http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=195938&mesg_id=195939My other thread on math and voting. Real interesting so far - a snippet:
We all appreciate the critical nature of this lost information. For proof, just recall almost any New Hampshire Presidential Primary where the news media describes how some candidate siphons votes away from another one with similar views. The candidates react with cries of ``Don't waste your vote, vote for me!" So, as captured by Borda's example, it could be that when most voters like several closely contending candidates, none wins because of how the voters split the vote!
Even worse, suppose after nine serious candidates are in the New Hampshire primary a radical, R, joins the election chase on, say, the ``Return to the Snowshoe" platform. Suppose R is so repulsive that almost 90% of the voters have R bottom-ranked. But if the views of the respectable candidates are sufficiently close to split the vote, R could win with only slightly over 10% of the vote coming from the lunatic fringe, the uninformed, or voters mistaking R for someone else (as once happened in Illinois). R, a candidate almost universally despised, could win! This is the problem that Borda worried about.
While this Snowshoe example is extreme, it suggests that we might find elections exhibiting Borda's phenomenon when the views of top-ranked candidates represent a small portion of the electorate. By checking, we discover that again Roemer and certain other candidates in the 1995 Louisiana gubernatorial election may have been victimized by the plurality vote. While the final two candidates did not suffer from scandal, it is difficult to claim that both represented the views of a vast number of voters. Similarly, recent Russian elections provide a rich source of examples.