Counting unranked candidate pairs as zero votes each, Bart's example results are:
        wv ABC    mar ABC
C       T   NT      T   NT
B T    A/C B       A/C A
A NT   B   B       C   B

Where A/C is a tie.  Both ABC and CBA either gain nothing or lose by truncating
(for both wv and margins) given that A/C and B have equal utility.  Adam's claim
survives Bart's challange. 

If you give the unranked pairs 1/2 vote each then the wv result becomes identical to the margin result as shown above.  My conjecture is:  If you give multiple unranked candidates 1/2 vote each then both wv and margins have the no strategic truncation incentive (NSTI) property.  If, as I maintain, there is no overall disadvantage to giving 1/2 vote each to multiple unranked candidates, and if my conjecture is correct, then there is no basis for claiming that this NSTI property favors wv over margins.

Reply via email to