On Fri, Jan 28, 2011 at 1:59 AM, Kristofer Munsterhjelm <[email protected]> wrote: > But to sum up after that digression: unless you add in voter ignorance, > equal-rank would vanish under the model, and truncation seems too arbitrary > to implement. Thus I'm staying with full rank. I may consider the other > things later, but I haven't even written the program to do full rank yet.
I'm not sure you can attribute equal rankings to "voter ignorance", I think it's perfectly rational to rate two candidates as roughly equal from your point of view, even if they are rather different. And truncation could either be apathy, or a rational voter strategy depending on the method used to pick a winner. But for the purpose of high-performance generation of Yee diagrams using the "standard" voter model, I think sticking to non-truncated fully-ranked ballots is a perfectly reasonable simplification for the time being. The FFT approach sounds rather interesting as well, though. Also, I wonder if there is a good method for knowing where the noise will be and, after the main computation is, sample those regions to add the noise back in. This seems rather tricky, given Warren's observation of the noise patterns in a pentagonal candidate configuration using IRV. But that's putting the cart before the horse. Best, Leon ---- Election-Methods mailing list - see http://electorama.com/em for list info
