[EM] Markus, 2 March, 1349 GMT

[Markus Schulze]

Dear Mike,

I wrote (28 Feb 2005):
> Suppose your sincere preference is A>B>C>D>E. Suppose in
> situation #1, candidate A is elected with a probability of
> 60% and candidate B with a probability of 40%. Suppose in
> situation #2, candidate A is elected with a probability of
> 70%, candidate B with a probability of 20%, and candidate C
> with a probability of 10%. How does the used election method
> know which situation you like better, when you can cast only
> rankings and not ratings?

You wrote (2 March 2005):
> I don't understand what you're trying to say, what your
> example means. But I've never said that "the used method"
> should know what situation (?) you like better. Check the
> definition of FBC, and you won't find mention of that.
>
> There are 2 answers to Markus's question, depending on whether
> we call his lotteries the outcomes, or whether we call the
> single winner after the lotteries the outcomes. I prefer the
> latter definition of outcome.
>
> If we define outcome in that latter way, then we don't have
> to answer Markus's question about which lottery the voter
> prefers to the other, because they aren't outcomes, as I
> define outcomes.
> 
> I define an outcome as a single-winner. Markus's lottery
> outcomes, are, by my definition, not outcomes, but only
> intermediate points in the method's selection.
>
> We can call that a clarification of FBC, but I suggest that
> it's also the presumptive default definition of "outcome".
> How can Markus say that there's an outcome, when people
> are waiting to find out who will be president? The lottery
> clearly is not an outcome. The selection process clearly
> is not over till the lottery has given a winner. That´s
> when there´s an outcome. When it isn't known who the
> president will be, there's no outcome, by any reasonable 
> definition of "outcome".
>
> If a method chooses lotteries such as Markus described, and
> the method stipulates that the lotteries be carried out,
> to find the single winner, then clearly the lotteries are
> part of the method, and not the outcome of the method.

In short: You use the resolute model.

I suggest (1) that you shouldn't use the term "outcome"
in your definition of FBC when you only mean the winner
or (2) that you should write explicitly that you use
the resolute model.

Markus Schulze