[EM] Introduction (cont.)

[Richard Moore]

Roy wrote:
> That's a rather slavish devotion to unreliable probabilities. If 
> those candidates turn out to be contenders, that means our 
> probability assumptions were drastically wrong, and making strong 
> statements based on faulty data is a strategically bad move. At that 
> point, it would be more prudent (IMO) to vote as if you had zero 
> information.

After reading your arguments, it's clear that what we've 
talked about on this list in the past with regards to 
sincere and strategic CR voting has been in the context of 
zero-information cases vs. high-quality information cases, 
assuming that we'd either be dealing with one or the other. 
We weren't really thinking much about low-quality 
information cases. It's always good to have someone come 
along with fresh insight on these matters, so thanks!

Richard