[EM] Approval-Strategy article at CES website

[Juho Laatu]

> On 04 Nov 2016, at 21:21, Michael Ossipoff <email9648742 at gmail.com> wrote:

> Better to just use the previous election as the poll.

Some individuals might do so. But I'm sure there will be some kind of poll style information available, and some voters will see them. There's thus an opportunity to influence the outcome of typical approval elections by influencing the polls. Poll strategies should be seen as one type of vulnerability of approval elections.

> You keep saying that, voting only for the strong top-set, you might not vote between the actual two frontrunners. And then I keep repeating that it's optimal to approve (only) your strong top-set, because, by the definition of the strong top-set and strong bottom-set, the merit difference _within_ those 2 sets is negligible in comparison to the merit-difference _between_ those 2 sets.

Yes, this seems to be about the only difference in how we talk about optimal strategy.

If all the potential winners are either in the top-set or in the bottom-set, and all candidates within those sets are equal, then the optimal strategy could as well be to stay home and not vote (following the normal approval strategy, and "default target", see below). On the other hand, if there is some difference in preferences within those sets, then the optimal approval cutoff point can be within those sets (still assuming that some voters believe that the frontrunners are in one of those sets). Uncertainty of poll information does not change this.

In some of your comments you seem to say also that the voter does not want to approve any candidates in the bottom-set. If the voter behaves that way (which is probably quite common), then his/her strategy is not the normal optimal approval strategy, but something different. In typical technical EM discussions the "default target" of the voters is to maximize the expected quality of the winner. That means that by default they do not to pay any attention to their other feelings (= other targets), like whether they should approve / not approve some candidates in the bottom-set / top-set. Many voters certainly do vote according to their strong feelings, but that is not in line with the optimal normal approval strategy (in the "default target" sense). It would be good to point out in the discussions what targets the voters are expected to follow (at least if they differ from the "default target").

Juho