[EM] [Fwd: No response was requested from you]

[Richard Moore]

Dave Ketchum wrote:

> On the other side, "Two levels" puzzles me - if I had more time for this
> group I might have got this from context, but I could have been asking
> the same question (e.g., this could have been tied to something for
> which Craig and I cannot picture levels being involved).


I'm not sure where the difficulty is; it's not a major 
abstraction, but perhaps the term has multiple uses. Buddha 
certainly got it in his post. Approval voting uses a 
two-level ballot: Each candidate can get either a mark 
(approved, or Buddha's "A" level) or no mark (disapproved, 
or Buddha's "B" level).

Condorcet methods use ballots with N levels when there are N 
candidates. So do IRV, Borda, or any other fully-ranked 
method. Ratings methods use M levels, where M is arbitrary 
(and again, Approval is a special case where M=2).

If you constrain voters to a two-level ballot, and tabulate 
the results into a pairwise matrix and evaluate the matrix 
with the same method that you would evaluate it in a 
Condorcet method, then you will pick the Approval winner. So 
Condorcet and Approval both can use the same counting 
method, but they use different ballots.

Craig may have a genuine problem understanding some of the 
statements made on this list (though most of the time I 
think it's plain disingenuity, or that's how it comes 
across), but his bigger problem is that he just doesn't know 
how to ask for clarification nicely. Start with ad hominems 
-- against the entire list, mind you -- and add in 
non-sequiturs such as those about Martin and "P" to confuse 
the issue, and exclude any contextual information that would 
help express what his question is, and you can see why 
rational people get frustrated with this behavior. If 
there's a genuine inquiry in there, why must it be wrapped 
up in the language and tactics of a troll? And what's with 
the unsolicited off-list messages?

Richard