Having ideas is great.
What we do should be based on whether we see them as worth the effort.
My proposal for adding an array to the N*N matrix was based on both
reducing effort and expecting less errors in recording desires.
Dave Ketchum
On Jul 7, 2010, at 11:33 AM, robert bristow-johnson wrote:
On Jul 7, 2010, at 2:36 AM, Kristofer Munsterhjelm wrote:
robert bristow-johnson wrote:
i still think that rectangular N*N matrix is sorta useless. it's
hard to read. each pair should be grouped together for visual
inspection.
How do you handle the case where some voters have no preference
between certain candidates? In that case -- say someone voted C =
D, then (C>D + D>C) is not equal to the number of voters. If you
publish only C>D, then there's no way to know whether someone equal-
ranked him or not.
so, maybe instead of two little lines per pair, there are three.
(actually another line for defeat strength would be useful.)
C>D 2936
D>C 1592
C=D 129
<1344>
so, i guess that would mean that 129 voters ranked C and D equally
somewhere above the absolute bottom (since unranked and unmarked
candidates are tied for last place on any particular ballot). if
those 3 numbers add up to less than the number of voters who
received and cast ballots, then the difference would have to be
voters who didn't rank C and D at all (and left tied for last place).
the point is that everything you have regarding a pair of candidates
is grouped together in one place for easy inspection. i do not
think that rectangular NxN matrix lends itself to quick and easy
inspection by a lay person.
--
r b-j [email protected]
----
Election-Methods mailing list - see http://electorama.com/em for list info
