[EM] Black box election method tests - is it possible?

[Markus Schulze]

Dear Jim,

you wrote (13 Feb 2005):
> Would it be possible to design a series of black box tests (using
> standardized inputs and expected outputs) which would determine whether
> or not a given election method satisfies the  list of various criteria
> that are deemed essential  for an acceptable election method?

No, unfortunately.

You wrote (13 Feb 2005):
> The tests should reveal flaws/ shortcomings/ failed criteria  in the
> results produced by the election method under test. Which of the
> criteria, if any, require knowledge of the internals of the counting
> method used?

In my opinion, the underlying heuristic and the used algorithm are
not part of the proposed method.

You wrote (13 Feb 2005):
> Would some criteria need to be weighted (given more importance) more
> than others?

Which criterion is how important is, of course, a subjective question.

Markus Schulze