[EM] Free riding

[Raph Frank]

On Thu, Sep 4, 2008 at 10:13 PM, Juho <juho4880 at yahoo.co.uk> wrote:
> The point was really that the
> ordering of all the candidates should be re-evaluated based on the estimated
> probabilities and utilities.

Yeah, lower probability candidates should be moved upwards and higher
probability candidates should be moved downwards.

I would say that the first is Woodall and the 2nd is Hylland.
However, in a sense, moving high probability candidates downwards has
the same effect as moving low probability candidates upwards.

> Meek like systems will give different
> probabilities than those that allow also Woodall style free riding.

True.  However, I think it also reduces the benefit of Woodall.

> One simple approach would be to follow a candidate given inheritance order
> (=> trees or explicit candidate given inheritance lists) instead of multiple
> voter given inheritance orders. That would not be really STV any more, but
> at least a reference point to compare with (and maybe to start finding some
> intermediate solutions with optimal properties from both sides).

I think the benefit of STV is that it allows the voters to decide if
parties or candidates are more important.

This seems to be at the cost of potential Hylland free riding.