On Mon, 13 Dec 2004, Shane Rush wrote:
Julian Mehnle says in spf-council,
The state of the art of voting methods are ranked voting methods, more
specifically Condorcet voting. Every voter may declare a ranking of all
the available options, from most desired to least desired. Then a
somewhat complex tallying process is performed, possibly using some
tie-breaking method, resulting in a set of winning options of the desired
size.
I am very much against this idea, though I understand a desire by some
to make changes to produce solid results. My initial thought is that
the last thing that is needed is 'complex tallying process' -
simplicity and an open understandable tally means that like the first
vote everyone can check their own vote and do a tally themselves -
even if/when next time the origin of each vote is aliased.
Would you be willing to consider using Condorcet only as a way to
determine winner in case of a tie? It would work as follows:
1. Each voter selects 5 candidates and puts them in order of preference
2. The total votes are counted and if there are no ties the 5 candidates
that got most votes are declared a winner.
3. If several candidates had same number of votes, then condorcet is used
to determine their ranking among each other (i.e. who was prefered more
over the other). If there is still a tie (not very likely) a runoff
is held.
Condorcet voting represents the voter's preferences optimally and
is strategy-free.
I obviously need to do some proper research for Condorcet but a quick
back of envelope test shows me that while Condorcet voting may be
strategy-free for a large constituency when it comes to the small number
of votes being cast here then strategy is instead magnified.
I agree. We do not have enough voters and condorcet also has some
weaknesses when it comes down to that more then one candidate is to
be selected.
--
William Leibzon
Elan Networks
william(_at_)elan(_dot_)net