3 Candidate Condorcet Simulations E results

Steve Eppley seppley at alumni.caltech.edu
Tue Dec 3 12:44:07 PST 1996


DEMOREP1 wrote:
>Mr. Eppley wrote:
>>If it doesn't take much time, how about doing these 3-candidate random
>>simulations:
-snip-
>Again, such 24.0, 25.3 and 50.7 split is highly suspect in a real election.

I thank the distinguished gentleman from aol for running those
simulations, and for his quick response.

I agree the splits are suspect.  I think people who bother to vote
would rarely be utterly indifferent in any pairing, and I'd hope
enough people would understand ranked ballots to be able to represent
their preferences accurately if they want.

Also suspect is the randomness of the preference orders in these
simulations.  For example, I'd expect in a real election the numbers
of A>B>C votes and A>C>B votes would be highly unequal, since voters
who like A the most would often tend to analyze B vs C in similar
ways.  But in a random simulation, the number of A>B>C votes is
going to be close to the number of A>C>B votes, just as the number
of tails and heads are close in a series of coin tosses. 

---Steve     (Steve Eppley    seppley at alumni.caltech.edu)




More information about the Election-Methods mailing list