[EM] IRV strategy correction

[MIKE OSSIPOFF]

It was reasonable to suggest ranking the candidates in order of their
probability of being the beat compromise you can get, the best one
who has a pairwise majority against those whom you like less, but part
of my paragraph about that wasn't worded that well in my previous
message.

Of course that's easier said than done, and one must simply
guess at the candidates' order in that regard. Of course that isn't at all 
complete because it doesn't account for candidate utilities. Of course 
there's a
more detailed and accurate strategy for IRV, but it would be
quite complicated, a sort of Runoff to the nth power. Runoff's
mathematical strategy is already complicated enough, but IRV takes
it to a much greater extreme. It isn't just the computation, but
also it's all the probabilities that would have to be estimated.

But, when there are completely unacceptable candidates, ranking
the acceptable candidates in order of their probability of being
able, with your help, to deny victory to an unacceptable candidate
is a valid good strategy.

As with other methods, strategy is much simpler when there are
completely unacceptable candidates, or, at least, when there are
2 sets of candidates for whom the merit differences within the sets
are insignificant compared to the merit difference between the sets.

But I'd be ashamed if I'd started a discussion of IRV strategy, so
I hope no one but Don or Craig will reply to this message.

I'd probably not vote at all in an IRV election.

Mike Ossipoff


_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em