator before going the polls once the word gets out.  So 
on balance I am still not convinced that the truncation stability benefit of wv 
outweighs the costs.

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em