[EM] A different "strategyproofness" notion

[Warren Smith]

I did not claim naive-exag-strategy is always rational.

I merely asked what are the consequences of assuming people do it.

In fact, in Australia, it appears at least about 85% of the voters use
naive-exag-strategy with IRV voting -- despite the Lundell claim it is
not rational since IRV satisfies later-no-harm.  This percentage is
easily sufficient to prevent third-party candidates from winning IRV
seats.  And indeed, in the last 3 Australian house election cycles
(150 IRV seat-elections per cycle) combined, the total number of
third-party candidates elected, was zero out of 450, despite an
average of about 7 or 8 candidates per seat.

It appears real-world voters do not care much about what Jonathan
Lundell considers to be rational.   Personally, I wish (and I daresay
he wishes) voters were more like me (him).  But they are not.





On 10/19/09, Jonathan Lundell <jlundell at pobox.com> wrote:
> On Oct 19, 2009, at 8:50 AM, Warren Smith wrote:
>
>> But this leads to another interesting idea.   Consider this "naive-
>> exag"
>> voter strategy:  rank the two frontrunners A&B (where you prefer A>B)
>> top & bottom, then anybody better than A is ranked co-equal top (or if
>> that forbidden, then just below A)
>> anybody worse than B is ranked co-equal bottom (or if that forbidden
>> then just above B)
>> and anybody between A&B is ranked between using honest normalized
>> utility.
>
> IIRC, this is a variation on the "bury B" strategy that Clay uses for
> his regret simulations. But such a strategy would be anti-rational for
> any method (STV, for example) that satisfies later-no-harm. For
> example, if my sincere preference is A>B>everyone-else, why would I
> strategize A>everyone-else>B in, say, an IRV election?
>


-- 
Warren D. Smith
http://RangeVoting.org  <-- add your endorsement (by clicking
"endorse" as 1st step)
and
math.temple.edu/~wds/homepage/works.html