If the voters know that ties are going to result in the speaker of the house, say, deciding the winner, then those voters who like the speaker's likely choice better than the likely CW may try to arrange a tie.
They can still try to arrange a tie with Eppley's method as well, forcing a random choice to be made.
I don't see how this forms a convincing argument either way.
I would still want to see the results expressed as a tie, rather then have them hidden from me.
---- Election-methods mailing list - see http://electorama.com/em for list info
