[EM] monotonicity criteria

[Michael Ossipoff]

We've been discussing four added-ballot monotonicity criteria:

Listing them in order of worst ones (to fail) first:

Mono-Add-Plump
Mono-Add-Solo-Top
Mono-Add-Top
Participation

Woodall describes more than that, including some "append" criteria
dealing with bottom-end changes, which I don't consider as important.

I didn't find Mono-Add-Solo-Top in Woodall's paper, though maybe you
will. I'm continuing to use that name for it here, because it fits in
well with Woodall's naming.

It's of interest to compare some methods with regard to the four
above-listed criteria.

In the table below, the methods are on the left, and the criteria are
on the right.

A method fails a criterion if it is separated from it by a horizontal
line that is below the method:

TUC, ICT, MJ.....| Mono-Add-Plump
------------------------------------------------
IRV.....................| Mono-Add-Solo-Top
---------------------------------------------------
...........................| Mono-Add-Top
Approval, Score.| Participation


Note that TUC fails, but Approval passes, Mono-Add-Solo-Top, which
could be called "Make-Your-Favorite-Lose"

Now, someone could argue that IRV fails Mono-Raise.

Mono-Raise and Partilcipation are exact changed-ballot and
added-ballot counterparts to eachother.

Someone could say, "IRV fails that changed ballot criterion and some
added-ballot criteriia, while TUC only fails added-ballot criteria."
But, for one thing, TUC fails more added-ballot criteria. For
anotother thing, based on the paragraph before this one, the criteria
that TUC fails are worse than the ones that IRV fails.

Mike Ossipoff