Dear Forest! Your grand compromise sounds very interesting and will make a good occupation for me this weekend I guess :-) By the way, I was quite annoyed to find a very simple example of only six candidates in which the condorcet lottery probabilities are all but monotonic (which I will also post in the next days). Perhaps your proposal has better monotonicity properties...
Yours, Jobst PS: I don't know of an efficient way to determine the Dutta set but perhaps we will find one. However, whether there is a 3-element Dutta set can be established in O(n^3) time by testing each pair for the existence of a third which beats both. ---- Election-methods mailing list - see http://electorama.com/em for list info