[EM] A different "strategyproofness" notion

[Warren Smith]

>> A voting system is "naive-exag-proof" if its winners do not change when the
> labeling of two candidates as "frontrunners" is altered to another two
> and all (or in another version, just some, the rest being honest) voters use
> naive-exag strategy only.
> This is a far weaker notion than Gibbard/Satterthwaite "strategyproof."
>
> Are there any interesting naive-exag-proof voting systems?

>Systems that allow for continuous vote shifting, where the voters can
see the changing results in real time, are naive-exag-proof.  In such
systems, the differences between perception and reality vanish.  A
candidate is either a frontrunner, or is not.  Labels to the contrary
will be ignored, and will have no effect on the results.

--not necessarily.
Your suggestion can get into a "perpetual oscillatory cycle" without
ever reaching a steady state.

Some such systems might be immune-to-oscillation though.  But
certainly not all of them.

--And: possible replies to those worried naive-exag-strat is not
necessarily "rational" are:
(1) in real life, naive-exag-strat is highly used, wheres what the
inviligators claim are "rational" strategies, are considerably less
common.   I made no claim whatever naive-exag-strat was or was not
rational.  (And I continue to make no claim on that right now,
so do not tell me I just did. OK?)
I AM claiming that naive-exag-strat is IMPORTANT (and more so than the
inviligators notions of rational) hence worth studying.

(2) If you want to study rational strategy, there already has been
such a study. The concept is called "strategyproofness."  The whole
point of the study I am suggesting, is
to avoid that historical dead-end by NOT considering
strategyproofness, but merely immunity to naive-exag-strat.

Mind you, even my suggestion might similarly still be a dead-end
leading to an impossibility theorem.  But that is an open question.  I
was simply raising the question.

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