[Election-Methods] Range voting simulations

[Ian Fellows]

Kevin,
The candidates that are considered frontrunners are not independent of voter
preferences. In fact front runner status can be thought of as based solely
on voter preferences. Strategy in a voting system is best thought of in
terms of game theory rather than as a single "best response strategy" to an
arbitrarily chosen "front runner" and "runner up."

Ian
http://thefell.googlepages.com


Hi Ian,

--- Ian Fellows <ifellows at ucsd.edu> a ecrit :
> Kevin,
>
> That seems to me to be an unrealistic assumption of voter behavior.

With my implementation it isn't about voter behavior, but see below:

> Strategic votes under this definition of "frontrunner" could then be
> based
> on the two least likely to be elected candidates. Admittedly, finding a
> strategic equilibrium for a voting system is difficult, and in most cases
> computationally expensive. The solution to this problem is not to pick
> two
> candidates at random, and base your strategies on the assumption that
> these
> are the most viable candidates. This is especially true of approval
> voting,
> as there is no non-strategic voting pattern (that is not to say that a
> strategic voting pattern is necessarily dishonest).

Perhaps I didn't understand your original post. I thought the identities of
candidates 0 and 1 were fixed, and did not vary by voter. Thus the fact
that candidates 0 and 1 are considered "frontrunners" is not a strategic
decision made by the voters, it is the information provided to them.

Kevin Venzke