Solutions for Open Elections

[Mike Ossipoff]

First of all, in Bruce's example, there doesn't seem to be any
strategy that the C voters should have used, even if they'd known
what would happen. The real C voters, who voted C in 1st place
effecively didn't vote a 2nd choice. The voters who voted B & C
equal in 1st place wouldn't care which won, and would have no reason
to try to thwart the B strategy.

So the real problem there is that, though C is Condorcet winner,
it's real 1st choice support is poor, and that's why it lost.
As I said, with only 26/101 of the people preferring it to B,
no one has a right to complain when it loses to B. At least no
such complaint can be based on majority rule.

***

Anyway though, if an election is an open election where the 
results are posted as they come, in terms of how many people
voted what ranking, and what is beating what by how much, it's
easy to avoid a problem in situations where it _does_ matter:

If, at the scheduled time for closing the polls, every alternative
in the set of alternatives from which Condorcet's method is to 
choose (this could be the Smith set or it could be the entire set
of alternatives) has another alterntive ranked over it by more
than half of the voters, then the voting period is extended by
a specified amount, and voters are allowed to change their votes.

That way, if order-reversal were used by voters who knew how others
had voted, their information would be useless, since their intended
victims would observe what happened and could change their votes.

But it might be simpler to instead, under the conditions named in
the paragraph before last, to simply hold a 2nd election, by
Approval, among the alternatives from which Condorcet's method
was to choose. That wouldn't require puntive strategy. On the
other hand, the vote-changing solution would provide better
deterrence.

***

Obviously these 2 solutions could also be used, under the same
conditions, in a meeting show-of-hands Condorcet election.

***

Mike





--