[EM] manipulation free method?

[Jan Kok]

This is a fascinating idea.

There might be a slight "bug" in the description. If a person ranks #1
a candidate who is estimated to be more than 50% likely to win, would
that ballot count as an approval for that  #1 choice? I would think
the #1 choice on each ballot should ALWAYS be approved.

As a person who likes Approval Voting (in large part because it passes
FBC), this method appeals to me. It might appeal to IRVists who like
IRV because they like ranking the candidates and because of the
perceived simplicity of IRV. It might appeal to those who don't like
plain Approval Voting because they don't like the idea that approving
a 2nd choice could cause the 2nd choice to win when their 1st choice
might have won otherwise. It might also appeal to those who don't like
Approval because it requires voters to who want to maximize their
voting power to estimate the likely winners, which might be difficult
in small nonpartisan races (such as city council seats) where there
can be several candidates of similar strength, and no one is taking
polls.

Cheers,
- Jan